CDO User Guide

Climate Data Operator

Version 2.3.0
October 2023

Uwe Schulzweida – MPI for Meteorology

Contents
1. Introduction 7
1.1. Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.1. Unix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.2. MacOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.1.3. Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2. Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.1. Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.2. Environment variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.3. Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.4. Parallelized operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.5. Operator parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.6. Operator chaining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.7. Chaining Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3. Advanced Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3.1. Wildcards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3.2. Argument Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3.3. Apply Keyword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4. Memory Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5. Horizontal grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.1. Grid area weights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.2. Grid description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.3. ICON - Grid File Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6. Z-axis description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.7. Time axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7.1. Absolute time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7.2. Relative time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7.3. Conversion of the time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.8. Parameter table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.9. Missing values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.9.1. Mean and average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.10. Percentile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.10.1. Percentile over timesteps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.11. Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2. Reference manual 28
2.1. Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.1.1. INFO - Information and simple statistics . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.2. SINFO - Short information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.1.3. XSINFO - Extra short information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.1.4. DIFF - Compare two datasets field by field . . . . . . . . . . . . . . . . . . . . . . . 33
2.1.5. NINFO - Print the number of parameters, levels or times . . . . . . . . . . . . . . . 34
2.1.6. SHOWINFO - Show variables, levels or times . . . . . . . . . . . . . . . . . . . . . . 35
2.1.7. SHOWATTRIBUTE - Show attributes . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.1.8. FILEDES - Dataset description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.2. File operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2.1. APPLY - Apply operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2.2. COPY - Copy datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.2.3. TEE - Duplicate a data stream and write it to file . . . . . . . . . . . . . . . . . . . 41
2.2.4. PACK - Pack data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2.5. UNPACK - Unpack data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2
Contents Contents

2.2.6. BITROUNDING - Bit rounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

2.2.7. REPLACE - Replace variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2.8. DUPLICATE - Duplicates a dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2.9. MERGEGRID - Merge grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2.10. MERGE - Merge datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.11. SPLIT - Split a dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.2.12. SPLITTIME - Split timesteps of a dataset . . . . . . . . . . . . . . . . . . . . . . . . 47
2.2.13. SPLITSEL - Split selected timesteps . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.2.14. SPLITDATE - Splits a file into dates . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.2.15. DISTGRID - Distribute horizontal grid . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.2.16. COLLGRID - Collect horizontal grid . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.3. Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.3.1. SELECT - Select fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.3.2. SELMULTI - Select multiple fields via GRIB1 parameters . . . . . . . . . . . . . . . 54
2.3.3. SELVAR - Select fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.3.4. SELTIME - Select timesteps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.3.5. SELBOX - Select a box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.3.6. SELREGION - Select horizontal regions . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.3.7. SELGRIDCELL - Select grid cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.3.8. SAMPLEGRID - Resample grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.3.9. SELYEARIDX - Select year by index . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.3.10. SELSURFACE - Extract surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4. Conditional selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.4.1. COND - Conditional select one field . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.4.2. COND2 - Conditional select two fields . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.4.3. CONDC - Conditional select a constant . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.4.4. MAPREDUCE - Reduce fields to user-defined mask . . . . . . . . . . . . . . . . . . 66
2.5. Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.5.1. COMP - Comparison of two fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.5.2. COMPC - Comparison of a field with a constant . . . . . . . . . . . . . . . . . . . . 69
2.5.3. YMONCOMP - Multi-year monthly comparison . . . . . . . . . . . . . . . . . . . . 70
2.6. Modification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.6.1. SETATTRIBUTE - Set attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.6.2. SETPARTAB - Set parameter table . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.6.3. SET - Set field info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.6.4. SETTIME - Set time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
2.6.5. CHANGE - Change field header . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.6.6. SETGRID - Set grid information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.6.7. SETZAXIS - Set z-axis information . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.6.8. INVERT - Invert latitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.6.9. INVERTLEV - Invert levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.6.10. SHIFTXY - Shift field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
2.6.11. MASKREGION - Mask regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2.6.12. MASKBOX - Mask a box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
2.6.13. SETBOX - Set a box to constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
2.6.14. ENLARGE - Enlarge fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
2.6.15. SETMISS - Set missing value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
2.6.16. VERTFILLMISS - Vertical filling of missing values . . . . . . . . . . . . . . . . . . . 91
2.6.17. TIMFILLMISS - Temporal filling of missing values . . . . . . . . . . . . . . . . . . . 91
2.6.18. SETGRIDCELL - Set the value of a grid cell . . . . . . . . . . . . . . . . . . . . . . 92
2.7. Arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
2.7.1. EXPR - Evaluate expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
2.7.2. MATH - Mathematical functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
2.7.3. ARITHC - Arithmetic with a constant . . . . . . . . . . . . . . . . . . . . . . . . . . 101
2.7.4. ARITH - Arithmetic on two datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
2.7.5. DAYARITH - Daily arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
2.7.6. MONARITH - Monthly arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.7.7. YEARARITH - Yearly arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

3
Contents Contents

2.7.8. YHOURARITH - Multi-year hourly arithmetic . . . . . . . . . . . . . . . . . . . . . 106

2.7.9. YDAYARITH - Multi-year daily arithmetic . . . . . . . . . . . . . . . . . . . . . . . 107
2.7.10. YMONARITH - Multi-year monthly arithmetic . . . . . . . . . . . . . . . . . . . . . 108
2.7.11. YSEASARITH - Multi-year seasonal arithmetic . . . . . . . . . . . . . . . . . . . . . 109
2.7.12. ARITHDAYS - Arithmetic with days . . . . . . . . . . . . . . . . . . . . . . . . . . 110
2.7.13. ARITHLAT - Arithmetic with latitude . . . . . . . . . . . . . . . . . . . . . . . . . . 110
2.8. Statistical values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.8.1. TIMCUMSUM - Cumulative sum over all timesteps . . . . . . . . . . . . . . . . . . 119
2.8.2. CONSECSTAT - Consecute timestep periods . . . . . . . . . . . . . . . . . . . . . . 119
2.8.3. VARSSTAT - Statistical values over all variables . . . . . . . . . . . . . . . . . . . . 120
2.8.4. ENSSTAT - Statistical values over an ensemble . . . . . . . . . . . . . . . . . . . . . 121
2.8.5. ENSSTAT2 - Statistical values over an ensemble . . . . . . . . . . . . . . . . . . . . 123
2.8.6. ENSVAL - Ensemble validation tools . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
2.8.7. FLDSTAT - Statistical values over a field . . . . . . . . . . . . . . . . . . . . . . . . 126
2.8.8. ZONSTAT - Zonal statistical values . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
2.8.9. MERSTAT - Meridional statistical values . . . . . . . . . . . . . . . . . . . . . . . . 130
2.8.10. GRIDBOXSTAT - Statistical values over grid boxes . . . . . . . . . . . . . . . . . . 132
2.8.11. REMAPSTAT - Remaps source points to target cells . . . . . . . . . . . . . . . . . . 133
2.8.12. VERTSTAT - Vertical statistical values . . . . . . . . . . . . . . . . . . . . . . . . . 135
2.8.13. TIMSELSTAT - Time range statistical values . . . . . . . . . . . . . . . . . . . . . . 136
2.8.14. TIMSELPCTL - Time range percentile values . . . . . . . . . . . . . . . . . . . . . . 137
2.8.15. RUNSTAT - Running statistical values . . . . . . . . . . . . . . . . . . . . . . . . . . 138
2.8.16. RUNPCTL - Running percentile values . . . . . . . . . . . . . . . . . . . . . . . . . 139
2.8.17. TIMSTAT - Statistical values over all timesteps . . . . . . . . . . . . . . . . . . . . . 140
2.8.18. TIMPCTL - Percentile values over all timesteps . . . . . . . . . . . . . . . . . . . . 141
2.8.19. HOURSTAT - Hourly statistical values . . . . . . . . . . . . . . . . . . . . . . . . . 142
2.8.20. HOURPCTL - Hourly percentile values . . . . . . . . . . . . . . . . . . . . . . . . . 143
2.8.21. DAYSTAT - Daily statistical values . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
2.8.22. DAYPCTL - Daily percentile values . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
2.8.23. MONSTAT - Monthly statistical values . . . . . . . . . . . . . . . . . . . . . . . . . 146
2.8.24. MONPCTL - Monthly percentile values . . . . . . . . . . . . . . . . . . . . . . . . . 147
2.8.25. YEARMONSTAT - Yearly mean from monthly data . . . . . . . . . . . . . . . . . . 148
2.8.26. YEARSTAT - Yearly statistical values . . . . . . . . . . . . . . . . . . . . . . . . . . 149
2.8.27. YEARPCTL - Yearly percentile values . . . . . . . . . . . . . . . . . . . . . . . . . . 150
2.8.28. SEASSTAT - Seasonal statistical values . . . . . . . . . . . . . . . . . . . . . . . . . 151
2.8.29. SEASPCTL - Seasonal percentile values . . . . . . . . . . . . . . . . . . . . . . . . . 152
2.8.30. YHOURSTAT - Multi-year hourly statistical values . . . . . . . . . . . . . . . . . . 153
2.8.31. DHOURSTAT - Multi-day hourly statistical values . . . . . . . . . . . . . . . . . . . 155
2.8.32. YDAYSTAT - Multi-year daily statistical values . . . . . . . . . . . . . . . . . . . . 157
2.8.33. YDAYPCTL - Multi-year daily percentile values . . . . . . . . . . . . . . . . . . . . 159
2.8.34. YMONSTAT - Multi-year monthly statistical values . . . . . . . . . . . . . . . . . . 160
2.8.35. YMONPCTL - Multi-year monthly percentile values . . . . . . . . . . . . . . . . . . 162
2.8.36. YSEASSTAT - Multi-year seasonal statistical values . . . . . . . . . . . . . . . . . . 163
2.8.37. YSEASPCTL - Multi-year seasonal percentile values . . . . . . . . . . . . . . . . . . 165
2.8.38. YDRUNSTAT - Multi-year daily running statistical values . . . . . . . . . . . . . . . 166
2.8.39. YDRUNPCTL - Multi-year daily running percentile values . . . . . . . . . . . . . . 168
2.9. Correlation and co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
2.9.1. FLDCOR - Correlation in grid space . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
2.9.2. TIMCOR - Correlation over time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
2.9.3. FLDCOVAR - Covariance in grid space . . . . . . . . . . . . . . . . . . . . . . . . . 171
2.9.4. TIMCOVAR - Covariance over time . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
2.10. Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
2.10.1. REGRES - Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
2.10.2. DETREND - Detrend time series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
2.10.3. TREND - Trend of time series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
2.10.4. TRENDARITH - Add or subtract a trend . . . . . . . . . . . . . . . . . . . . . . . . 175
2.11. EOFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
2.11.1. EOFS - Empirical Orthogonal Functions . . . . . . . . . . . . . . . . . . . . . . . . . 177

4
Contents Contents

2.11.2. EOFCOEFF - Principal coefficients of EOFs . . . . . . . . . . . . . . . . . . . . . . 179

2.12. Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
2.12.1. REMAPBIL - Bilinear interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
2.12.2. REMAPBIC - Bicubic interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
2.12.3. REMAPNN - Nearest neighbor remapping . . . . . . . . . . . . . . . . . . . . . . . . 183
2.12.4. REMAPDIS - Distance weighted average remapping . . . . . . . . . . . . . . . . . . 184
2.12.5. REMAPCON - First order conservative remapping . . . . . . . . . . . . . . . . . . . 185
2.12.6. REMAPCON2 - Second order conservative remapping . . . . . . . . . . . . . . . . . 187
2.12.7. REMAPLAF - Largest area fraction remapping . . . . . . . . . . . . . . . . . . . . . 189
2.12.8. REMAP - Grid remapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
2.12.9. REMAPETA - Remap vertical hybrid level . . . . . . . . . . . . . . . . . . . . . . . 191
2.12.10.VERTINTML - Vertical interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . 193
2.12.11.VERTINTAP - Vertical pressure interpolation . . . . . . . . . . . . . . . . . . . . . 194
2.12.12.VERTINTGH - Vertical height interpolation . . . . . . . . . . . . . . . . . . . . . . 195
2.12.13.INTLEVEL - Linear level interpolation . . . . . . . . . . . . . . . . . . . . . . . . . 196
2.12.14.INTLEVEL3D - Linear level interpolation from/to 3D vertical coordinates . . . . . 196
2.12.15.INTTIME - Time interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
2.12.16.INTYEAR - Year interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
2.13. Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
2.13.1. SPECTRAL - Spectral transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 200
2.13.2. SPECCONV - Spectral conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
2.13.3. WIND2 - D and V to velocity potential and stream function . . . . . . . . . . . . . 202
2.13.4. WIND - Wind transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
2.13.5. FOURIER - Fourier transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
2.14. Import/Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
2.14.1. IMPORTBINARY - Import binary data sets . . . . . . . . . . . . . . . . . . . . . . 207
2.14.2. IMPORTCMSAF - Import CM-SAF HDF5 files . . . . . . . . . . . . . . . . . . . . 208
2.14.3. IMPORTAMSR - Import AMSR binary files . . . . . . . . . . . . . . . . . . . . . . 209
2.14.4. INPUT - Formatted input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
2.14.5. OUTPUT - Formatted output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2.14.6. OUTPUTTAB - Table output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.14.7. OUTPUTGMT - GMT output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
2.15. Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
2.15.1. GRADSDES - GrADS data descriptor file . . . . . . . . . . . . . . . . . . . . . . . . 217
2.15.2. AFTERBURNER - ECHAM standard post processor . . . . . . . . . . . . . . . . . 218
2.15.3. FILTER - Time series filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
2.15.4. GRIDCELL - Grid cell quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
2.15.5. SMOOTH - Smooth grid points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
2.15.6. DELTAT - Difference between timesteps . . . . . . . . . . . . . . . . . . . . . . . . . 222
2.15.7. REPLACEVALUES - Replace variable values . . . . . . . . . . . . . . . . . . . . . . 223
2.15.8. GETGRIDCELL - Get grid cell index . . . . . . . . . . . . . . . . . . . . . . . . . . 223
2.15.9. VARGEN - Generate a field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
2.15.10.TIMSORT - Timsort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
2.15.11.WINDTRANS - Wind Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 226
2.15.12.ROTUVB - Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2.15.13.MROTUVB - Backward rotation of MPIOM data . . . . . . . . . . . . . . . . . . . 227
2.15.14.MASTRFU - Mass stream function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2.15.15.DERIVEPAR - Derived model parameters . . . . . . . . . . . . . . . . . . . . . . . . 228
2.15.16.ADISIT - Potential temperature to in-situ temperature and vice versa . . . . . . . . 229
2.15.17.RHOPOT - Calculates potential density . . . . . . . . . . . . . . . . . . . . . . . . . 229
2.15.18.HISTOGRAM - Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
2.15.19.SETHALO - Set the bounds of a field . . . . . . . . . . . . . . . . . . . . . . . . . . 230
2.15.20.WCT - Windchill temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
2.15.21.FDNS - Frost days where no snow index per time period . . . . . . . . . . . . . . . . 231
2.15.22.STRWIN - Strong wind days index per time period . . . . . . . . . . . . . . . . . . . 231
2.15.23.STRBRE - Strong breeze days index per time period . . . . . . . . . . . . . . . . . . 232
2.15.24.STRGAL - Strong gale days index per time period . . . . . . . . . . . . . . . . . . . 232
2.15.25.HURR - Hurricane days index per time period . . . . . . . . . . . . . . . . . . . . . 232

5
Contents Contents

2.15.26.CMORLITE - CMOR lite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

2.15.27.VERIFYGRID - Verify grid coordinates . . . . . . . . . . . . . . . . . . . . . . . . . 234
2.15.28.HEALPIX - Change healpix resolution . . . . . . . . . . . . . . . . . . . . . . . . . . 235

3. Contributors 236
3.1. History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
3.2. External sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
3.3. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

A. Environment Variables 240

B. Parallelized operators 241

C. Standard name table 243

D. Grid description examples 244

D.1. Example of a curvilinear grid description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
D.2. Example description for an unstructured grid . . . . . . . . . . . . . . . . . . . . . . . . . . 245

Index 246

6
1. Introduction
The Climate Data Operator (CDO) software is a collection of many operators for standard processing of
climate and forecast model data. The operators include simple statistical and arithmetic functions, data
selection and subsampling tools, and spatial interpolation. CDO was developed to have the same set of
processing functions for GRIB [GRIB] and NetCDF [NetCDF] datasets in one package.
The Climate Data Interface [CDI] is used for the fast and file format independent access to GRIB and
NetCDF datasets. The local MPI-MET data formats SERVICE, EXTRA and IEG are also supported.
There are some limitations for GRIB and NetCDF datasets:
GRIB datasets have to be consistent, similar to NetCDF. That means all time steps need to have the same
variables, and within a time step each variable may occur only once. Multiple fields in single GRIB2
messages are not supported!
NetCDF datasets are only supported for the classic data model and arrays up to 4 dimensions. These
dimensions should only be used by the horizontal and vertical grid and the time. The NetCDF
attributes should follow the GDT, COARDS or CF Conventions.
The main CDO features are:
• More than 700 operators available
• Modular design and easily extendable with new operators
• Very simple UNIX command line interface
• A dataset can be processed by several operators, without storing the interim results in files
• Most operators handle datasets with missing values
• Fast processing of large datasets
• Support of many different grid types
• Tested on many UNIX/Linux systems, Cygwin, and MacOS-X
Latest pdf documentation be found here.

1.1. Installation
CDO is supported in different operative systems such as Unix, macOS and Windows. This section describes
how to install CDO in those platforms. More examples are found on the main website (https://code.
mpimet.mpg.de/projects/cdo/wiki)

1.1.1. Unix
1.1.1.1. Prebuilt CDO packages

Prebuilt CDO versions are available in online Unix repositories, and you can install them by typing on the
Unix terminal

apt-get install cdo

Note that prebuilt libraries do not offer the most recent version, and their version might vary with the
Unix system (see table below). It is recommended to build from the source or Conda environment for an
updated version or a customised setting.

7
Installation Introduction

Unix OS CDO Version

11 (Bullseye) 1.9.10-1
Debian 10 (Buster) 1.9.6-1
Sid 2.0.6-2
13 2.0.6
FreeBSD
12 2.0.6
Leap 15.3 2.0.6
openSUSE
Tumbleweed 2.0.6-1
18.04 LTS 1.9.3
Ubuntu 20.04 LTS 1.9.9
22.04 LTS 2.0.4-1

1.1.1.2. Building from sources

CDO uses the GNU configure and build system for compilation. The only requirement is a working ISO
C++17 and C11 compiler.
First go to the download page (https://code.mpimet.mpg.de/projects/cdo) to get the latest distribu-
tion, if you do not have it yet.
To take full advantage of CDO features the following additional libraries should be installed:

• Unidata NetCDF library (https://www.unidata.ucar.edu/software/netcdf) version 3 or higher.

This library is needed to process NetCDF [NetCDF] files with CDO.
• ECMWF ecCodes library (https://software.ecmwf.int/wiki/display/ECC/ecCodes+Home) ver-
sion 2.3.0 or higher. This library is needed to process GRIB2 files with CDO.
• HDF5 szip library (https://www.hdfgroup.org/doc_resource/SZIP) version 2.1 or higher.
This library is needed to process szip compressed GRIB [GRIB] files with CDO.
• HDF5 library (https://www.hdfgroup.org) version 1.6 or higher.
This library is needed to import CM-SAF [CM-SAF] HDF5 files with the CDO operator im-
port_cmsaf.
• PROJ library (https://proj.org) version 5.0 or higher.
This library is needed to convert Sinusoidal and Lambert Azimuthal Equal Area coordinates to
geographic coordinates, for e.g. remapping.
• Magics library (https://software.ecmwf.int/wiki/display/MAGP/Magics) version 2.18 or higher.
This library is needed to create contour, vector and graph plots with CDO.

CDO is a multi-threaded application. Therefore all the above libraries should be compiled thread safe.
Using non-threadsafe libraries could cause unexpected errors!

Compilation

Compilation is done by performing the following steps:

1. Unpack the archive, if you haven’t done that yet:

gunzip cdo-$VERSION.tar.gz # uncompress the archive

tar xf cdo-$VERSION.tar # unpack it
cd cdo-$VERSION

2. Run the configure script:

./configure

• Optionaly with NetCDF [NetCDF] support:

8
Introduction Installation

./configure --with-netcdf=<NetCDF root directory>

• and with ecCodes:

./configure --with-eccodes=<ecCodes root directory>

For an overview of other configuration options use

./configure --help

3. Compile the program by running make:

make

The program should compile without problems and the binary (cdo) should be available in the src
directory of the distribution.

Installation

After the compilation of the source code do a make install, possibly as root if the destination
permissions require that.

make install

The binary is installed into the directory <prefix>/bin. <prefix> defaults to /usr/local but
can be changed with the --prefix option of the configure script.
Alternatively, you can also copy the binary from the src directory manually to some bin directory
in your search path.

1.1.1.3. Conda

Conda is an open-source package manager and environment management system for various languages
(Python, R, etc.). Conda is installed via Anaconda or Miniconda. Unlike Anaconda, miniconda is a
lightweight conda distribution. They can be dowloaded from the main conda Website (https://conda.
io/projects/conda/en/latest/user-guide/install/linux.html) or on the terminal

wget https://repo.anaconda.com/archive/Anaconda3-2021.11-Linux-x86_64.sh
bash Anaconda3-2021.11-Linux-x86_64.sh
source ~/.bashrc

and

wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh

Upon setting your conda environment, you can install CDO using conda

conda install cdo

conda install python-cdo

1.1.2. MacOS

Among the MacOS package managers, CDO can be installed from Homebrew and Macports. The instal-
lation via Homebrew is straight forward process on the terminal

brew install cdo

9
Installation Introduction

Similarly, Macports

port install cdo

In contrast to Homebrew, Macport allows you to enable GRIB2, szip compression and Magics++ graphic
in CDO installation.

port install cdo +grib_api +magicspp +szip

In addition, you could also set CDO via Conda as Unix. You can follow this tutorial to install anaconda
or miniconda in your computer (https://conda.io/projects/conda/en/latest/user-guide/install/
macos.html). Then, you can install cdo by

conda install -c conda-forge cdo

1.1.3. Windows

Currently, CDO is not supported in Windows system and the binary is not available in the windows conda
repository. Therefore, CDO needs to be set in a virtual environment. Here, it covers the installation of
CDO using Windows Subsystem Linux (WSL) and virtual machines.

1.1.3.1. WSL

WSL emulates Unix in your Windows system. Then, you can install Unix libraries and software such
as CDO or the linux conda distribution in your computer. Also, it allows you to directly share your
files between your Windows and the WSL environment. However, more complex functions that require a
graphic interface are not allowed.
In Windows 10 or newer, WSL can be readily set in your cmd by typing

wsl --install

This command will install, by default, Ubuntu 20.04 in WSL2. You could also choose a different system
from this list.

wsl -l -o

Then, you can install your WSL environment as

wsl --install -d NAME

1.1.3.2. Virtual machine

Virtual machines can emulate different operative systems in your computer. Virtual machines are guest
computers mounted inside your host computer. You can set a Linux distribution in your Windows device
in this particular case. The advantages of Virtual machines to WSL are the graphical interface and the
fully operational Linux system. You can follow any tutorial on the internet such as this one

https://ubuntu.com/tutorials/how-to-run-ubuntu-desktop-on-a-virtual-machine-using-virtualbox#
1-overview

Finally, you can install CDO following any method listed in the section 1.1.1.

10
Introduction Usage

1.2. Usage
This section descibes how to use CDO. The syntax is:

cdo [ Options ] Operator1 [ -Operator2 [ -OperatorN ] ]

1.2.1. Options

All options have to be placed before the first operator. The following options are available for all operators:

-a Generate an absolute time axis.

-b <nbits> Set the number of bits for the output precision. The valid precisions depend
on the file format:
<format> <nbits>
grb1, grb2 P1 - P24
nc1, nc2, nc4, nc4c, nc5 I8/I16/I32/F32/F64
nc4, nc4c, nc5 U8/U16/U32
grb2, srv, ext, ieg F32/F64
For srv, ext and ieg format the letter L or B can be added to set the byteorder
to Little or Big endian.
--cmor CMOR conform NetCDF output.
-C, --color Colorized output messages.
--double Using double precision floats for data in memory.
--eccodes Use ecCodes to decode/encode GRIB1 messages.
--filter <filterId,params>
NetCDF4/HDF5 filter description.
-f <format> Set the output file format. The valid file formats are:
File format <format>
GRIB version 1 grb1/grb
GRIB version 2 grb2
NetCDF nc1
NetCDF version 2 (64-bit offset) nc2/nc
NetCDF-4 (HDF5) nc4
NetCDF-4 classic nc4c
NetCDF version 5 (64-bit data) nc5
SERVICE srv
EXTRA ext
IEG ieg
GRIB2 is only available if CDO was compiled with ecCodes support and all
NetCDF file types are only available if CDO was compiled with NetCDF support!
-g <grid> Define the default grid description by name or from file (see chapter 1.3 on page 18).
Available grid names are: r<NX>x<NY>, lon=<LON>/lat=<LAT>, F<XXX>, gme<NI>
-h, --help Help information for the operators.
--no_history Do not append to NetCDF history global attribute.
--netcdf_hdr_pad, --hdr_pad, --header_pad <nbr>
Pad NetCDF output header with nbr bytes.
-k <chunktype> NetCDF4 chunk type: auto, grid or lines.
-L Lock I/O (sequential access).
-m <missval> Set the missing value of non NetCDF files (default: -9e+33).
-O Overwrite existing output file, if checked.
Existing output file is checked only for: ens<STAT>, merge, mergetime
--operators List of all operators.
-P <nthreads> Set number of OpenMP threads (Only available if OpenMP support was compiled in).
--pedantic Warnings count as errors.

11
Usage Introduction

--percentile <method>
Methods: nrank, nist, rtype8, <NumPy method (linear|lower|higher|nearest|...)>
--reduce_dim Reduce NetCDF dimensions.
-R, --regular Convert GRIB1 data from global reduced to regular Gaussian grid (only with cgribex lib).
-r Generate a relative time axis.
-S Create an extra output stream for the module TIMSTAT. This stream contains
the number of non missing values for each output period.
-s, --silent Silent mode.
--shuffle Specify shuffling of variable data bytes before compression (NetCDF).
--single Using single precision floats for data in memory.
--sortname Alphanumeric sorting of NetCDF parameter names.
-t <partab> Set the GRIB1 (cgribex) default parameter table name or file (see chapter 1.6 on page 24).
Predefined tables are: echam4 echam5 echam6 mpiom1 ecmwf remo
--timestat_date <srcdate>
Target timestamp (temporal statistics): first, middle, midhigh or last source timestep.
-V, --version Print the version number.
-v, --verbose Print extra details for some operators.
-w Disable warning messages.
--worker <num> Number of worker to decode/decompress GRIB records.
-z aec AEC compression of GRIB1 records.
jpeg JPEG compression of GRIB2 records.
zip[_1-9] Deflate compression of NetCDF4 variables.
zstd[_1-19] Zstandard compression of NetCDF4 variables.

1.2.2. Environment variables

There are some environment variables which influence the behavior of CDO. An incomplete list can be
found in Appendix A.
Here is an example to set the envrionment variable CDO_RESET_HISTORY for different shells:

Bourne shell (sh): CDO_RESET_HISTORY=1 ; export CDO_RESET_HISTORY

Korn shell (ksh): export CDO_RESET_HISTORY=1
C shell (csh): setenv CDO_RESET_HISTORY 1

1.2.3. Operators

There are more than 700 operators available. A detailed description of all operators can be found in the
Reference Manual section.

1.2.4. Parallelized operators

Some of the CDO operators are shared memory parallelized with OpenMP. An OpenMP-enabled C compiler
is needed to use this feature. Users may request a specific number of OpenMP threads nthreads with the
’ -P’ switch.
Here is an example to distribute the bilinear interpolation on 8 OpenMP threads:

cdo -P 8 remapbil,targetgrid infile outfile

Many CDO operators are I/O-bound. This means most of the time is spend in reading and writing the
data. Only compute intensive CDO operators are parallelized. An incomplete list of OpenMP parallelized
operators can be found in Appendix B.

12
Introduction Usage

1.2.5. Operator parameter

Some operators need one or more parameter. A list of parameter is indicated by the seperator ’,’.

• STRING
String parameters require quotes if the string contains blanks or other characters interpreted by the
shell. The following command select variables with the name pressure and tsurf:
cdo selvar,pressure,tsurf infile outfile
• FLOAT
Floating point number in any representation. The following command sets the range between 0 and
273.15 of all fields to missing value:
cdo setrtomiss,0,273.15 infile outfile
• BOOL
Boolean parameter in the following representation TRUE/FALSE, T/F or 0/1. To disable the weight-
ing by grid cell area in the calculation of a field mean, use:
cdo fldmean,weights=FALSE infile outfile
• INTEGER
A range of integer parameter can be specified by first/last[/inc]. To select the days 5, 6, 7, 8 and 9
use:
cdo selday,5/9 infile outfile
The result is the same as:
cdo selday,5,6,7,8,9 infile outfile

1.2.6. Operator chaining

Operator chaining allows to combine two or more operators on the command line into a single CDO call.
This allows the creation of complex operations out of more simple ones: reductions over several dimensions,
file merges and all kinds of analysis processes. All operators with a fixed number of input streams and
one output stream can pass the result directly to an other operator. For differentiation between files and
operators all operators must be written with a prepended "–" when chaining.

cdo -monmean -add -mulc,2.0 infile1 -daymean infile2 outfile (CDO example call)

Here monmean will have the output of add while add takes the output of mulc,2.0 and daymean. infile1
and infile2 are inputs for their predecessor. When mixing operators with an arbitrary number of input
streams extra care needs to be taken. The following examples illustrates why.
1. cdo info -timavg infile1 infile2
2. cdo info -timavg infile?
3. cdo timavg infile1 tmpfile
cdo info tmpfile infile2
rm tmpfile
All three examples produce identical results. The time average will be computed only on the first input file.

Note(1): In section 1.3.2 we introduce argument groups which will make this a lot easier and less er-
ror prone.

Note(2): Operator chaining is implemented over POSIX Threads (pthreads). Therefore this CDO feature
is not available on operating systems without POSIX Threads support!

13
Advanced Usage Introduction

1.2.7. Chaining Benefits

Combining operators can have several benefits. The most obvious is a performance increase through
reducing disk I/O:

cdo sub -dayavg infile2 -timavg infile1 outfile

instead of

cdo timavg infile1 tmp1

cdo dayavg infile2 tmp2
cdo sub tmp2 tmp1 outfile
rm tmp1 tmp2

Especially with large input files the reading and writing of intermediate files can have a big influence on
the overall performance.
A second aspect is the execution of operators: Limited by the algorythms potentially all operators of a
chain can run in parallel.

1.3. Advanced Usage

In this section we will introduce advanced features of CDO. These include operator grouping which allows
to write more complex CDO calls and the apply keyword which allows to shorten calls that need an operator
to be executed on multiple files as well as wildcards which allow to search paths for file signatures. These
features have several restrictions and follow rules that depend on the input/output properties. These
required properties of operators can be investigated with the following commands which will output a list
of operators that have selected properties:

cdo --attribs [arbitrary/filesOnly/onlyFirst/noOutput/obase]

• arbitrary describes all operators where the number of inputs is not defined.
• filesOnly are operators that can have other operators as input.
• onlyFirst shows which operators can only be at the most left position of the polish notation argument
chain.
• noOutput are all operators that do not print to any file (e.g info)
• obase Here obase describes an operator that does not use the output argument as file but e.g as a file
name base (output base). This is almost exclusivly used for operators the split input files.

cdo -splithour baseName_

could result in: baseName_1 baseName_2 ... baseName_N

For checking a single or multiple operator directly the following usage of --attribs can be used:

cdo --attribs operatorName

1.3.1. Wildcards

Wildcards are a standard feature of command line interpreters (shells) on many operating systems. They
are placeholder characters used in file paths that are expanded by the interpreter into file lists. For further
information the Advance Bash Scripting Guide is a valuable source of information. Handling of input is
a central issue for CDO and in some circumstances it is not enough to use the wildcards from the shell.
That’s why CDO can handle them on its own.

14
Introduction Advanced Usage

all files 2020-2-01.txt 2020-2-11.txt 2020-2-15.txt 2020-3-01.txt 2020-3-02.txt

2020-3-12.txt 2020-3-13.txt 2020-3-15.txt 2021.grb 2022.grb
wildcard filelist results
2020-3* and 2020-3-??.txt 2020-3-01.txt 2020-3-02.txt 2020-3-12.txt 2020-3-13.txt 2020-3-15.txt
2020-3-?1.txt 2020-3-01.txt
*.grb 2021.grb 2020.grb
Use single quotes if the input stream names matched to a single wildcard expression. In this case CDO
will do the pattern matching and the output can be combined with other operators. Here is an example
for this feature:

cdo timavg -select,name=temperature ’infile?’ outfile

In earlier versions of CDO this was necessary to have the right files parsed to the right operator. Newer
version support this with the argument grouping feature (see 1.3.2). We advice the use of the grouping
mechanism instead of the single quoted wildcards since this feature could be deprecated in future versions.

Note: Wildcard expansion is not available on operating systems without the glob() function!

1.3.2. Argument Groups

In section 1.2.6 we described that it is not possible to chain operators with an arbitrary number of inputs.
In this section we want to show how this can be achieved through the use of operator grouping with angled
brackets []. Using these brackets CDO can assigned the inputs to their corresponding operators during
the execution of the command line. The ability to write operator combination in a parenthis-free way is
partly given up in favor of allowing operators with arbitrary number of inputs. This allows a much more
compact way to handle large number of input files.
The following example shows an example which we will transform from a non-working solution to a working
one.

cdo -infon -div -fldmean -cat infileA -mulc,2.0 infileB -fldmax infileC

This example will throw the following error:

cdo (Warning): Did you forget to use ’[’ and/or ’]’ for multiple variable input operators?
cdo (Warning): use option --variableInput, for description

cdo (Abort): Too few streams specified! Operator div needs 2 input streams and 1 output stream!

The error is raised by the operator div. This operator needs two input streams and one output stream,
but the cat operator has claimed all possible streams on its right hand side as input because it accepts an
arbitrary number of inputs. Hence it didn’t leave anything for the remaining input or output streams of
div. For this we can declare a group which will be passed to the operator left of the group.

cdo -infon -div -fldmean -cat [ infileA -mulc,2.0 infileB ] -fldmax infileC

For full flexibility it is possible to have groups inside groups:

cdo -infon -div -fldmean -cat [ infileA infileB -merge [ infileC1 infileC2 ] ] -fldmax infileD

15
Advanced Usage Introduction

1.3.3. Apply Keyword

When working with medium or large number of similar files there is a common problem of a processing
step (often a reduction) which needs to be performed on all of them before a more specific analysis can be
applied. Ususally this can be done in two ways: One option is to use merge to glue everything together and
chain the reduction step after it. The second option is to write a for-loop over all inputs which perform
the basic processing on each of the files separately and call merge one the results. Unfortunately both
options have side-effects: The first one needs a lot of memory because all files are read in completely and
reduced afterwards while the latter one creates a lot of temporary files. Both memory and disk IO can be
bottlenecks and should be avoided.
The apply keyword was introduced for that purpose. It can be used as an operator, but it needs at least
one operator as a parameter, which is applied in parallel to all related input streams in a parallel way
before all streams are passed to operator next in the chain.
The following is an example with three input files:

cdo -merge -apply,-daymean [ infile1 infile2 infile3 ] outfile

would result in:

cdo -merge -daymean infile1 -daymean infile2 -daymean infile3 outfile

Figure 1.1.: Usage and result of apply keyword

Apply is especially useful when combined with wildcards. The previous example can be shortened further.

cdo -merge -apply,-daymean [ infile? ] outfile

As shown this feature allows to simplify commands with medium amount of files and to move reductions
further back. This can also have a positive impact on the performance.

An example where performance can take a hit.

cdo -yearmean -daymean -merge [ f1 ... f40 ]

An improved but ugly to write example.

cdo -yearmean -merge [ -daymean f1 -daymean f2 ... -daymean f40 ]

Apply saves the day. And creates the call above with much less typing.
cdo -yearmean -merge [ -apply,-daymean [ f1 ... f40 ] ]

Figure 1.2.: Apply keyword simplifies command and execution

In the example in figure 1.2 the resulting call will dramatically save process interaction as well as execution
times since the reduction (daymean) is applied on the files first. That means that the merge operator will
receive the reduced files and the operations for merging the whole data is saved. For other CDO calls
further improvements can be made by adding more arguments to apply (1.3)

16
Introduction Memory Requirements

A less performant example.

cdo -aReduction -anotherReduction -daymean -merge [ f1 ... f40 ]

cdo -merge -apply,"-aReduction -anotherReduction -daymean" [ f1 ... f40 ]

Figure 1.3.: Multi argument apply

Restrictions: While the apply keyword can be extremely helpful it has several restrictions (for now!).
• Apply inputs can only be files, wildcards and operators that have 0 inputs and 1 output.
• Apply can not be used as the first CDO operator.
• Apply arguments can only be operators with 1 input and 1 output.
• Grouping inside the Apply argument or input is not allowed.

1.4. Memory Requirements

This section roughly describes the memory requirements of CDO. CDO tries to use as little memory as
possible. The smallest unit that is read by all operators is a horizontal field. The required memory depends
mainly on the used operators, the data format, the data type and the size of the fields.
The operators have partly very different memory requirements. Many CDO modules like FLDSTAT process
one horizontal field at a time. Memory-intensive modules such as ENSSTAT and TIMSTAT require all
fields of a time step to be held in memory. Of course, the memory requirements of each operator add
up when they are combined. Some operators are parallelized with OpenMP. In multi-threaded mode (see
option -P) the memory requirement can increase for these operators. This increase grows with the number
of threads used.
The data type determines the number of bytes per value. Single precision floating point data occupies
4 bytes per value. All other data types are read as double precision floats and thus occupy 8 bytes per
value. With the CDO option --single all data is read as single precision floats. This can reduce the memory
requirement by a factor of 2.

1.5. Horizontal grids

Physical quantities of climate models are typically stored on a horizonal grid. CDO supports structured
grids like regular lon/lat or curvilinear grids and also unstructured grids.

1.5.1. Grid area weights

One single point of a horizontal grid represents the mean of a grid cell. These grid cells are typically of
different sizes, because the grid points are of varying distance.
Area weights are individual weights for each grid cell. They are needed to compute the area weighted
mean or variance of a set of grid cells (e.g. fldmean - the mean value of all grid cells). In CDO the area
weights are derived from the grid cell area. If the cell area is not available then it will be computed from
the geographical coordinates via spherical triangles. This is only possible if the geographical coordinates of
the grid cell corners are available or derivable. Otherwise CDO gives a warning message and uses constant
area weights for all grid cells.
The cell area is read automatically from a NetCDF input file if a variable has the corresponding “cell_measures”
attribute, e.g.:

17
Horizontal grids Introduction

var : c e l l _ m e a s u r e s = " a r e a : c e l l _ a r e a " ;

If the computed cell area is not desired then the CDO operator setgridarea can be used to set or overwrite
the grid cell area.

1.5.2. Grid description

In the following situations it is necessary to give a description of a horizontal grid:

• Changing the grid description (operator: setgrid)
• Horizontal interpolation (all remapping operators)
• Generating of variables (operator: const, random)
As now described, there are several possibilities to define a horizontal grid.

1.5.2.1. Predefined grids

Predefined grids are available for global regular, gaussian, HEALPix or icosahedral-hexagonal GME grids.

Global regular grid: global_<DXY>

global_<DXY> defines a global regular lon/lat grid. The grid increment <DXY> can be chosen arbitrarily.
The longitudes start at <DXY>/2 - 180◦ and the latitudes start at <DXY>/2 - 90◦ .

Regional regular grid: dcw:<CountryCode>[_<DXY>]

dcw:<CountryCode>[_<DXY>] defines a regional regular lon/lat grid from the country code. The default
value of the optional grid increment <DXY> is 0.1 degree. The ISO two-letter country codes can be found
on https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2. To define a state, append the state code to the
country code, e.g. USAK for Alaska. For the coordinates of a country CDO uses the DCW (Digital Chart
of the World) dataset from GMT. This dataset must be installed on the system and the environment
variable DIR_DCW must point to it.

Zonal latitudes: zonal_<DY>

zonal_<DY> defines a grid with zonal latitudes only. The latitude increment <DY> can be chosen arbitrarily.
The latitudes start at <DY>/2 - 90◦ . The boundaries of each latitude are also generated. The number of
longitudes is 1. A grid description of this type is needed to calculate the zonal mean (zonmean) for data
on an unstructured grid.

Global regular grid: r<NX>x<NY>

r<NX>x<NY> defines a global regular lon/lat grid. The number of the longitudes <NX> and the latitudes
<NY> can be chosen arbitrarily. The longitudes start at 0◦ with an increment of (360/<NX>)◦ . The latitudes
go from south to north with an increment of (180/<NY>)◦ .

One grid point: lon=<LON>/lat=<LAT>

lon=<LON>/lat=<LAT> defines a lon/lat grid with only one grid point.

18
Introduction Horizontal grids

Full regular Gaussian grid: F<XXX>

F<XXX> defines a global regular Gaussian grid. XXX specifies the number of latitudes lines between the Pole
and the Equator. The longitudes start at 0◦ with an increment of (360/nlon)◦ . The gaussian latitudes go
from north to south.

Global icosahedral-hexagonal GME grid: gme<NI>

gme<NI> defines a global icosahedral-hexagonal GME grid. NI specifies the number of intervals on a main
triangle side.

HEALPix grid: hp<NSIDE>[_<ORDER>]

HEALPix is an acronym for Hierarchical Equal Area isoLatitude Pixelization of a sphere.

hp<NSIDE>[_<ORDER>] defines the parameter of a global HEALPix grid. The NSIDE parameter controls the
resolution of the pixellization. It is the number of pixels on the side of each of the 12 top-level HEALPix
pixels. The total number of grid pixels is 12*NSIDE*NSIDE. NSIDE=1 generates the 12 (H=4, K=3) equal
sized top-level HEALPix pixels. ORDER sets the index ordering convention of the pixels, available are nested
(default) or ring ordering. A shortcut for hp<NSIDE>_nested is hpz<ZOOM>. ZOOM is the zoom level and
the relation to NSIDE is zoom = log2 (nside).
If the geographical coordinates are required in CDO, they are calculated from the HEALPix parameters.
For this calculation the astropy-healpix C library is used.

1.5.2.2. Grids from data files

You can use the grid description from an other datafile. The format of the datafile and the grid of the data
field must be supported by CDO. Use the operator ’sinfo’ to get short informations about your variables
and the grids. If there are more then one grid in the datafile the grid description of the first variable will
be used. Add the extension :N to the name of the datafile to select grid number N.

1.5.2.3. SCRIP grids

SCRIP (Spherical Coordinate Remapping and Interpolation Package) uses a common grid description for
curvilinear and unstructured grids. For more information about the convention see [SCRIP]. This grid
description is stored in NetCDF. Therefor it is only available if CDO was compiled with NetCDF support!

SCRIP grid description example of a curvilinear MPIOM [MPIOM] GROB3 grid (only the NetCDF header):
netcdf grob3s {
dimensions :
g r i d _ s i z e = 12120 ;
grid_corners = 4 ;
grid_rank = 2 ;
variables :
i n t grid_dims ( g r i d _ r a n k ) ;
double grid_center_lat ( g r i d _ s i z e ) ;
grid_center_lat : units = " degrees " ;
g r i d _ c e n t e r _ l a t : bounds = " g r i d _ c o r n e r _ l a t " ;
double grid_center_lon ( g r i d _ s i z e ) ;
grid_center_lon : units = " degrees " ;
g r i d _ c e n t e r _ l o n : bounds = " g r i d _ c o r n e r _ l o n " ;
i n t grid_imask ( g r i d _ s i z e ) ;
grid_imask : u n i t s = " u n i t l e s s " ;
grid_imask : c o o r d i n a t e s = " g r i d _ c e n t e r _ l o n g r i d _ c e n t e r _ l a t " ;
double grid_corner_lat ( grid_size , grid_corners ) ;
grid_corner_lat : units = " degrees " ;
double grid_corner_lon ( grid_size , grid_corners ) ;
grid_corner_lon : units = " degrees " ;

19
Horizontal grids Introduction

// g l o b a l a t t r i b u t e s :
: t i t l e = " grob3s " ;
}

1.5.2.4. CDO grids

All supported grids can also be described with the CDO grid description. The following keywords can be
used to describe a grid:

Keyword Datatype Description

gridtype STRING Type of the grid (gaussian, lonlat, curvilinear, unstructured).
gridsize INTEGER Size of the grid.
xsize INTEGER Size in x direction (number of longitudes).
ysize INTEGER Size in y direction (number of latitudes).
xvals FLOAT ARRAY X values of the grid cell center.
yvals FLOAT ARRAY Y values of the grid cell center.
nvertex INTEGER Number of the vertices for all grid cells.
xbounds FLOAT ARRAY X bounds of each gridbox.
ybounds FLOAT ARRAY Y bounds of each gridbox.
xfirst, xinc FLOAT, FLOAT Macros to define xvals with a constant increment,
xfirst is the x value of the first grid cell center.
yfirst, yinc FLOAT, FLOAT Macros to define yvals with a constant increment,
yfirst is the y value of the first grid cell center.
xunits STRING units of the x axis
yunits STRING units of the y axis

Which keywords are necessary depends on the gridtype. The following table gives an overview of the
default values or the size with respect to the different grid types.

gridtype lonlat gaussian projection curvilinear unstructured

gridsize xsize*ysize xsize*ysize xsize*ysize xsize*ysize ncell
xsize nlon nlon nx nlon gridsize
ysize nlat nlat ny nlat gridsize
xvals xsize xsize xsize gridsize gridsize
yvals ysize ysize ysize gridsize gridsize
nvertex 2 2 2 4 nv
xbounds 2*xsize 2*xsize 2*xsize 4*gridsize nv*gridsize
ybounds 2*ysize 2*ysize 2*xsize 4*gridsize nv*gridsize
xunits degrees degrees m degrees degrees
yunits degrees degrees m degrees degrees

The keywords nvertex, xbounds and ybounds are optional if area weights are not needed. The grid cell
corners xbounds and ybounds have to rotate counterclockwise.

CDO grid description example of a T21 gaussian grid:

gridtype = gaussian
xsize = 64
ysize = 32
xfirst = 0
xinc = 5.625
yvals = 85.76 80.27 74.75 69.21 63.68 58.14 52.61 47.07
41.53 36.00 30.46 24.92 19.38 13.84 8.31 2.77
−2.77 −8.31 −13.84 −19.38 −24.92 −30.46 −36.00 −41.53
−47.07 −52.61 −58.14 −63.68 −69.21 −74.75 −80.27 −85.76

20
Introduction Horizontal grids

CDO grid description example of a global regular grid with 60x30 points:
gridtype = lonlat
xsize = 60
ysize = 30
xfirst = −177
xinc = 6
yfirst = −87
yinc = 6

The description for a projection is somewhat more complicated. Use the first section to describe the
coordinates of the projection with the above keywords. Add the keyword grid_mapping_name to
descibe the mapping between the given coordinates and the true latitude and longitude coordinates.
grid_mapping_name takes a string value that contains the name of the projection. A list of attributes
can be added to define the mapping. The name of the attributes depend on the projection. The valid
names of the projection and there attributes follow the NetCDF CF-Convention.
CDO supports the special grid mapping attribute proj_params. These parameter will be passed directly
to the PROJ library to generate the geographic coordinates if needed.
The geographic coordinates of the following projections can be generated without the attribute proj_params,
if all other attributes are available:
• rotated_latitude_longitude
• lambert_conformal_conic
• lambert_azimuthal_equal_area
• sinusoidal
• polar_stereographic
It is recommend to set the attribute proj_params also for the above projections to make sure all PROJ
parameter are set correctly.

Here is an example of a CDO grid description using the attribute proj_params to define the PROJ
parameter of a polar stereographic projection:
gridtype = projection
xsize = 11
ysize = 11
xunits = " meter "
yunits = " meter "
xfirst = −638000
xinc = 150
yfirst = −3349350
yinc = 150
grid_mapping = c r s
grid_mapping_name = p o l a r _ s t e r e o g r a p h i c
proj_params = "+ p r o j=s t e r e +lon_0=−45 +l a t _ t s =70 +l a t _ 0 =90 +x_0=0 +y_0=0"

The result is the same as using the CF conform Grid Mapping Attributes:
gridtype = projection
xsize = 11
ysize = 11
xunits = " meter "
yunits = " meter "
xfirst = −638000
xinc = 150
yfirst = −3349350
yinc = 150
grid_mapping = c r s
grid_mapping_name = p o l a r _ s t e r e o g r a p h i c
s t r a i g h t _ v e r t i c a l _ l o n g i t u d e _ f r o m _ p o l e = −45.
standard_parallel = 70.
latitude_of_projection_origin = 90.
false_easting = 0.
false_northing = 0.

21
Z-axis description Introduction

CDO grid description example of a regional rotated lon/lat grid:

gridtype = projection
xsize = 81
ysize = 91
xunits = " degrees "
yunits = " degrees "
xfirst = −19.5
xinc = 0.5
yfirst = −25.0
yinc = 0.5
grid_mapping_name = r o t a t e d _ l a t i t u d e _ l o n g i t u d e
g r i d _ n o r t h _ p o l e _ l o n g i t u d e = −170
grid_north_pole_latitude = 32.5

Example CDO descriptions of a curvilinear and an unstructured grid can be found in Appendix D.

1.5.3. ICON - Grid File Server

The geographic coordinates of the ICON model are located on an unstructured grid. This grid is stored in
a separate grid file independent of the model data. The grid files are made available to the general public
via a file server. Furthermore, these grid files are located at DKRZ under /pool/data/ICON/grids.
With the CDO function setgrid,<gridfile> this grid information can be added to the data if needed.
Here is an example:
cdo sellonlatbox,-20,60,10,70 -setgrid,<path_to_gridfile> icondatafile result

ICON model data in NetCDF format contains the global attribute grid_file_uri. This attribute contains
a link to the appropriate grid file on the ICON grid file server. If the global attribute grid_file_uri is
present and valid, the grid information can be added automatically. The setgrid function is then no longer
required. The environment variable CDO_DOWNLOAD_PATH can be used to select a directory for storing the
grid file. If this environment variable is set, the grid file will be automatically downloaded from the grid file
server to this directory if needed. If the grid file already exists in the current directory, the environment
variable does not need to be set.
If the grid files are available locally, like at DKRZ, they do not need to be fetched from the grid file server.
Use the environment variable CDO_ICON_GRIDS to set the root directory of the ICON grids. Here is an
example for the ICON grids at DKRZ:
CDO_ICON_GRIDS=/pool/data/ICON

1.6. Z-axis description

Sometimes it is necessary to change the description of a z-axis. This can be done with the operator setzaxis.
This operator needs an ASCII formatted file with the description of the z-axis. The following keywords
can be used to describe a z-axis:

Keyword Datatype Description

zaxistype STRING type of the z-axis
size INTEGER number of levels
levels FLOAT ARRAY values of the levels
lbounds FLOAT ARRAY lower level bounds
ubounds FLOAT ARRAY upper level bounds
vctsize INTEGER number of vertical coordinate parameters
vct FLOAT ARRAY vertical coordinate table

The keywords lbounds and ubounds are optional. vctsize and vct are only necessary to define hybrid
model levels.

22
Introduction Time axis

Available z-axis types:

Z-axis type Description Units

surface Surface
pressure Pressure level pascal
hybrid Hybrid model level
height Height above ground meter
depth_below_sea Depth below sea level meter
depth_below_land Depth below land surface centimeter
isentropic Isentropic (theta) level kelvin

Z-axis description example for pressure levels 100, 200, 500, 850 and 1000 hPa:
zaxistype = pressure
size = 5
levels = 10000 20000 50000 85000 100000

Z-axis description example for ECHAM5 L19 hybrid model levels:

zaxistype = hybrid
size = 19
levels = 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
vctsize = 40
vct = 0 2000 4000 6 0 4 6 . 1 0 9 3 8 8 2 6 7 . 9 2 5 7 8 1 0 6 0 9 . 5 1 1 7 1 2 8 5 1 . 1 0 1 6 1 4 6 9 8 . 5
15861.125 16116.2383 15356.9258 13621.4609 11101.5625 8127.14453
5125.14062 2549.96875 783.195068 0 0 0
0 0 0 0.000338993268 0.00335718691 0.0130700432 0.0340771675
0.0706498027 0.12591666 0.201195419 0.295519829 0.405408859
0.524931908 0.646107674 0.759697914 0.856437683 0.928747177
0.972985268 0.992281914 1

Note that the vctsize is twice the number of levels plus two and the vertical coordinate table must be
specified for the level interfaces.

1.7. Time axis

A time axis describes the time for every timestep. Two time axis types are available: absolute time and
relative time axis. CDO tries to maintain the actual type of the time axis for all operators.

1.7.1. Absolute time

An absolute time axis has the current time to each time step. It can be used without knowledge of the
calendar. This is preferably used by climate models. In NetCDF files the absolute time axis is represented
by the unit of the time: "day as %Y%m%d.%f".

1.7.2. Relative time

A relative time is the time relative to a fixed reference time. The current time results from the reference time
and the elapsed interval. The result depends on the calendar used. CDO supports the standard Gregorian,
proleptic Gregorian, 360 days, 365 days and 366 days calendars. The relative time axis is preferably used
by numerical weather prediction models. In NetCDF files the relative time axis is represented by the unit
of the time: "time-units since reference-time", e.g "days since 1989-6-15 12:00".

23
Parameter table Introduction

1.7.3. Conversion of the time

Some programs which work with NetCDF data can only process relative time axes. Therefore it may be
necessary to convert from an absolute into a relative time axis. This conversion can be done for each
operator with the CDO option ’-r’. To convert a relative into an absolute time axis use the CDO option
’-a’.

1.8. Parameter table

A parameter table is an ASCII formated file to convert code numbers to variable names. Each variable
has one line with its code number, name and a description with optional units in a blank separated list. It
can only be used for GRIB, SERVICE, EXTRA and IEG formated files. The CDO option ’-t <partab>’
sets the default parameter table for all input files. Use the operator ’setpartab’ to set the parameter table
for a specific file.

Example of a CDO parameter table:

134 aps s u r f a c e p r e s s u r e [ Pa ]
141 sn snow depth [m]
147 ahfl l a t e n t h e a t f l u x [W/m∗ ∗ 2 ]
172 slm l a n d s e a mask
175 albedo s u r f a c e albedo
211 siced i c e depth [m]

1.9. Missing values

Missing values are data points that are missing or invalid. Such data points are treated in a different way
than valid data. Most CDO operators can handle missing values in a smart way. But if the missing value
is within the range of valid data, it can lead to incorrect results. This applies to all arithmetic operations,
but especially to logical operations when the missing value is 0 or 1.

The default missing value for GRIB, SERVICE, EXTRA and IEG files is −9.e33 . The CDO option ’-m
<missval>’ overwrites the default missing value. In NetCDF files the variable attribute ’_FillValue’ is
used as a missing value. The operator ’setmissval’ can be used to set a new missing value.

The CDO use of the missing value is shown in the following tables, where one table is printed for each
operation. The operations are applied to arbitrary numbers a, b, the special case 0, and the missing value
miss. For example the table named "addition" shows that the sum of an arbitrary number a and the
missing value is the missing value, and the table named "multiplication" shows that 0 multiplied by missing
value results in 0.

24
Introduction Percentile

addition b miss
a a+b miss
miss miss miss
subtraction b miss
a a−b miss
miss miss miss
multiplication b 0 miss
a a∗b 0 miss
0 0 0 0
miss miss 0 miss
division b 0 miss
a a/b miss miss
0 0 miss miss
miss miss miss miss
maximum b miss
a max(a, b) a
miss b miss
minimum b miss
a min(a, b) a
miss b miss
sum b miss
a a+b a
miss b miss

The handling of missing values by the operations "minimum" and "maximum" may be surprising, but the
definition given here is more consistent with that expected in practice. Mathematical functions (e.g. log,
sqrt, etc.) return the missing value if an argument is the missing value or an argument is out of range.

All statistical functions ignore missing values, treading them as not belonging to the sample, with the
side-effect of a reduced sample size.

1.9.1. Mean and average

An artificial distinction is made between the notions mean and average. The mean is regarded as a
statistical function, whereas the average is found simply by adding the sample members and dividing the
result by the sample size. For example, the mean of 1, 2, miss and 3 is (1 + 2 + 3)/3 = 2, whereas the
average is (1 + 2 + miss + 3)/4 = miss/4 = miss. If there are no missing values in the sample, the average
and mean are identical.

1.10. Percentile

There is no standard definition of percentile. All definitions yield to similar results when the number of
values is very large. The following percentile methods are available in CDO:

25
Regions Introduction

Percentile
Description
method
nrank Nearest Rank method [default in CDO]
nist The primary method recommended by NIST
rtype8 R’s type=8 method
inverted_cdf NumPy with percentile method=’inverted_cdf’ (R type=1)
averaged_inverted_cdf NumPy with percentile method=’averaged_inverted_cdf’ (R type=2)
closest_observation NumPy with percentile method=’closest_observation’ (R type=3)
interpolated_inverted_cdf NumPy with percentile method=’interpolated_inverted_cdf’ (R type=4)
hazen NumPy with percentile method=’hazen’ (R type=5)
weibull NumPy with percentile method=’weibull’ (R type=6)
linear NumPy with percentile method=’linear’ (R type=7) [default in NumPy and R]
median_unbiased NumPy with percentile method=’median_unbiased’ (R type=8)
normal_unbiased NumPy with percentile method=’normal_unbiased’ (R type=9)
lower NumPy with percentile method=’lower’
higher NumPy with percentile method=’higher’
midpoint NumPy with percentile method=’midpoint’
nearest NumPy with percentile method=’nearest’

The percentile method can be selected with the CDO option - -percentile. The Nearest Rank method
is the default percentile method in CDO.

The different percentile methods can lead to different results, especially for small number of data values.
Consider the ordered list {15, 20, 35, 40, 50, 55}, which contains six data values. Here is the result for the
30th, 40th, 50th, 75th and 100th percentiles of this list using the different percentile methods:

Percentile NumPy NumPy NumPy NumPy

nrank nist rtype8
P linear lower higher nearest
30th 20 21.5 23.5 27.5 20 35 35
40th 35 32 33 35 35 35 35
50th 35 37.5 37.5 37.5 35 40 40
75th 50 51.25 50.42 47.5 40 50 50
100th 55 55 55 55 55 55 55

1.10.1. Percentile over timesteps

The amount of data for time series can be very large. All data values need to held in memory to calculate
the percentile. The percentile over timesteps uses a histogram algorithm, to limit the amount of required
memory. The default number of histogram bins is 101. That means the histogram algorithm is used,
when the dataset has more than 101 time steps. The default can be overridden by setting the environment
variable CDO_PCTL_NBINS to a different value. The histogram algorithm is implemented only for the Nearest
Rank method.

1.11. Regions

The CDO operators maskregion and selregion can be used to mask and select regions. For this purpose,
the region needs to be defined by the user. In CDO there are two possibilities to define regions.

One possibility is to define the regions with an ASCII file. Each region is defined by a convex polygon.
Each line of the polygon contains the longitude and latitude coordinates of a point. A description file for
regions can contain several polygons, these must be separated by a line with the character &.

26
Introduction Regions

Here is a simple example of a polygon for a box with longitudes from 120W to 90E and latitudes from 20N
to 20S:
120 20
120 −20
270 −20
270 20

With the second option, predefined regions can be used via country codes. A country is specified with
dcw:<CountryCode>. Country codes can be combined with the plus sign.

Here is an example to select the region Spain and Portugal:

cdo selregion,dcw:ES+PT infile outfile

The ISO two-letter country codes can be found on https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2.

To define a state, append the state code to the country code, e.g. USAK for Alaska. For the coordinates
of a country CDO uses the DCW (Digital Chart of the World) dataset from GMT. This dataset must be
installed on the system and the environment variable DIR_DCW must point to it.

27
2. Reference manual
This section gives a description of all operators. Related operators are grouped to modules. For easier
description all single input files are named infile or infile1, infile2, etc., and an arbitrary number of
input files are named infiles. All output files are named outfile or outfile1, outfile2, etc. Further
the following notion is introduced:
i(t) Timestep t of infile
i(t, x) Element number x of the field at timestep t of infile
o(t) Timestep t of outfile
o(t, x) Element number x of the field at timestep t of outfile

28
Reference manual Information

2.1. Information
This section contains modules to print information about datasets. All operators print there results to
standard output.
Here is a short overview of all operators in this section:

info Dataset information listed by parameter identifier

infon Dataset information listed by parameter name
map Dataset information and simple map

sinfo Short information listed by parameter identifier

sinfon Short information listed by parameter name

xsinfo Extra short information listed by parameter name

xsinfop Extra short information listed by parameter identifier

diff Compare two datasets listed by parameter id

diffn Compare two datasets listed by parameter name

npar Number of parameters

nlevel Number of levels
nyear Number of years
nmon Number of months
ndate Number of dates
ntime Number of timesteps
ngridpoints Number of gridpoints
ngrids Number of horizontal grids

showformat Show file format

showcode Show code numbers
showname Show variable names
showstdname Show standard names
showlevel Show levels
showltype Show GRIB level types
showyear Show years
showmon Show months
showdate Show date information
showtime Show time information
showtimestamp Show timestamp

showattribute Show a global attribute or a variable attribute

partab Parameter table

codetab Parameter code table
griddes Grid description
zaxisdes Z-axis description
vct Vertical coordinate table

29
Information Reference manual

2.1.1. INFO - Information and simple statistics

Synopsis

< operator > infiles

Description
This module writes information about the structure and contents for each field of all input files to
standard output. A field is a horizontal layer of a data variable. All input files need to have the same
structure with the same variables on different timesteps. The information displayed depends on the
chosen operator.

Operators

info Dataset information listed by parameter identifier

Prints information and simple statistics for each field of all input datasets. For each field
the operator prints one line with the following elements:
• Date and Time
• Level, Gridsize and number of Missing values
• Minimum, Mean and Maximum
The mean value is computed without the use of area weights!
• Parameter identifier
infon Dataset information listed by parameter name
The same as operator info but using the name instead of the identifier to label the param-
eter.
map Dataset information and simple map
Prints information, simple statistics and a map for each field of all input datasets. The
map will be printed only for fields on a regular lon/lat grid.

Example
To print information and simple statistics for each field of a dataset use:
cdo infon infile

This is an example result of a dataset with one 2D parameter over 12 timesteps:

−1 : Date Time L e v e l Size Miss : Minimum Mean Maximum : Name
1 : 1987−01−31 12:00:00 0 2048 1361 : 232.77 2 6 6 . 6 5 3 0 5 . 3 1 : SST
2 : 1987−02−28 12:00:00 0 2048 1361 : 233.64 2 6 7 . 1 1 3 0 7 . 1 5 : SST
3 : 1987−03−31 12:00:00 0 2048 1361 : 225.31 2 6 7 . 5 2 3 0 7 . 6 7 : SST
4 : 1987−04−30 12:00:00 0 2048 1361 : 215.68 2 6 8 . 6 5 3 1 0 . 4 7 : SST
5 : 1987−05−31 12:00:00 0 2048 1361 : 215.78 2 7 1 . 5 3 3 1 2 . 4 9 : SST
6 : 1987−06−30 12:00:00 0 2048 1361 : 212.89 2 7 2 . 8 0 3 1 4 . 1 8 : SST
7 : 1987−07−31 12:00:00 0 2048 1361 : 209.52 2 7 4 . 2 9 3 1 6 . 3 4 : SST
8 : 1987−08−31 12:00:00 0 2048 1361 : 210.48 2 7 4 . 4 1 3 1 5 . 8 3 : SST
9 : 1987−09−30 12:00:00 0 2048 1361 : 210.48 2 7 2 . 3 7 3 1 2 . 8 6 : SST
10 : 1987−10−31 12:00:00 0 2048 1361 : 219.46 2 7 0 . 5 3 3 0 9 . 5 1 : SST
11 : 1987−11−30 12:00:00 0 2048 1361 : 230.98 2 6 9 . 8 5 3 0 8 . 6 1 : SST
12 : 1987−12−31 12:00:00 0 2048 1361 : 241.25 2 6 9 . 9 4 3 0 9 . 2 7 : SST

30
Reference manual Information

2.1.2. SINFO - Short information

Synopsis
< operator > infiles

Description
This module writes information about the structure of infiles to standard output. infiles is an
arbitrary number of input files. All input files need to have the same structure with the same variables
on different timesteps. The information displayed depends on the chosen operator.

Operators
sinfo Short information listed by parameter identifier
Prints short information of a dataset. The information is divided into 4 sections. Section
1 prints one line per parameter with the following information:
• institute and source
• time c=constant v=varying
• type of statistical processing
• number of levels and z-axis number
• horizontal grid size and number
• data type
• parameter identifier
Section 2 and 3 gives a short overview of all grid and vertical coordinates. And the last
section contains short information of the time coordinate.
sinfon Short information listed by parameter name
The same as operator sinfo but using the name instead of the identifier to label the
parameter.

Example
To print short information of a dataset use:
cdo sinfon infile

This is the result of an ECHAM5 dataset with 3 parameter over 12 timesteps:

−1 : I n s t i t u t S o u r c e T S t e p t y p e L e v e l s Num P o i n t s Num Dtype : Name
1 : MPIMET ECHAM5 c i n s t a n t 1 1 2048 1 F32 : GEOSP
2 : MPIMET ECHAM5 v i n s t a n t 4 2 2048 1 F32 : T
3 : MPIMET ECHAM5 v i n s t a n t 1 1 2048 1 F32 : TSURF
Grid coordinates :
1 : gaussian : p o i n t s =2048 ( 6 4 x32 ) F16
l o n g i t u d e : 0 t o 3 5 4 . 3 7 5 by 5 . 6 2 5 d e g r e e s _ e a s t circular
l a t i t u d e : 8 5 . 7 6 0 6 t o −85.7606 d e g r e e s _ n o r t h
Vertical coordinates :
1 : surface : l e v e l s =1
2 : pressure : l e v e l s =4
l e v e l : 92500 t o 20000 Pa
Time c o o r d i n a t e :
time : 12 s t e p s
YYYY−MM−DD hh :mm: s s YYYY−MM−DD hh :mm: s s YYYY−MM−DD hh :mm: s s YYYY−MM−DD hh :mm: s s
1987−01−31 1 2 : 0 0 : 0 0 1987−02−28 1 2 : 0 0 : 0 0 1987−03−31 1 2 : 0 0 : 0 0 1987−04−30 1 2 : 0 0 : 0 0
1987−05−31 1 2 : 0 0 : 0 0 1987−06−30 1 2 : 0 0 : 0 0 1987−07−31 1 2 : 0 0 : 0 0 1987−08−31 1 2 : 0 0 : 0 0
1987−09−30 1 2 : 0 0 : 0 0 1987−10−31 1 2 : 0 0 : 0 0 1987−11−30 1 2 : 0 0 : 0 0 1987−12−31 1 2 : 0 0 : 0 0

31
Information Reference manual

2.1.3. XSINFO - Extra short information

Synopsis
< operator > infiles

Description
This module writes information about the structure of infiles to standard output. infiles is an
arbitrary number of input files. All input files need to have the same structure with the same variables
on different timesteps. The information displayed depends on the chosen operator.

Operators
xsinfo Extra short information listed by parameter name
Prints short information of a dataset. The information is divided into 4 sections. Section
1 prints one line per parameter with the following information:
• institute and source
• time c=constant v=varying
• type of statistical processing
• number of levels and z-axis number
• horizontal grid size and number
• data type
• memory type (float or double)
• parameter name
Section 2 to 4 gives a short overview of all grid, vertical and time coordinates.
xsinfop Extra short information listed by parameter identifier
The same as operator xsinfo but using the identifier instead of the name to label the
parameter.

Example
To print extra short information of a dataset use:
cdo xsinfo infile

This is the result of an ECHAM5 dataset with 3 parameter over 12 timesteps:

−1 : I n s t i t u t Source T S t e p t y p e L e v e l s Num P o i n t s Num Dtype Mtype : Name
1 : MPIMET ECHAM5 c instant 1 1 2048 1 F32 F32 : GEOSP
2 : MPIMET ECHAM5 v instant 4 2 2048 1 F32 F32 : T
3 : MPIMET ECHAM5 v instant 1 1 2048 1 F32 F32 : TSURF
Grid coordinates :
1 : gaussian : p o i n t s =2048 ( 6 4 x32 ) F16
longitude : 0 t o 3 5 4 . 3 7 5 by 5 . 6 2 5 d e g r e e s _ e a s t circular
latitude : 8 5 . 7 6 0 6 t o −85.7606 d e g r e e s _ n o r t h
Vertical coordinates :
1 : surface : l e v e l s =1
2 : pressure : l e v e l s =4
level : 92500 t o 20000 Pa
Time c o o r d i n a t e :
steps : 12
time : 1987−01−31T18 : 0 0 : 0 0 t o 1987−12−31T18 : 0 0 : 0 0 by 1 month
units : days s i n c e 1987−01−01T00 : 0 0 : 0 0
calendar : proleptic_gregorian

32
Reference manual Information

2.1.4. DIFF - Compare two datasets field by field

Synopsis

< operator >[,options] infile1 infile2

Description
Compares the contents of two datasets field by field. The input datasets need to have the same
structure and its fields need to have the dimensions. Try the option names if the number of variables
differ. Exit status is 0 if inputs are the same and 1 if they differ.

Operators

diff Compare two datasets listed by parameter id

Provides statistics on differences between two datasets. For each pair of fields the operator
prints one line with the following information:
• Date and Time
• Level, Gridsize and number of Missing values
• Number of different values
• Occurrence of coefficient pairs with different signs (S)
• Occurrence of zero values (Z)
• Maxima of absolute difference of coefficient pairs
• Maxima of relative difference of non-zero coefficient pairs with equal signs
• Parameter identifier

Absdif f (t, x) = |i1 (t, x) − i2 (t, x)|

|i1 (t, x) − i2 (t, x)|

Reldif f (t, x) =
max(|i1 (t, x)| , |i2 (t, x)|)
diffn Compare two datasets listed by parameter name
The same as operator diff. Using the name instead of the identifier to label the parameter.

Parameter
maxcount INTEGER Stop after maxcount different fields
abslim FLOAT Limit of the maximum absolute difference (default: 0)
rellim FLOAT Limit of the maximum relative difference (default: 1)
names STRING Consideration of the variable names of only one input file (left/right) or
the intersection of both (intersect).

Example
To print the difference for each field of two datasets use:
cdo diffn infile1 infile2

This is an example result of two datasets with one 2D parameter over 12 timesteps:

33
Information Reference manual

Date Time L e v e l Size Miss D i f f : S Z Max_Absdiff Max_Reldiff : Name

1 : 1987−01−31 1 2 : 0 0 : 0 0 0 2048 1361 273 : F F 0 . 0 0 0 1 0 6 8 1 4 . 1 6 6 0 e −07 : SST
2 : 1987−02−28 1 2 : 0 0 : 0 0 0 2048 1361 309 : F F 6 . 1 0 3 5 e −05 2 . 3 7 4 2 e −07 : SST
3 : 1987−03−31 1 2 : 0 0 : 0 0 0 2048 1361 292 : F F 7 . 6 2 9 4 e −05 3 . 3 7 8 4 e −07 : SST
4 : 1987−04−30 1 2 : 0 0 : 0 0 0 2048 1361 183 : F F 7 . 6 2 9 4 e −05 3 . 5 1 1 7 e −07 : SST
5 : 1987−05−31 1 2 : 0 0 : 0 0 0 2048 1361 207 : F F 0 . 0 0 0 1 0 6 8 1 4 . 0 3 0 7 e −07 : SST
7 : 1987−07−31 1 2 : 0 0 : 0 0 0 2048 1361 317 : F F 9 . 1 5 5 3 e −05 3 . 5 6 3 4 e −07 : SST
8 : 1987−08−31 1 2 : 0 0 : 0 0 0 2048 1361 219 : F F 7 . 6 2 9 4 e −05 2 . 8 8 4 9 e −07 : SST
9 : 1987−09−30 1 2 : 0 0 : 0 0 0 2048 1361 188 : F F 7 . 6 2 9 4 e −05 3 . 6 1 6 8 e −07 : SST
10 : 1987−10−31 1 2 : 0 0 : 0 0 0 2048 1361 297 : F F 9 . 1 5 5 3 e −05 3 . 5 0 0 1 e −07 : SST
11 : 1987−11−30 1 2 : 0 0 : 0 0 0 2048 1361 234 : F F 6 . 1 0 3 5 e −05 2 . 3 8 3 9 e −07 : SST
12 : 1987−12−31 1 2 : 0 0 : 0 0 0 2048 1361 267 : F F 9 . 3 5 5 3 e −05 3 . 7 6 2 4 e −07 : SST
11 o f 12 r e c o r d s d i f f e r

2.1.5. NINFO - Print the number of parameters, levels or times

Synopsis

< operator > infile

Description
This module prints the number of variables, levels or times of the input dataset.

Operators

npar Number of parameters

Prints the number of parameters (variables).
nlevel Number of levels
Prints the number of levels for each variable.
nyear Number of years
Prints the number of different years.
nmon Number of months
Prints the number of different combinations of years and months.
ndate Number of dates
Prints the number of different dates.
ntime Number of timesteps
Prints the number of timesteps.
ngridpoints Number of gridpoints
Prints the number of gridpoints for each variable.
ngrids Number of horizontal grids
Prints the number of horizontal grids.

Example
To print the number of parameters (variables) in a dataset use:
cdo npar infile

To print the number of months in a dataset use:

cdo nmon infile

34
Reference manual Information

2.1.6. SHOWINFO - Show variables, levels or times

Synopsis

< operator > infile

Description
This module prints the format, variables, levels or times of the input dataset.

Operators

showformat Show file format

Prints the file format of the input dataset.
showcode Show code numbers
Prints the code number of all variables.
showname Show variable names
Prints the name of all variables.
showstdname Show standard names
Prints the standard name of all variables.
showlevel Show levels
Prints all levels for each variable.
showltype Show GRIB level types
Prints the GRIB level type for all z-axes.
showyear Show years
Prints all years.
showmon Show months
Prints all months.
showdate Show date information
Prints date information of all timesteps (format YYYY-MM-DD).
showtime Show time information
Prints time information of all timesteps (format hh:mm:ss).
showtimestamp Show timestamp
Prints timestamp of all timesteps (format YYYY-MM-DDThh:mm:ss).

Example
To print the code number of all variables in a dataset use:
cdo showcode infile

This is an example result of a dataset with three variables:

129 130 139

To print all months in a dataset use:

cdo showmon infile

This is an examples result of a dataset with an annual cycle:

1 2 3 4 5 6 7 8 9 10 11 12

35
Information Reference manual

2.1.7. SHOWATTRIBUTE - Show attributes

Synopsis

showattribute[,attributes] infile

Description
This operator prints the attributes of the data variables of a dataset.
Each attribute has the following structure:
[var_nm@][att_nm]

var_nm Variable name (optional). Example: pressure

att_nm Attribute name (optional). Example: units

The value of var_nm is the name of the variable containing the attribute (named att_nm) that
you want to print. Use wildcards to print the attribute att_nm of more than one variable. A value
of var_nm of ’*’ will print the attribute att_nm of all data variables. If var_nm is missing then
att_nm refers to a global attribute.
The value of att_nm is the name of the attribute you want to print. Use wildcards to print more
than one attribute. A value of att_nm of ’*’ will print all attributes.

Parameter
attributes STRING Comma-separated list of attributes.

36
Reference manual Information

2.1.8. FILEDES - Dataset description

Synopsis

< operator > infile

Description
This module provides operators to print meta information about a dataset. The printed meta-data
depends on the chosen operator.

Operators

partab Parameter table

Prints all available meta information of the variables.
codetab Parameter code table
Prints a code table with a description of all variables. For each variable the operator
prints one line listing the code, name, description and units.
griddes Grid description
Prints the description of all grids.
zaxisdes Z-axis description
Prints the description of all z-axes.
vct Vertical coordinate table
Prints the vertical coordinate table.

Example
Assume all variables of the dataset are on a Gausssian N16 grid. To print the grid description of this
dataset use:
cdo griddes infile

Result:
gridtype : gaussian
gridsize : 2048
xname : lon
xlongname : longitude
xunits : degrees_east
yname : lat
ylongname : latitude
yunits : degrees_north
xsize : 64
ysize : 32
xfirst : 0
xinc : 5.625
yvals : 85.76058 80.26877 74.74454 69.21297 63.67863 58.1429 52.6065
47.06964 41.53246 35.99507 30.4575 24.91992 19.38223 13.84448
8 . 3 0 6 7 0 2 2 . 7 6 8 9 0 3 −2.768903 −8.306702 −13.84448 −19.38223
−24.91992 −30.4575 −35.99507 −41.53246 −47.06964 −52.6065
−58.1429 −63.67863 −69.21297 −74.74454 −80.26877 −85.76058

37
File operations Reference manual

2.2. File operations

This section contains modules to perform operations on files.
Here is a short overview of all operators in this section:

apply Apply operators on each input file.

copy Copy datasets

clone Clone datasets
cat Concatenate datasets

tee Duplicate a data stream

pack Pack data

unpack Unpack data

bitrounding Bit rounding

replace Replace variables

duplicate Duplicates a dataset

mergegrid Merge grid

merge Merge datasets with different fields

mergetime Merge datasets sorted by date and time

splitcode Split code numbers

splitparam Split parameter identifiers
splitname Split variable names
splitlevel Split levels
splitgrid Split grids
splitzaxis Split z-axes
splittabnum Split parameter table numbers

splithour Split hours

splitday Split days
splitseas Split seasons
splityear Split years
splityearmon Split in years and months
splitmon Split months

splitsel Split time selection

splitdate Splits a file into dates

distgrid Distribute horizontal grid

collgrid Collect horizontal grid

38
Reference manual File operations

2.2.1. APPLY - Apply operators

Synopsis

apply,operators infiles

Description
The apply utility runs the named operators on each input file. The input files must be enclosed in
square brackets. This utility can only be used on a series of input files. These are all operators with
more than one input file (infiles). Here is an incomplete list of these operators: copy, cat, merge,
mergetime, select, ENSSTAT. The parameter operators is a blank-separated list of CDO operators.
Use quotation marks if more than one operator is needed. Each operator may have only one input
and output stream.

Parameter
operators STRING Blank-separated list of CDO operators.

Example
Suppose we have multiple input files with multiple variables on different time steps. The input files
contain the variables U and V, among others. We are only interested in the absolute windspeed on
all time steps. Here is the standard CDO solution for this task:
cdo expr,wind="sqrt(u*u+v*v)" -mergetime infile1 infile2 infile3 outfile

This first joins all the time steps together and then calculates the wind speed. If there are many
variables in the input files, this procedure is ineffective. In this case it is better to first calculate the
wind speed:
cdo mergetime -expr,wind="sqrt(u*u+v*v)" infile1 \
-expr,wind="sqrt(u*u+v*v)" infile2 \
-expr,wind="sqrt(u*u+v*v)" infile3 outfile

However, this can quickly become very confusing with more than 3 input files. The apply operator
solves this problem:
cdo mergetime -apply,-expr,wind="sqrt(u*u+v*v)" [ infile1 infile2 infile3 ] outfile

Another example is the calculation of the mean value over several input files with ensmean. The input
files contain several variables, but we are only interested in the variable named XXX:
cdo ensmean -apply,-selname,XXX [ infile1 infile2 infile3 ] outfile

39
File operations Reference manual

2.2.2. COPY - Copy datasets

Synopsis

< operator > infiles outfile

Description
This module contains operators to copy, clone or concatenate datasets. infiles is an arbitrary
number of input files. All input files need to have the same structure with the same variables on
different timesteps.

Operators

copy Copy datasets

Copies all input datasets to outfile.
clone Clone datasets
Copies all input datasets to outfile. In contrast to the copy operator, clone tries not to
change the input data. GRIB records are neither decoded nor decompressed.
cat Concatenate datasets
Concatenates all input datasets and appends the result to the end of outfile. If outfile
does not exist it will be created.

Example
To change the format of a dataset to NetCDF use:
cdo -f nc copy infile outfile.nc

Add the option ’-r’ to create a relative time axis, as is required for proper recognition by GrADS or
Ferret:
cdo -r -f nc copy infile outfile.nc

To concatenate 3 datasets with different timesteps of the same variables use:

cdo copy infile1 infile2 infile3 outfile

If the output dataset already exists and you wish to extend it with more timesteps use:
cdo cat infile1 infile2 infile3 outfile

40
Reference manual File operations

2.2.3. TEE - Duplicate a data stream and write it to file

Synopsis

tee,outfile2 infile outfile1

Description
This operator copies the input dataset to outfile1 and outfile2. The first output stream in
outfile1 can be further processesd with other cdo operators. The second output outfile2 is written
to disk. It can be used to store intermediate results to a file.

Parameter
outfile2 STRING Destination filename for the copy of the input file

Example
To compute the daily and monthy average of a dataset use:
cdo monavg -tee,outfile_dayavg dayavg infile outfile_monavg

2.2.4. PACK - Pack data

Synopsis

pack infile outfile

Description
Packing reduces the data volume by reducing the precision of the stored numbers. It is implemented
using the NetCDF attributes add_offset and scale_factor. The operator pack calculates the
attributes add_offset and scale_factor for all variables. The default data type for all variables is
automatically changed to 16-bit integer. Use the CDO option -b to change the data type to a different
integer precision, if needed. Missing values are automatically transformed to the current data type.

2.2.5. UNPACK - Unpack data

Synopsis

unpack infile outfile

Description
Packing reduces the data volume by reducing the precision of the stored numbers. It is implemented
using the NetCDF attributes add_offset and scale_factor. The operator unpack unpack all packed
variables. The default data type for all variables is automatically changed to 32-bit floats. Use the
CDO option -b F64 to change the data type to 64-bit floats, if needed.

41
File operations Reference manual

2.2.6. BITROUNDING - Bit rounding

Synopsis

bitrounding[,parameter] infile outfile

Description
This operator calculates for each field the number of necessary mantissa bits to get a certain infor-
mation level in the data. With this number of significant bits (numbits) a rounding of the data is
performed. This allows the data to be compressed to a higher level.
The default value of the information level is 0.9999 and can be adjusted with the parameter inflevel.
That means 99.99% of the information in the mantissa bits is preserved.
Alternatively, the number of significant bits can be set for all variables with the numbits parameter.
Furthermore, numbits can be assigned for each variable via the filename parameter. In this case,
numbits is still calculated for all variables if they are not present in the file.
The analysis of the bit information is based on the Julia library BitInformation.jl. The procedure to
derive the number of significant mantissa bits was adapted from the Python library xbitinfo. Quantize
to the number of mantissa bits is done with IEEE rounding using code from NetCDF 4.9.0.
Currently only 32-bit float data is rounded. Data with missing values are not yet supported for the
calculation of significant bits.

Parameter
inflevel FLOAT Information level (0 - 1) [default: 0.9999]
addbits INTEGER Add bits to the number of significant bits [default: 0]
minbits INTEGER Minimum value of the number of bits [default: 1]
maxbits INTEGER Maximum value of the number of bits [default: 23]
numsteps INTEGER Set to 1 to run the calculation only in the first time step
numbits INTEGER Set number of significant bits
printbits BOOL Print max. numbits per variable of 1st timestep to stdout [format: name=numbits]
filename STRING Read number of significant bits per variable from file [format: name=numbits]

Example
Apply bit rounding to all 32-bit float fields, preserving 99.9% of the information, followed by com-
pression and storage to NetCDF4:
cdo -f nc4 -z zip bitrounding,inflevel=0.999 infile outfile

Add the option ’-v’ to view used number of mantissa bits for each field:
cdo -v -f nc4 -z zip bitrounding,inflevel=0.999 infile outfile

42
Reference manual File operations

2.2.7. REPLACE - Replace variables

Synopsis

replace infile1 infile2 outfile

Description
This operator replaces variables in infile1 by variables from infile2 and write the result to outfile.
Both input datasets need to have the same number of timesteps. All variable names may only occur
once!

Example
Assume the first input dataset infile1 has three variables with the names geosp, t and tslm1 and the
second input dataset infile2 has only the variable tslm1. To replace the variable tslm1 in infile1
by tslm1 from infile2 use:
cdo replace infile1 infile2 outfile

2.2.8. DUPLICATE - Duplicates a dataset

Synopsis

duplicate[,ndup] infile outfile

Description
This operator duplicates the contents of infile and writes the result to outfile. The optional
parameter sets the number of duplicates, the default is 2.

Parameter
ndup INTEGER Number of duplicates, default is 2.

2.2.9. MERGEGRID - Merge grid

Synopsis

mergegrid infile1 infile2 outfile

Description
Merges grid points of all variables from infile2 to infile1 and write the result to outfile. Only
the non missing values of infile2 will be used. The horizontal grid of infile2 should be smaller
or equal to the grid of infile1 and the resolution must be the same. Only rectilinear grids are
supported. Both input files need to have the same variables and the same number of timesteps.

43
File operations Reference manual

2.2.10. MERGE - Merge datasets

Synopsis

< operator > infiles outfile

Description
This module reads datasets from several input files, merges them and writes the resulting dataset to
outfile.

Operators

merge Merge datasets with different fields

Merges time series of different fields from several input datasets. The number of fields
per timestep written to outfile is the sum of the field numbers per timestep in all
input datasets. The time series on all input datasets are required to have different
fields and the same number of timesteps. The fields in each different input file either
have to be different variables or different levels of the same variable. A mixture of
different variables on different levels in different input files is not allowed.
mergetime Merge datasets sorted by date and time
Merges all timesteps of all input files sorted by date and time. All input files need
to have the same structure with the same variables on different timesteps. After this
operation every input timestep is in outfile and all timesteps are sorted by date
and time.

Environment
SKIP_SAME_TIME If set to 1, skips all consecutive timesteps with a double entry of the same
timestamp.

Note
Operators of this module need to open all input files simultaneously. The maximum number of open
files depends on the operating system!

Example
Assume three datasets with the same number of timesteps and different variables in each dataset. To
merge these datasets to a new dataset use:
cdo merge infile1 infile2 infile3 outfile

Assume you split a 6 hourly dataset with splithour. This produces four datasets, one for each hour.
The following command merges them together:
cdo mergetime infile1 infile2 infile3 infile4 outfile

44
Reference manual File operations

2.2.11. SPLIT - Split a dataset

Synopsis

< operator >[,parameter] infile obase

Description
This module splits infile into pieces. The output files will be named <obase><xxx><suffix>
where suffix is the filename extension derived from the file format. xxx and the contents of the output
files depends on the chosen operator. params is a comma-separated list of processing parameters.

Operators

splitcode Split code numbers

Splits a dataset into pieces, one for each different code number. xxx will have three
digits with the code number.
splitparam Split parameter identifiers
Splits a dataset into pieces, one for each different parameter identifier. xxx will be
a string with the parameter identifier.
splitname Split variable names
Splits a dataset into pieces, one for each variable name. xxx will be a string with
the variable name.
splitlevel Split levels
Splits a dataset into pieces, one for each different level. xxx will have six digits
with the level.
splitgrid Split grids
Splits a dataset into pieces, one for each different grid. xxx will have two digits
with the grid number.
splitzaxis Split z-axes
Splits a dataset into pieces, one for each different z-axis. xxx will have two digits
with the z-axis number.
splittabnum Split parameter table numbers
Splits a dataset into pieces, one for each GRIB1 parameter table number. xxx will
have three digits with the GRIB1 parameter table number.

Parameter
swap STRING Swap the position of obase and xxx in the output filename
uuid=<attname> STRING Add a UUID as global attribute <attname> to each output file

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

Note
Operators of this module need to open all output files simultaneously. The maximum number of open
files depends on the operating system!

45
File operations Reference manual

Example
Assume an input GRIB1 dataset with three variables, e.g. code number 129, 130 and 139. To split
this dataset into three pieces, one for each code number use:
cdo splitcode infile code

Result of ’dir code*’:

code129 . grb code130 . grb code139 . grb

46
Reference manual File operations

2.2.12. SPLITTIME - Split timesteps of a dataset

Synopsis

< operator > infile obase

splitmon[,format] infile obase

Description
This module splits infile into timesteps pieces. The output files will be named <obase><xxx><suffix>
where suffix is the filename extension derived from the file format. xxx and the contents of the out-
put files depends on the chosen operator.

Operators

splithour Split hours

Splits a file into pieces, one for each different hour. xxx will have two digits with
the hour.
splitday Split days
Splits a file into pieces, one for each different day. xxx will have two digits with
the day.
splitseas Split seasons
Splits a file into pieces, one for each different season. xxx will have three characters
with the season.
splityear Split years
Splits a file into pieces, one for each different year. xxx will have four digits with
the year (YYYY).
splityearmon Split in years and months
Splits a file into pieces, one for each different year and month. xxx will have six
digits with the year and month (YYYYMM).
splitmon Split months
Splits a file into pieces, one for each different month. xxx will have two digits with
the month.

Parameter
format STRING C-style format for strftime() (e.g. %B for the full month name)

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

Note
Operators of this module need to open all output files simultaneously. The maximum number of open
files depends on the operating system!

47
File operations Reference manual

Example
Assume the input GRIB1 dataset has timesteps from January to December. To split each month with
all variables into one separate file use:
cdo splitmon infile mon

Result of ’dir mon*’:

mon01 . grb mon02 . grb mon03 . grb mon04 . grb mon05 . grb mon06 . grb
mon07 . grb mon08 . grb mon09 . grb mon10 . grb mon11 . grb mon12 . grb

2.2.13. SPLITSEL - Split selected timesteps

Synopsis

splitsel,nsets[,noffset[,nskip]] infile obase

Description
This operator splits infile into pieces, one for each adjacent sequence t_1, ...., t_n of timesteps of
the same selected time range. The output files will be named <obase><nnnnnn><suffix> where
nnnnnn is the sequence number and suffix is the filename extension derived from the file format.

Parameter
nsets INTEGER Number of input timesteps for each output file
noffset INTEGER Number of input timesteps skipped before the first timestep range (optional)
nskip INTEGER Number of input timesteps skipped between timestep ranges (optional)

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

2.2.14. SPLITDATE - Splits a file into dates

Synopsis

splitdate infile obase

Description
This operator splits infile into pieces, one for each different date. The output files will be named
<obase><YYYY-MM-DD><suffix> where YYYY-MM-DD is the date and suffix is the filename exten-
sion derived from the file format.

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

48
Reference manual File operations

2.2.15. DISTGRID - Distribute horizontal grid

Synopsis

distgrid,nx[,ny] infile obase

Description
This operator distributes a dataset into smaller pieces. Each output file contains a different region
of the horizontal source grid. 2D Lon/Lat grids can be split into nx*ny pieces, where a target grid
region contains a structured longitude/latitude box of the source grid. Data on an unstructured grid
is split into nx pieces. The output files will be named <obase><xxx><suffix> where suffix is the
filename extension derived from the file format. xxx will have five digits with the number of the target
region.

Parameter
nx INTEGER Number of regions in x direction, or number of pieces for unstructured grids
ny INTEGER Number of regions in y direction [default: 1]

Note
This operator needs to open all output files simultaneously. The maximum number of open files
depends on the operating system!

Example
Distribute data on a 2D Lon/Lat grid into 6 smaller files, each output file receives one half of x and
a third of y of the source grid:
cdo distgrid,2,3 infile.nc obase

Below is a schematic illustration of this example:

−20 20 20 60
−20 20 60 45 45
45 45

15 15
15 15 15 15

−15 −15 −15 −15

−15 −15

−45 −45
−20 20 60 −45 −45
−20 20 20 60

On the left side is the data of the input file and on the right side is the data of the six output files.

49
File operations Reference manual

2.2.16. COLLGRID - Collect horizontal grid

Synopsis

collgrid[,nx[,names]] infiles outfile

Description
This operator collects the data of the input files to one output file. All input files need to have the
same variables and the same number of timesteps on a different horizonal grid region. If the source
regions are on a structured lon/lat grid, all regions together must result in a new structured lat/long
grid box. Data on an unstructured grid is concatenated in the order of the input files. The parameter
nx needs to be specified only for curvilinear grids.

Parameter
nx INTEGER Number of regions in x direction [default: number of input files]
names STRING Comma-separated list of variable names [default: all variables]

Note
This operator needs to open all input files simultaneously. The maximum number of open files depends
on the operating system!

Example
Collect the horizonal grid of 6 input files. Each input file contains a lon/lat region of the target grid:
cdo collgrid infile[1-6] outfile

Below is a schematic illustration of this example:

−20 20 20 60
45 45 −20 20 60
45 45

15 15
15 15 15 15

−15 −15 −15 −15

−15 −15

−45 −45
−45 −45 −20 20 60
−20 20 20 60

On the left side is the data of the six input files and on the right side is the collected data of the
output file.

50
Reference manual Selection

2.3. Selection
This section contains modules to select time steps, fields or a part of a field from a dataset.
Here is a short overview of all operators in this section:

select Select fields

delete Delete fields

selmulti Select multiple fields

delmulti Delete multiple fields
changemulti Change identication of multiple fields

selparam Select parameters by identifier

delparam Delete parameters by identifier
selcode Select parameters by code number
delcode Delete parameters by code number
selname Select parameters by name
delname Delete parameters by name
selstdname Select parameters by standard name
sellevel Select levels
sellevidx Select levels by index
selgrid Select grids
selzaxis Select z-axes
selzaxisname Select z-axes by name
selltype Select GRIB level types
seltabnum Select parameter table numbers

seltimestep Select timesteps

seltime Select times
selhour Select hours
selday Select days
selmonth Select months
selyear Select years
selseason Select seasons
seldate Select dates
selsmon Select single month

sellonlatbox Select a longitude/latitude box

selindexbox Select an index box

selregion Select cells inside regions

selcircle Select cells inside a circle

selgridcell Select grid cells

delgridcell Delete grid cells

samplegrid Resample grid

selyearidx Select year by index

bottomvalue Extract bottom level

topvalue Extract top level
isosurface Extract isosurface

51
Selection Reference manual

2.3.1. SELECT - Select fields

Synopsis

< operator >,parameter infiles outfile

Description
This module selects some fields from infiles and writes them to outfile. infiles is an arbitrary
number of input files. All input files need to have the same structure with the same variables on
different timesteps. The fields selected depends on the chosen parameters. Parameter is a comma-
separated list of "key=value" pairs. A range of integer values can be specified by first/last[/inc].
Wildcards are supported for string values.

Operators

select Select fields

Selects all fields with parameters in a user given list.
delete Delete fields
Deletes all fields with parameters in a user given list.

Parameter
name STRING Comma-separated list of variable names.
param STRING Comma-separated list of parameter identifiers.
code INTEGER Comma-separated list or first/last[/inc] range of code numbers.
level FLOAT Comma-separated list of vertical levels.
levrange FLOAT First and last value of the level range.
levidx INTEGER Comma-separated list or first/last[/inc] range of index of levels.
zaxisname STRING Comma-separated list of zaxis names.
zaxisnum INTEGER Comma-separated list or first/last[/inc] range of zaxis numbers.
ltype INTEGER Comma-separated list or first/last[/inc] range of GRIB level types.
gridname STRING Comma-separated list of grid names.
gridnum INTEGER Comma-separated list or first/last[/inc] range of grid numbers.
steptype STRING Comma-separated list of timestep types (constant, avg, accum,
min, max, range, diff, sum)
date STRING Comma-separated list of dates (format YYYY-MM-DDThh:mm:ss).
startdate STRING Start date (format YYYY-MM-DDThh:mm:ss).
enddate STRING End date (format YYYY-MM-DDThh:mm:ss).
minute INTEGER Comma-separated list or first/last[/inc] range of minutes.
hour INTEGER Comma-separated list or first/last[/inc] range of hours.
day INTEGER Comma-separated list or first/last[/inc] range of days.
month INTEGER Comma-separated list or first/last[/inc] range of months.
season STRING Comma-separated list of seasons (substring of DJFMAMJJA-
SOND or ANN).

52
Reference manual Selection

year INTEGER Comma-separated list or first/last[/inc] range of years.

dom STRING Comma-separated list of the day of month (e.g. 29feb).
timestep INTEGER Comma-separated list or first/last[/inc] range of timesteps. Neg-
ative values select timesteps from the end (NetCDF only).
timestep_of_year INTEGER Comma-separated list or first/last[/inc] range of timesteps of year.
timestepmask STRING Read timesteps from a mask file.

Example
Assume you have 3 inputfiles. Each inputfile contains the same variables for a different time period.
To select the variable T,U and V on the levels 200, 500 and 850 from all 3 input files, use:
cdo select,name=T,U,V,level=200,500,850 infile1 infile2 infile3 outfile

To remove the February 29th use:

cdo delete,dom=29feb infile outfile

53
Selection Reference manual

2.3.2. SELMULTI - Select multiple fields via GRIB1 parameters

Synopsis
< operator >,selection-specification infile outfile

Description
This module selects multiple fields from infile and writes them to outfile. selection-specification
is a filename or in-place string with the selection specification. Each selection-specification has the
following compact notation format:
<type>(parameters; leveltype(s); levels)

type sel for select or del for delete (optional)

parameters GRIB1 parameter code number
leveltype GRIB1 level type
levels value of each level
Examples:
( 1 ; 103; 0)
( 3 3 , 3 4 ; 105; 10)
( 1 1 , 1 7 ; 105; 2)
(71 ,73 ,74 ,75 ,61 ,62 ,65 ,117 ,67 ,122 ,121 ,11 ,131 ,66 ,84 ,111 ,112; 105; 0)

The following descriptive notation can also be used for selection specification from a file:
SELECT/DELETE, PARAMETER=parameters, LEVTYPE=leveltye(s), LEVEL=levels

Examples:
SELECT, PARAMETER=1 , LEVTYPE=103 , LEVEL=0
SELECT, PARAMETER=33/34 , LEVTYPE=105 , LEVEL=10
SELECT, PARAMETER=11/17 , LEVTYPE=105 , LEVEL=2
SELECT, PARAMETER=71/73/74/75/61/62/65/117/67/122 , LEVTYPE=105 , LEVEL=0
DELETE, PARAMETER=128 , LEVTYPE=109 , LEVEL=∗

The following will convert Pressure from Pa into hPa; Temp from Kelvin to Celsius:
SELECT, PARAMETER=1 , LEVTYPE= 1 0 3 , LEVEL=0, SCALE=0.01
SELECT, PARAMETER=11 , LEVTYPE=105 , LEVEL=2 , OFFSET=273.15

If SCALE and/or OFFSET are defined, then the data values are scaled as SCALE*(VALUE-OFFSET).

Operators
selmulti Select multiple fields

delmulti Delete multiple fields

changemulti Change identication of multiple fields

Example
Change ECMWF GRIB code of surface pressure to Hirlam notation:
cdo changemulti,’{(134;1;*|1;105;*)}’ infile outfile

54
Reference manual Selection

2.3.3. SELVAR - Select fields

Synopsis

< operator >,parameter infile outfile

selcode,codes infile outfile
delcode,codes infile outfile
selname,names infile outfile
delname,names infile outfile
selstdname,stdnames infile outfile
sellevel,levels infile outfile
sellevidx,levidx infile outfile
selgrid,grids infile outfile
selzaxis,zaxes infile outfile
selzaxisname,zaxisnames infile outfile
selltype,ltypes infile outfile
seltabnum,tabnums infile outfile

Description
This module selects some fields from infile and writes them to outfile. The fields selected de-
pends on the chosen operator and the parameters. A range of integer values can be specified by
first/last[/inc].

Operators

selparam Select parameters by identifier

Selects all fields with parameter identifiers in a user given list.
delparam Delete parameters by identifier
Deletes all fields with parameter identifiers in a user given list.
selcode Select parameters by code number
Selects all fields with code numbers in a user given list or range.
delcode Delete parameters by code number
Deletes all fields with code numbers in a user given list or range.
selname Select parameters by name
Selects all fields with parameter names in a user given list.
delname Delete parameters by name
Deletes all fields with parameter names in a user given list.
selstdname Select parameters by standard name
Selects all fields with standard names in a user given list.
sellevel Select levels
Selects all fields with levels in a user given list.
sellevidx Select levels by index
Selects all fields with index of levels in a user given list or range.
selgrid Select grids
Selects all fields with grids in a user given list.

55
Selection Reference manual

selzaxis Select z-axes

Selects all fields with z-axes in a user given list.
selzaxisname Select z-axes by name
Selects all fields with z-axis names in a user given list.
selltype Select GRIB level types
Selects all fields with GRIB level type in a user given list or range.
seltabnum Select parameter table numbers
Selects all fields with parameter table numbers in a user given list or range.

Parameter
parameter STRING Comma-separated list of parameter identifiers.
codes INTEGER Comma-separated list or first/last[/inc] range of code numbers.
names STRING Comma-separated list of variable names.
stdnames STRING Comma-separated list of standard names.
levels FLOAT Comma-separated list of vertical levels.
levidx INTEGER Comma-separated list or first/last[/inc] range of index of levels.
ltypes INTEGER Comma-separated list or first/last[/inc] range of GRIB level types.
grids STRING Comma-separated list of grid names or numbers.
zaxes STRING Comma-separated list of z-axis types or numbers.
zaxisnames STRING Comma-separated list of z-axis names.
tabnums INTEGER Comma-separated list or range of parameter table numbers.

Example
Assume an input dataset has three variables with the code numbers 129, 130 and 139. To select the
variables with the code number 129 and 139 use:
cdo selcode,129,139 infile outfile

You can also select the code number 129 and 139 by deleting the code number 130 with:
cdo delcode,130 infile outfile

56
Reference manual Selection

2.3.4. SELTIME - Select timesteps

Synopsis

seltimestep,timesteps infile outfile

seltime,times infile outfile

selhour,hours infile outfile

selday,days infile outfile

selmonth,months infile outfile

selyear,years infile outfile

selseason,seasons infile outfile

seldate,startdate[,enddate] infile outfile

selsmon,month[,nts1[,nts2]] infile outfile

Description
This module selects user specified timesteps from infile and writes them to outfile. The timesteps
selected depends on the chosen operator and the parameters. A range of integer values can be specified
by first/last[/inc].

Operators

seltimestep Select timesteps

Selects all timesteps with a timestep in a user given list or range.

seltime Select times

Selects all timesteps with a time in a user given list or range.

selhour Select hours

Selects all timesteps with a hour in a user given list or range.

selday Select days

Selects all timesteps with a day in a user given list or range.

selmonth Select months

Selects all timesteps with a month in a user given list or range.

selyear Select years

Selects all timesteps with a year in a user given list or range.

selseason Select seasons

Selects all timesteps with a month of a season in a user given list.

seldate Select dates

Selects all timesteps with a date in a user given range.

selsmon Select single month

Selects a month and optional an arbitrary number of timesteps before and after this
month.

57
Selection Reference manual

Parameter
timesteps INTEGER Comma-separated list or first/last[/inc] range of timesteps. Negative val-
ues select timesteps from the end (NetCDF only).
times STRING Comma-separated list of times (format hh:mm:ss).
hours INTEGER Comma-separated list or first/last[/inc] range of hours.
days INTEGER Comma-separated list or first/last[/inc] range of days.
months INTEGER Comma-separated list or first/last[/inc] range of months.
years INTEGER Comma-separated list or first/last[/inc] range of years.
seasons STRING Comma-separated list of seasons (substring of DJFMAMJJASOND or
ANN).
startdate STRING Start date (format YYYY-MM-DDThh:mm:ss).
enddate STRING End date (format YYYY-MM-DDThh:mm:ss) [default: startdate].
nts1 INTEGER Number of timesteps before the selected month [default: 0].
nts2 INTEGER Number of timesteps after the selected month [default: nts1].

58
Reference manual Selection

2.3.5. SELBOX - Select a box

Synopsis

sellonlatbox,lon1,lon2,lat1,lat2 infile outfile

selindexbox,idx1,idx2,idy1,idy2 infile outfile

Description
Selects grid cells inside a lon/lat or index box.

Operators

sellonlatbox Select a longitude/latitude box

Selects grid cells inside a lon/lat box. The user must specify the longitude and
latitude of the edges of the box. Only those grid cells are considered whose grid
center lies within the lon/lat box. For rotated lon/lat grids the parameters must
be specified in rotated coordinates.
selindexbox Select an index box
Selects grid cells within an index box. The user must specify the indices of the
edges of the box. The index of the left edge can be greater then the one of the
right edge. Use negative indexing to start from the end. The input grid must be a
regular lon/lat or a 2D curvilinear grid.

Parameter
lon1 FLOAT Western longitude in degrees
lon2 FLOAT Eastern longitude in degrees
lat1 FLOAT Southern or northern latitude in degrees
lat2 FLOAT Northern or southern latitude in degrees
idx1 INTEGER Index of first longitude (1 - nlon)
idx2 INTEGER Index of last longitude (1 - nlon)
idy1 INTEGER Index of first latitude (1 - nlat)
idy2 INTEGER Index of last latitude (1 - nlat)

Example
To select the region with the longitudes from 30W to 60E and latitudes from 30N to 80N from all
input fields use:
cdo sellonlatbox,-30,60,30,80 infile outfile

If the input dataset has fields on a Gaussian N16 grid, the same box can be selected with selindexbox
by:
cdo selindexbox,60,11,3,11 infile outfile

59
Selection Reference manual

2.3.6. SELREGION - Select horizontal regions

Synopsis

selregion,regions infile outfile

selcircle[,parameter] infile outfile

Description
Selects all grid cells with the center point inside user defined regions or a circle. The resulting grid is
unstructured.

Operators

selregion Select cells inside regions

Selects all grid cells with the center point inside the regions. Regions can be defined
by the user via an ASCII file. Each region consists of the geographic coordinates of
a convex polygon. Each line of a polygon description file contains the longitude and
latitude of one point. Each polygon description file can contain one or more polygons
separated by a line with the character &.
Predefined regions of countries can be specified via the country codes. A country is
specified with dcw:<CountryCode>. Country codes can be combined with the plus
sign.
selcircle Select cells inside a circle
Selects all grid cells with the center point inside a circle. The circle is described by
geographic coordinates of the center and the radius of the circle.

Parameter
regions STRING Comma-separated list of ASCII formatted files with different regions
lon FLOAT Longitude of the center of the circle in degrees, default lon=0.0
lat FLOAT Latitude of the center of the circle in degrees, default lat=0.0
radius STRING Radius of the circle, default radius=1deg (units: deg, rad, km, m)

Example
To select all grid cells of a country use the country code with data from the Digital Chart of the
World. Here is an example for Spain with the country code ES:
cdo selregion,dcw:ES infile outfile

60
Reference manual Selection

2.3.7. SELGRIDCELL - Select grid cells

Synopsis

< operator >,indices infile outfile

Description
The operator selects grid cells of all fields from infile. The user must specify the index of each grid
cell. The resulting grid in outfile is unstructured.

Operators

selgridcell Select grid cells

delgridcell Delete grid cells

Parameter
indices INTEGER Comma-separated list or first/last[/inc] range of indices

2.3.8. SAMPLEGRID - Resample grid

Synopsis

samplegrid,factor infile outfile

Description
This is a special operator for resampling the horizontal grid. No interpolation takes place. Resample
factor=2 means every second grid point is removed. Only rectilinear and curvilinear source grids are
supported by this operator.

Parameter
factor INTEGER Resample factor, typically 2, which will half the resolution

2.3.9. SELYEARIDX - Select year by index

Synopsis

selyearidx infile1 infile2 outfile

Description
Selects field elements from infile2 by a yearly time index from infile1. The yearly indices in
infile1 should be the result of corresponding yearminidx and yearmaxidx operations, respectively.

61
Selection Reference manual

2.3.10. SELSURFACE - Extract surface

Synopsis

< operator > infile outfile

isosurface,isovalue infile outfile

Description
This module computes a surface from all 3D variables. The result is a horizonal 2D field.

Operators

bottomvalue Extract bottom level

This operator selects the valid values at the bottom level. The NetCDF CF com-
pliant attribute positive is used to determine where top and bottom are. If this
attribute is missing, low values are bottom and high values are top.
topvalue Extract top level
This operator selects the valid values at the top level. The NetCDF CF compliant
attribute positive is used to determine where top and bottom are. If this attribute
is missing, low values are bottom and high values are top.
isosurface Extract isosurface
This operator computes an isosurface. The value of the isosurfce is specified by the
parameter isovalue. The isosurface is calculated by linear interpolation between
two layers.

Parameter
isovalue FLOAT Isosurface value

62
Reference manual Conditional selection

2.4. Conditional selection

This section contains modules to conditional select field elements. The fields in the first input file are
handled as a mask. A value not equal to zero is treated as "true", zero is treated as "false".
Here is a short overview of all operators in this section:

ifthen If then
ifnotthen If not then

ifthenelse If then else

ifthenc If then constant

ifnotthenc If not then constant

reducegrid Reduce input file variables to locations, where mask is non-zero.

63
Conditional selection Reference manual

2.4.1. COND - Conditional select one field

Synopsis

< operator > infile1 infile2 outfile

Description
This module selects field elements from infile2 with respect to infile1 and writes them to outfile.
The fields in infile1 are handled as a mask. A value not equal to zero is treated as "true", zero is
treated as "false". The number of fields in infile1 has either to be the same as in infile2 or the
same as in one timestep of infile2 or only one. The fields in outfile inherit the meta data from
infile2.

Operators

ifthen If then 
i2 (t, x) if i1 (t, x) ̸= 0 ∧ i1 (t, x) ̸= miss
o(t, x) =
miss if i1 (t, x) = 0 ∨ i1 (t, x) = miss
ifnotthen If not then

i2 (t, x) if i1 (t, x) = 0 ∧ i1 (t, x) ̸= miss
o(t, x) =
miss if i1 [t, x) ̸= 0 ∨ i1 (t, x) = miss

Example
To select all field elements of infile2 if the corresponding field element of infile1 is greater than
0 use:
cdo ifthen infile1 infile2 outfile

2.4.2. COND2 - Conditional select two fields

Synopsis

ifthenelse infile1 infile2 infile3 outfile

Description
This operator selects field elements from infile2 or infile3 with respect to infile1 and writes
them to outfile. The fields in infile1 are handled as a mask. A value not equal to zero is treated
as "true", zero is treated as "false". The number of fields in infile1 has either to be the same as in
infile2 or the same as in one timestep of infile2 or only one. infile2 and infile3 need to have
the same number of fields. The fields in outfile inherit the meta data from infile2.

 i2 (t, x) if i1 (t, x) ̸= 0 ∧ i1 (t, x) ̸= miss
o(t, x) = i3 (t, x) if i1 (t, x) = 0 ∧ i1 (t, x) ̸= miss

miss if i1 (t, x) = miss

Example
To select all field elements of infile2 if the corresponding field element of infile1 is greater than
0 and from infile3 otherwise use:
cdo ifthenelse infile1 infile2 infile3 outfile

64
Reference manual Conditional selection

2.4.3. CONDC - Conditional select a constant

Synopsis

< operator >,c infile outfile

Description
This module creates fields with a constant value or missing value. The fields in infile are handled
as a mask. A value not equal to zero is treated as "true", zero is treated as "false".

Operators

ifthenc If then constant


c if i(t, x) ̸= 0 ∧ i(t, x) ̸= miss
o(t, x) =
miss if i(t, x) = 0 ∨ i(t, x) = miss
ifnotthenc If not then
 constant
c if i(t, x) = 0 ∧ i(t, x) ̸= miss
o(t, x) =
miss if i(t, x) ̸= 0 ∨ i(t, x) = miss

Parameter
c FLOAT Constant

Example
To create fields with the constant value 7 if the corresponding field element of infile is greater than
0 use:
cdo ifthenc,7 infile outfile

65
Conditional selection Reference manual

2.4.4. MAPREDUCE - Reduce fields to user-defined mask

Synopsis

reducegrid,mask[,limitCoordsOutput] infile outfile

Description
This module holds an operator for data reduction based on a user defined mask. The output grid
is unstructured and includes coordinate bounds. Bounds can be avoided by using the additional
’nobounds’ keyword. With ’nocoords’ given, coordinates a completely suppressed.

Parameter
mask STRING file which holds the mask field
limitCoordsOutput STRING optional parameter to limit coordinates output: ’nobounds’ dis-
ables coordinate bounds, ’nocoords’ avoids all coordinate information

Example
To limit data fields to land values, a mask has to be created first with
cdo -gtc,0 -topo,ni96 lsm_gme96.grb

Here a GME grid is used. Say temp_gme96.grb contains a global temperture field. The following
command limits the global grid to landpoints.
cdo -f nc reduce,lsm_gme96.grb temp_gme96.grb tempOnLand_gme96.nc

Note that output file type is NetCDF, because unstructured grids cannot be stored in GRIB format.

66
Reference manual Comparison

2.5. Comparison
This section contains modules to compare datasets. The resulting field is a mask containing 1 if the
comparison is true and 0 if not.
Here is a short overview of all operators in this section:

eq Equal
ne Not equal
le Less equal
lt Less than
ge Greater equal
gt Greater than

eqc Equal constant

nec Not equal constant
lec Less equal constant
ltc Less than constant
gec Greater equal constant
gtc Greater than constant

ymoneq Compare time series with Equal

ymonne Compare time series with NotEqual
ymonle Compare time series with LessEqual
ymonlt Compares if time series with LessThan
ymonge Compares if time series with GreaterEqual
ymongt Compares if time series with GreaterThan

67
Comparison Reference manual

2.5.1. COMP - Comparison of two fields

Synopsis

< operator > infile1 infile2 outfile

Description
This module compares two datasets field by field. The resulting field is a mask containing 1 if the
comparison is true and 0 if not. The number of fields in infile1 should be the same as in infile2.
One of the input files can contain only one timestep or one field. The fields in outfile inherit the
meta data from infile1 or infile2. The type of comparison depends on the chosen operator.

Operators

eq Equal 
 1 if i1 (t, x) = i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) ̸= i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss
ne Not equal
 1 if i1 (t, x) ̸= i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) = i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss
le Less equal
 1 if i1 (t, x) ≤ i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) > i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss
lt Less than
 1 if i1 (t, x) < i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) ≥ i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss
ge Greater equal

 1 if i1 (t, x) ≥ i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) < i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss
gt Greater than

 1 if i1 (t, x) > i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
o(t, x) = 0 if i1 (t, x) ≤ i2 (t, x) ∧ i1 (t, x), i2 (t, x) ̸= miss
∨ i2 (t, x) = miss

miss if i1 (t, x) = miss

Example
To create a mask containing 1 if the elements of two fields are the same and 0 if the elements are
different use:
cdo eq infile1 infile2 outfile

68
Reference manual Comparison

2.5.2. COMPC - Comparison of a field with a constant

Synopsis

< operator >,c infile outfile

Description
This module compares all fields of a dataset with a constant. The resulting field is a mask containing
1 if the comparison is true and 0 if not. The type of comparison depends on the chosen operator.

Operators

eqc Equal constant


 1 if i(t, x) = c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) ̸= c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


nec Not equalconstant

 1 if i(t, x) ̸= c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) = c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


lec Less equalconstant

 1 if i(t, x) ≤ c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) > c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


ltc Less thanconstant

 1 if i(t, x) < c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) ≥ c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


gec Greater equal

 constant
 1 if i(t, x) ≥ c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) < c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


gtc Greater than

 constant
 1 if i(t, x) > c ∧ i(t, x), c ̸= miss
o(t, x) = 0 if i(t, x) ≤ c ∧ i(t, x), c ̸= miss
miss if i(t, x) = miss ∨ c = miss


Parameter
c FLOAT Constant

Example
To create a mask containing 1 if the field element is greater than 273.15 and 0 if not use:
cdo gtc,273.15 infile outfile

69
Comparison Reference manual

2.5.3. YMONCOMP - Multi-year monthly comparison

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs compaisons of a time series and one timestep with the same month of year.
For each field in infile1 the corresponding field of the timestep in infile2 with the same month of
year is used. The resulting field is a mask containing 1 if the comparison is true and 0 if not. The
type of comparison depends on the chosen operator. The input files need to have the same structure
with the same variables. Usually infile2 is generated by an operator of the module YMONSTAT.

Operators

ymoneq Compare time series with Equal

Compares whether a time series is equal to a multi-year monthly time series.
ymonne Compare time series with NotEqual
Compares whether a time series is not equal to a multi-year monthly time series.
ymonle Compare time series with LessEqual
Compares whether a time series is less than or equal to a multi-year monthly time series.
ymonlt Compares if time series with LessThan
Compares whether a time series is less than a multi-year monthly time series.
ymonge Compares if time series with GreaterEqual
Compares whether a time series is greater than or equal to a multi-year monthly time
series.
ymongt Compares if time series with GreaterThan
Compares whether a time series is greater than a multi-year monthly time series.

70
Reference manual Modification

2.6. Modification
This section contains modules to modify the metadata, fields or part of a field in a dataset.
Here is a short overview of all operators in this section:

setattribute Set attributes

setpartabp Set parameter table

setpartabn Set parameter table

setcodetab Set parameter code table

setcode Set code number
setparam Set parameter identifier
setname Set variable name
setunit Set variable unit
setlevel Set level
setltype Set GRIB level type
setmaxsteps Set max timesteps

setdate Set date

settime Set time of the day
setday Set day
setmon Set month
setyear Set year
settunits Set time units
settaxis Set time axis
settbounds Set time bounds
setreftime Set reference time
setcalendar Set calendar
shifttime Shift timesteps

chcode Change code number

chparam Change parameter identifier
chname Change variable or coordinate name
chunit Change variable unit
chlevel Change level
chlevelc Change level of one code
chlevelv Change level of one variable

setgrid Set grid

setgridtype Set grid type
setgridarea Set grid cell area
setgridmask Set grid mask

setzaxis Set z-axis

genlevelbounds Generate level bounds

invertlat Invert latitudes

invertlev Invert levels

shiftx Shift x
shifty Shift y

maskregion Mask regions

71
Modification Reference manual

masklonlatbox Mask a longitude/latitude box

maskindexbox Mask an index box

setclonlatbox Set a longitude/latitude box to constant

setcindexbox Set an index box to constant

enlarge Enlarge fields

setmissval Set a new missing value

setctomiss Set constant to missing value
setmisstoc Set missing value to constant
setrtomiss Set range to missing value
setvrange Set valid range
setmisstonn Set missing value to nearest neighbor
setmisstodis Set missing value to distance-weighted average

vertfillmiss Vertical filling of missing values

timfillmiss Temporal filling of missing values

setgridcell Set the value of a grid cell

72
Reference manual Modification

2.6.1. SETATTRIBUTE - Set attributes

Synopsis

setattribute,attributes infile outfile

Description
This operator sets attributes of a dataset and writes the result to outfile. The new attributes are
only available in outfile if the file format supports attributes.
Each attribute has the following structure:
[var_nm@]att_nm[:s|d|i]=[att_val|{[var_nm@]att_nm}]

var_nm Variable name (optional). Example: pressure

att_nm Attribute name. Example: units
att_val Comma-separated list of attribute values. Example: pascal

The value of var_nm is the name of the variable containing the attribute (named att_nm) that
you want to set. Use wildcards to set the attribute att_nm to more than one variable. A value
of var_nm of ’*’ will set the attribute att_nm to all data variables. If var_nm is missing then
att_nm refers to a global attribute.
The value of att_nm is the name of the attribute you want to set. For each attribute a string
(att_nm:s), a double (att_nm:d) or an integer (att_nm:i) type can be defined. By default the native
type is set.
The value of att_val is the contents of the attribute att_nm. att_val may be a single value or
one-dimensional array of elements. The type and the number of elements of an attribute will be
detected automatically from the contents of the values. An already existing attribute att_nm will
be overwritten or it will be removed if att_val is omitted. Alternatively, the values of an existing
attribute can be copied. This attribute must then be enclosed in curly brackets.
A special meaning has the attribute name FILE. If this is the 1st attribute then all attributes are
read from a file specified in the value of att_val.

Parameter
attributes STRING Comma-separated list of attributes.

Note
Attributes are evaluated by CDO when opening infile. Therefor the result of this operator is not
available for other operators when this operator is used in chaining operators.

Example
To set the units of the variable pressure to pascal use:
cdo setattribute,pressure@units=pascal infile outfile

To set the global text attribute "my_att" to "my contents", use:

cdo setattribute,my_att="my contents" infile outfile

Result of ’ncdump -h outfile’:

73
Modification Reference manual

netcdf o u t f i l e {
dimensions : . . .

variables : ...

// g l o b a l a t t r i b u t e s :
: my_att = "my c o n t e n t s " ;
}

74
Reference manual Modification

2.6.2. SETPARTAB - Set parameter table

Synopsis

< operator >,table[,convert] infile outfile

Description
This module transforms data and metadata of infile via a parameter table and writes the result
to outfile. A parameter table is an ASCII formatted file with a set of parameter entries for each
variable. Each new set have to start with "&parameter" and to end with "/".
The following parameter table entries are supported:

Entry Type Description

name WORD Name of the variable
out_name WORD New name of the variable
param WORD Parameter identifier (GRIB1: code[.tabnum]; GRIB2: num[.cat[.dis]])
out_param WORD New parameter identifier
type WORD Data type (real or double)
standard_name WORD As defined in the CF standard name table
long_name STRING Describing the variable
units STRING Specifying the units for the variable
comment STRING Information concerning the variable
cell_methods STRING Information concerning calculation of means or climatologies
cell_measures STRING Indicates the names of the variables containing cell areas and volumes
missing_value FLOAT Specifying how missing data will be identified
valid_min FLOAT Minimum valid value
valid_max FLOAT Maximum valid value
ok_min_mean_abs FLOAT Minimum absolute mean
ok_max_mean_abs FLOAT Maximum absolute mean
factor FLOAT Scale factor
delete INTEGER Set to 1 to delete variable
convert INTEGER Set to 1 to convert the unit if necessary

Unsupported parameter table entries are stored as variable attributes. The search key for the variable
depends on the operator. Use setpartabn to search variables by the name. This is typically used for
NetCDF datasets. The operator setpartabp searches variables by the parameter ID.

Operators

setpartabp Set parameter table

Search variables by the parameter identifier.
setpartabn Set parameter table
Search variables by name.

Parameter
table STRING Parameter table file or name
convert STRING Converts the units if necessary

75
Modification Reference manual

Example
Here is an example of a parameter table for one variable:
prompt> cat mypartab
&parameter
name = t
out_name = ta
standard_name = air_temperature
units = "K"
missing_value = 1.0e+20
valid_min = 157.1
valid_max = 336.3
/

To apply this parameter table to a dataset use:

cdo setpartabn,mypartab,convert infile outfile

This command renames the variable t to ta. The standard name of this variable is set to air_temperature
and the unit is set to [K] (converts the unit if necessary). The missing value will be set to 1.0e+20.
In addition it will be checked whether the values of the variable are in the range of 157.1 to 336.3.

76
Reference manual Modification

2.6.3. SET - Set field info

Synopsis

setcodetab,table infile outfile

setcode,code infile outfile
setparam,param infile outfile
setname,name infile outfile
setunit,unit infile outfile
setlevel,level infile outfile
setltype,ltype infile outfile
setmaxsteps,maxsteps infile outfile

Description
This module sets some field information. Depending on the chosen operator the parameter table,
code number, parameter identifier, variable name or level is set.

Operators

setcodetab Set parameter code table

Sets the parameter code table for all variables.
setcode Set code number
Sets the code number for all variables to the same given value.
setparam Set parameter identifier
Sets the parameter identifier of the first variable.
setname Set variable name
Sets the name of the first variable.
setunit Set variable unit
Sets the unit of the first variable.
setlevel Set level
Sets the first level of all variables.
setltype Set GRIB level type
Sets the GRIB level type of all variables.
setmaxsteps Set max timesteps
Sets maximum number of timesteps

Parameter
table STRING Parameter table file or name
code INTEGER Code number
param STRING Parameter identifier (GRIB1: code[.tabnum]; GRIB2: num[.cat[.dis]])
name STRING Variable name
level FLOAT New level
ltype INTEGER GRIB level type
maxsteps INTEGER Maximum number of timesteps

77
Modification Reference manual

2.6.4. SETTIME - Set time

Synopsis

setdate,date infile outfile

settime,time infile outfile
setday,day infile outfile
setmon,month infile outfile
setyear,year infile outfile
settunits,units infile outfile
settaxis,date,time[,inc] infile outfile
settbounds,frequency infile outfile
setreftime,date,time[,units] infile outfile
setcalendar,calendar infile outfile
shifttime,shiftValue infile outfile

Description
This module sets the time axis or part of the time axis. Which part of the time axis is overwrit-
ten/created depends on the chosen operator. The number of time steps does not change.

Operators

setdate Set date

Sets the date in every timestep to the same given value.
settime Set time of the day
Sets the time in every timestep to the same given value.
setday Set day
Sets the day in every timestep to the same given value.
setmon Set month
Sets the month in every timestep to the same given value.
setyear Set year
Sets the year in every timestep to the same given value.
settunits Set time units
Sets the base units of a relative time axis.
settaxis Set time axis
Sets the time axis.
settbounds Set time bounds
Sets the time bounds.
setreftime Set reference time
Sets the reference time of a relative time axis.
setcalendar Set calendar
Sets the calendar attribute of a relative time axis.
shifttime Shift timesteps
Shifts all timesteps by the parameter shiftValue.

78
Reference manual Modification

Parameter
day INTEGER Value of the new day
month INTEGER Value of the new month
year INTEGER Value of the new year
units STRING Base units of the time axis (seconds, minutes, hours, days, months, years)
date STRING Date (format: YYYY-MM-DD)
time STRING Time (format: hh:mm:ss)
inc STRING Optional increment (seconds, minutes, hours, days, months, years) [de-
fault: 1hour]
frequency STRING Frequency of the time series (hour, day, month, year)
calendar STRING Calendar (standard, proleptic_gregorian, 360_day, 365_day, 366_day)
shiftValue STRING Shift value (e.g. -3hour)

Example
To set the time axis to 1987-01-16 12:00:00 with an increment of one month for each timestep use:
cdo settaxis,1987-01-16,12:00:00,1mon infile outfile

Result of ’cdo showdate outfile’ for a dataset with 12 timesteps:

1987−01−16 1987−02−16 1987−03−16 1987−04−16 1987−05−16 1987−06−16 \
1987−07−16 1987−08−16 1987−09−16 1987−10−16 1987−11−16 1987−12−16

To shift this time axis by -15 days use:

cdo shifttime,-15days infile outfile

Result of ’cdo showdate outfile’:

1987−01−01 1987−02−01 1987−03−01 1987−04−01 1987−05−01 1987−06−01 \
1987−07−01 1987−08−01 1987−09−01 1987−10−01 1987−11−01 1987−12−01

79
Modification Reference manual

2.6.5. CHANGE - Change field header

Synopsis
chcode,oldcode,newcode[,...] infile outfile
chparam,oldparam,newparam,... infile outfile
chname,oldname,newname,... infile outfile
chunit,oldunit,newunit,... infile outfile
chlevel,oldlev,newlev,... infile outfile
chlevelc,code,oldlev,newlev infile outfile
chlevelv,name,oldlev,newlev infile outfile

Description
This module reads fields from infile, changes some header values and writes the results to outfile.
The kind of changes depends on the chosen operator.

Operators
chcode Change code number
Changes some user given code numbers to new user given values.
chparam Change parameter identifier
Changes some user given parameter identifiers to new user given values.
chname Change variable or coordinate name
Changes some user given variable or coordinate names to new user given names.
chunit Change variable unit
Changes some user given variable units to new user given units.
chlevel Change level
Changes some user given levels to new user given values.
chlevelc Change level of one code
Changes one level of a user given code number.
chlevelv Change level of one variable
Changes one level of a user given variable name.

Parameter
code INTEGER Code number
oldcode,newcode,... INTEGER Pairs of old and new code numbers
oldparam,newparam,... STRING Pairs of old and new parameter identifiers
name STRING Variable name
oldname,newname,... STRING Pairs of old and new variable names
oldlev FLOAT Old level
newlev FLOAT New level
oldlev,newlev,... FLOAT Pairs of old and new levels

Example
To change the code number 98 to 179 and 99 to 211 use:
cdo chcode,98,179,99,211 infile outfile

80
Reference manual Modification

2.6.6. SETGRID - Set grid information

Synopsis
setgrid,grid infile outfile
setgridtype,gridtype infile outfile
setgridarea,gridarea infile outfile
setgridmask,gridmask infile outfile

Description
This module modifies the metadata of the horizontal grid. Depending on the chosen operator a new
grid description is set, the coordinates are converted or the grid cell area is added.

Operators
setgrid Set grid
Sets a new grid description. The input fields need to have the same grid size as the
size of the target grid description.
setgridtype Set grid type
Sets the grid type of all input fields. The following grid types are available:
curvilinear Converts a regular grid to a curvilinear grid
unstructured Converts a regular or curvilinear grid to an unstructured grid
dereference Dereference a reference to a grid
regular Linear interpolation of a reduced Gaussian grid to a regular Gaus-
sian grid
regularnn Nearest neighbor interpolation of a reduced Gaussian grid to a
regular Gaussian grid
lonlat Converts a regular lonlat grid stored as a curvilinear grid back to
a lonlat grid
projection Removes the geographical coordinates if projection parameter avail-
able
setgridarea Set grid cell area
Sets the grid cell area. The parameter gridarea is the path to a data file, the first
field is used as grid cell area. The input fields need to have the same grid size as the
grid cell area. The grid cell area is used to compute the weights of each grid cell if
needed by an operator, e.g. for fldmean.
setgridmask Set grid mask
Sets the grid mask. The parameter gridmask is the path to a data file, the first field
is used as the grid mask. The input fields need to have the same grid size as the
grid mask. The grid mask is used as the target grid mask for remapping, e.g. for
remapbil.

Parameter
grid STRING Grid description file or name
gridtype STRING Grid type (curvilinear, unstructured, regular, lonlat, projection or derefer-
ence)
gridarea STRING Data file, the first field is used as grid cell area
gridmask STRING Data file, the first field is used as grid mask

81
Modification Reference manual

Example
Assuming a dataset has fields on a grid with 2048 elements without or with wrong grid description.
To set the grid description of all input fields to a Gaussian N32 grid (8192 gridpoints) use:
cdo setgrid,n32 infile outfile

2.6.7. SETZAXIS - Set z-axis information

Synopsis

setzaxis,zaxis infile outfile

genlevelbounds[,zbot[,ztop]] infile outfile

Description
This module modifies the metadata of the vertical grid.

Operators

setzaxis Set z-axis

This operator sets the z-axis description of all variables with the same number
of level as the new z-axis.
genlevelbounds Generate level bounds
Generates the layer bounds of the z-axis.

Parameter
zaxis STRING Z-axis description file or name of the target z-axis
zbot FLOAT Specifying the bottom of the vertical column. Must have the same units as
z-axis.
ztop FLOAT Specifying the top of the vertical column. Must have the same units as z-axis.

82
Reference manual Modification

2.6.8. INVERT - Invert latitudes

Synopsis

invertlat infile outfile

Description
This operator inverts the latitudes of all fields on a rectilinear grid.

Example
To invert the latitudes of a 2D field from N->S to S->N use:
cdo invertlat infile outfile

2.6.9. INVERTLEV - Invert levels

Synopsis

invertlev infile outfile

Description
This operator inverts the levels of all 3D variables.

83
Modification Reference manual

2.6.10. SHIFTXY - Shift field

Synopsis

< operator >,<nshift>,<cyclic>,<coord> infile outfile

Description
This module contains operators to shift all fields in x or y direction. All fields need to have the same
horizontal rectilinear or curvilinear grid.

Operators

shiftx Shift x
Shifts all fields in x direction.
shifty Shift y
Shifts all fields in y direction.

Parameter
nshift INTEGER Number of grid cells to shift (default: 1)
cyclic STRING If set, cells are filled up cyclic (default: missing value)
coord STRING If set, coordinates are also shifted

Example
To shift all input fields in the x direction by +1 cells and fill the new cells with missing value, use:
cdo shiftx infile outfile

To shift all input fields in the x direction by +1 cells and fill the new cells cyclic, use:
cdo shiftx,1,cyclic infile outfile

84
Reference manual Modification

2.6.11. MASKREGION - Mask regions

Synopsis

maskregion,regions infile outfile

Description
Masks different regions of the input fields. The grid cells inside a region are untouched, the cells
outside are set to missing value. Considered are only those grid cells with the grid center inside the
regions. All input fields must have the same horizontal grid.
Regions can be defined by the user via an ASCII file. Each region consists of the geographic coordinates
of a convex polygon. Each line of a polygon description file contains the longitude and latitude of one
point. Each polygon description file can contain one or more polygons separated by a line with the
character &.
Predefined regions of countries can be specified via the country codes. A country is specified with
dcw:<CountryCode>. Country codes can be combined with the plus sign.

Parameter
regions STRING Comma-separated list of ASCII formatted files with different regions

Example
To mask the region with the longitudes from 120E to 90W and latitudes from 20N to 20S on all input
fields use:
cdo maskregion,myregion infile outfile

For this example the description file of the region myregion should contain one polygon with the
following four coordinates:
120 20
120 −20
270 −20
270 20

To mask the region of a country use the country code with data from the Digital Chart of the World.
Here is an example for Spain with the country code ES:
cdo maskregion,dcw:ES infile outfile

85
Modification Reference manual

2.6.12. MASKBOX - Mask a box

Synopsis

masklonlatbox,lon1,lon2,lat1,lat2 infile outfile

maskindexbox,idx1,idx2,idy1,idy2 infile outfile

Description
Masks grid cells inside a lon/lat or index box. The elements inside the box are untouched, the
elements outside are set to missing value. All input fields need to have the same horizontal grid. Use
sellonlatbox or selindexbox if only the data inside the box are needed.

Operators

masklonlatbox Mask a longitude/latitude box

Masks grid cells inside a lon/lat box. The user must specify the longitude and
latitude of the edges of the box. Only those grid cells are considered whose grid
center lies within the lon/lat box. For rotated lon/lat grids the parameters must
be specified in rotated coordinates.
maskindexbox Mask an index box
Masks grid cells within an index box. The user must specify the indices of the
edges of the box. The index of the left edge can be greater then the one of the
right edge. Use negative indexing to start from the end. The input grid must
be a regular lon/lat or a 2D curvilinear grid.

Parameter
lon1 FLOAT Western longitude
lon2 FLOAT Eastern longitude
lat1 FLOAT Southern or northern latitude
lat2 FLOAT Northern or southern latitude
idx1 INTEGER Index of first longitude
idx2 INTEGER Index of last longitude
idy1 INTEGER Index of first latitude
idy2 INTEGER Index of last latitude

Example
To mask the region with the longitudes from 120E to 90W and latitudes from 20N to 20S on all input
fields use:
cdo masklonlatbox,120,-90,20,-20 infile outfile

If the input dataset has fields on a Gaussian N16 grid, the same box can be masked with maskindexbox
by:
cdo maskindexbox,23,48,13,20 infile outfile

86
Reference manual Modification

2.6.13. SETBOX - Set a box to constant

Synopsis

setclonlatbox,c,lon1,lon2,lat1,lat2 infile outfile

setcindexbox,c,idx1,idx2,idy1,idy2 infile outfile

Description
Sets a box of the rectangularly understood field to a constant value. The elements outside the box
are untouched, the elements inside are set to the given constant. All input fields need to have the
same horizontal grid.

Operators

setclonlatbox Set a longitude/latitude box to constant

Sets the values of a longitude/latitude box to a constant value. The user has to
give the longitudes and latitudes of the edges of the box.
setcindexbox Set an index box to constant
Sets the values of an index box to a constant value. The user has to give the
indices of the edges of the box. The index of the left edge can be greater than the
one of the right edge.

Parameter
c FLOAT Constant
lon1 FLOAT Western longitude
lon2 FLOAT Eastern longitude
lat1 FLOAT Southern or northern latitude
lat2 FLOAT Northern or southern latitude
idx1 INTEGER Index of first longitude
idx2 INTEGER Index of last longitude
idy1 INTEGER Index of first latitude
idy2 INTEGER Index of last latitude

Example
To set all values in the region with the longitudes from 120E to 90W and latitudes from 20N to 20S
to the constant value -1.23 use:
cdo setclonlatbox,-1.23,120,-90,20,-20 infile outfile

If the input dataset has fields on a Gaussian N16 grid, the same box can be set with setcindexbox by:
cdo setcindexbox,-1.23,23,48,13,20 infile outfile

87
Modification Reference manual

2.6.14. ENLARGE - Enlarge fields

Synopsis

enlarge,grid infile outfile

Description
Enlarge all fields of infile to a user given horizontal grid. Normally only the last field element is
used for the enlargement. If however the input and output grid are regular lon/lat grids, a zonal or
meridional enlargement is possible. Zonal enlargement takes place, if the xsize of the input field is 1
and the ysize of both grids are the same. For meridional enlargement the ysize have to be 1 and the
xsize of both grids should have the same size.

Parameter
grid STRING Target grid description file or name

Example
Assumed you want to add two datasets. The first dataset is a field on a global grid (n field elements)
and the second dataset is a global mean (1 field element). Before you can add these two datasets the
second dataset have to be enlarged to the grid size of the first dataset:
cdo enlarge,infile1 infile2 tmpfile
cdo add infile1 tmpfile outfile

Or shorter using operator piping:

cdo add infile1 -enlarge,infile1 infile2 outfile

88
Reference manual Modification

2.6.15. SETMISS - Set missing value

Synopsis
setmissval,newmiss infile outfile
setctomiss,c infile outfile
setmisstoc,c infile outfile
setrtomiss,rmin,rmax infile outfile
setvrange,rmin,rmax infile outfile
setmisstonn infile outfile
setmisstodis[,neighbors] infile outfile

Description
This module sets part of a field to missing value or missing values to a constant value. Which part of
the field is set depends on the chosen operator.

Operators
setmissval Set a newmissing value
newmiss if i(t, x) = miss
o(t, x) =
i(t, x) if i(t, x) ̸= miss
setctomiss Set constant
 to missing value
miss if i(t, x) = c
o(t, x) =
i(t, x) if i(t, x) ̸= c
setmisstoc Set missing
 value to constant
c if i(t, x) = miss
o(t, x) =
i(t, x) if i(t, x) ̸= miss
setrtomiss Set rangeto missing value
miss if i(t, x) ≥ rmin ∧ i(t, x) ≤ rmax
o(t, x) =
i(t, x) if i(t, x) < rmin ∨ i(t, x) > rmax
setvrange Set valid 
range
miss if i(t, x) < rmin ∨ i(t, x) > rmax
o(t, x) =
i(t, x) if i(t, x) ≥ rmin ∧ i(t, x) ≤ rmax
setmisstonn Set missing value to nearest neighbor
Set all missing values to the nearest non missing value.

i(t, y) if i(t, x) = miss ∧ i(t, y) ̸= miss
o(t, x) =
i(t, x) if i(t, x) ̸= miss
setmisstodis Set missing value to distance-weighted average
Set all missing values to the distance-weighted average of the nearest non missing
values. The default number of nearest neighbors is 4.

Parameter
neighbors INTEGER Number of nearest neighbors
newmiss FLOAT New missing value
c FLOAT Constant
rmin FLOAT Lower bound
rmax FLOAT Upper bound

89
Modification Reference manual

Example
setrtomiss
Assume an input dataset has one field with temperatures in the range from 246 to 304 Kelvin. To set
all values below 273.15 Kelvin to missing value use:
cdo setrtomiss,0,273.15 infile outfile

Result of ’cdo info infile’:

−1 : Date Time Code L e v e l Size Miss : Minimum Mean Maximum
1 : 1987−12−31 1 2 : 0 0 : 0 0 139 0 2048 0 : 246.27 276.75 303.71

Result of ’cdo info outfile’:

−1 : Date Time Code L e v e l Size Miss : Minimum Mean Maximum
1 : 1987−12−31 1 2 : 0 0 : 0 0 139 0 2048 871 : 273.16 287.08 303.71

setmisstonn
Set all missing values to the nearest non missing value:
cdo setmisstonn infile outfile

Below is a schematic illustration of this example:

−20˚ 20˚ 60˚ −20˚ 20˚ 60˚
45˚ 45˚

15˚ 15˚

−15˚ −15˚

−45˚ −45˚
−20˚ 20˚ 60˚ −20˚ 20˚ 60˚

On the left side is input data with missing values in grey and on the right side the result with the
filled missing values.

90
Reference manual Modification

2.6.16. VERTFILLMISS - Vertical filling of missing values

Synopsis

vertfillmiss[,parameter] infile outfile

Description
This operator fills in vertical missing values. The method parameter can be used to select the filling
method. The default method=nearest fills missing values with the nearest neighbor value. Other
options are f orward and backward to fill missing values by forward or backward propagation of
values. Use the limit parameter to set the maximum number of consecutive missing values to fill and
max_gaps to set the maximum number of gaps to fill.

Parameter
method STRING Fill method [nearest|linear|forward|backward] (default: nearest)
limit INTEGER The maximum number of consecutive missing values to fill (default: all)
max_gaps INTEGER The maximum number of gaps to fill (default: all)

2.6.17. TIMFILLMISS - Temporal filling of missing values

Synopsis

timfillmiss[,parameter] infile outfile

Description
This operator fills in temporally missing values. The method parameter can be used to select the
filling method. The default method=nearest fills missing values with the nearest neighbor value.
Other options are f orward and backward to fill missing values by forward or backward propagation
of values. Use the limit parameter to set the maximum number of consecutive missing values to fill
and max_gaps to set the maximum number of gaps to fill.

Parameter
method STRING Fill method [nearest|linear|forward|backward] (default: nearest)
limit INTEGER The maximum number of consecutive missing values to fill (default: all)
max_gaps INTEGER The maximum number of gaps to fill (default: all)

91
Modification Reference manual

2.6.18. SETGRIDCELL - Set the value of a grid cell

Synopsis

setgridcell,parameter infile outfile

Description
This operator sets the value of the selected grid cells. The grid cells can be selected by a comma-
separated list of grid cell indices or a mask. The mask is read from a data file, which may contain
only one field. If no grid cells are selected, all values are set.

Parameter
value FLOAT Value of the grid cell
cell INTEGER Comma-separated list of grid cell indices
mask STRING Name of the data file which contains the mask

92
Reference manual Arithmetic

2.7. Arithmetic

This section contains modules to arithmetically process datasets.

Here is a short overview of all operators in this section:

expr Evaluate expressions

exprf Evaluate expressions script
aexpr Evaluate expressions and append results
aexprf Evaluate expression script and append results

abs Absolute value

int Integer value
nint Nearest integer value
pow Power
sqr Square
sqrt Square root
exp Exponential
ln Natural logarithm
log10 Base 10 logarithm
sin Sine
cos Cosine
tan Tangent
asin Arc sine
acos Arc cosine
atan Arc tangent
reci Reciprocal value
not Logical NOT

addc Add a constant

subc Subtract a constant
mulc Multiply with a constant
divc Divide by a constant
minc Minimum of a field and a constant
maxc Maximum of a field and a constant

add Add two fields

sub Subtract two fields
mul Multiply two fields
div Divide two fields
min Minimum of two fields
max Maximum of two fields
atan2 Arc tangent of two fields

dayadd Add daily time series

daysub Subtract daily time series
daymul Multiply daily time series
daydiv Divide daily time series

monadd Add monthly time series

monsub Subtract monthly time series
monmul Multiply monthly time series
mondiv Divide monthly time series

93
Arithmetic Reference manual

yearadd Add yearly time series

yearsub Subtract yearly time series
yearmul Multiply yearly time series
yeardiv Divide yearly time series

yhouradd Add multi-year hourly time series

yhoursub Subtract multi-year hourly time series
yhourmul Multiply multi-year hourly time series
yhourdiv Divide multi-year hourly time series

ydayadd Add multi-year daily time series

ydaysub Subtract multi-year daily time series
ydaymul Multiply multi-year daily time series
ydaydiv Divide multi-year daily time series

ymonadd Add multi-year monthly time series

ymonsub Subtract multi-year monthly time series
ymonmul Multiply multi-year monthly time series
ymondiv Divide multi-year monthly time series

yseasadd Add multi-year seasonal time series

yseassub Subtract multi-year seasonal time series
yseasmul Multiply multi-year seasonal time series
yseasdiv Divide multi-year seasonal time series

muldpm Multiply with days per month

divdpm Divide by days per month
muldpy Multiply with days per year
divdpy Divide by days per year

mulcoslat Multiply with the cosine of the latitude

divcoslat Divide by cosine of the latitude

94
Reference manual Arithmetic

2.7.1. EXPR - Evaluate expressions

Synopsis

expr,instr infile outfile

exprf ,filename infile outfile
aexpr,instr infile outfile
aexprf ,filename infile outfile

Description
This module arithmetically processes every timestep of the input dataset. Each individual assignment
statement have to end with a semi-colon. The special key _ALL_ is used as a template. A statement
with a template is replaced for all variable names. Unlike regular variables, temporary variables are
never written to the output stream. To define a temporary variable simply prefix the variable name
with an underscore (e.g. _varname) when the variable is declared.
The following operators are supported:

Operator Meaning Example Result

= assignment x=y Assigns y to x
+ addition x+y Sum of x and y
- subtraction x-y Difference of x and y
* multiplication x*y Product of x and y
/ division x/y Quotient of x and y
ˆ exponentiation xˆy Exponentiates x with y
== equal to x == y 1, if x equal to y; else 0
!= not equal to x != y 1, if x not equal to y; else 0
> greater than x>y 1, if x greater than y; else 0
< less than x<y 1, if x less than y; else 0
>= greater equal x >= y 1, if x greater equal y; else 0
<= less equal x <= y 1, if x less equal y; else 0
<=> less equal greater x <=> y -1, if x less y; 1, if x greater y; else 0
&& logical AND x && y 1, if x and y not equal 0; else 0
|| logical OR x || y 1, if x or y not equal 0; else 0
! logical NOT !x 1, if x equal 0; else 0
?: ternary conditional x? y: z y, if x not equal 0, else z

The following functions are supported:

Math intrinsics:
abs(x) Absolute value of x
floor(x) Round to largest integral value not greater than x
ceil(x) Round to smallest integral value not less than x
float(x) 32-bit float value of x
int(x) Integer value of x
nint(x) Nearest integer value of x
sqr(x) Square of x
sqrt(x) Square Root of x
exp(x) Exponential of x

95
Arithmetic Reference manual

ln(x) Natural logarithm of x

log10(x) Base 10 logarithm of x
sin(x) Sine of x, where x is specified in radians
cos(x) Cosine of x, where x is specified in radians
tan(x) Tangent of x, where x is specified in radians
asin(x) Arc-sine of x, where x is specified in radians
acos(x) Arc-cosine of x, where x is specified in radians
atan(x) Arc-tangent of x, where x is specified in radians
sinh(x) Hyperbolic sine of x, where x is specified in radians
cosh(x) Hyperbolic cosine of x, where x is specified in radians
tanh(x) Hyperbolic tangent of x, where x is specified in radians
asinh(x) Inverse hyperbolic sine of x, where x is specified in radians
acosh(x) Inverse hyperbolic cosine of x, where x is specified in radians
atanh(x) Inverse hyperbolic tangent of x, where x is specified in radians
rad(x) Convert x from degrees to radians
deg(x) Convert x from radians to degrees
rand(x) Replace x by pseudo-random numbers in the range of 0 to 1
isMissval(x) Returns 1 where x is missing

mod(x,y) Floating-point remainder of x/ y

min(x,y) Minimum value of x and y
max(x,y) Maximum value of x and y
pow(x,y) Power function
hypot(x,y) Euclidean distance function, sqrt(x*x + y*y)
atan2(x,y) Arc tangent function of y/x, using signs to determine quadrants

Coordinates:

clon(x) Longitude coordinate of x (available only if x has geographical coordinates)

clat(x) Latitude coordinate of x (available only if x has geographical coordinates)
gridarea(x) Grid cell area of x (available only if x has geographical coordinates)
gridindex(x) Grid cell indices of x
clev(x) Level coordinate of x (0, if x is a 2D surface variable)
clevidx(x) Level index of x (0, if x is a 2D surface variable)
cthickness(x) Layer thickness, upper minus lower level bound of x (1, if level bounds are missing)
ctimestep() Timestep number (1 to N)
cdate() Verification date as YYYYMMDD
ctime() Verification time as HHMMSS.millisecond
cdeltat() Difference between current and last timestep in seconds
cday() Day as DD
cmonth() Month as MM

96
Reference manual Arithmetic

cyear() Year as YYYY

csecond() Second as SS.millisecond
cminute() Minute as MM
chour() Hour as HH
Constants:
ngp(x) Number of horizontal grid points
nlev(x) Number of vertical levels
size(x) Total number of elements (ngp(x)*nlev(x))
missval(x) Returns the missing value of variable x
Statistical values over a field:
fldmin(x), fldmax(x), fldrange(x), fldsum(x), fldmean(x), fldavg(x), fldstd(x), fldstd1(x), fldvar(x),
fldvar1(x), fldskew(x), fldkurt(x), fldmedian(x)
Zonal statistical values for regular 2D grids:
zonmin(x), zonmax(x), zonrange(x), zonsum(x), zonmean(x), zonavg(x), zonstd(x), zonstd1(x), zon-
var(x), zonvar1(x), zonskew(x), zonkurt(x), zonmedian(x)
Vertical statistical values:
vertmin(x), vertmax(x), vertrange(x), vertsum(x), vertmean(x), vertavg(x), vertstd(x), vertstd1(x),
vertvar(x), vertvar1(x)
Miscellaneous:
sellevel(x,k) Select level k of variable x
sellevidx(x,k) Select level index k of variable x
sellevelrange(x,k1,k2) Select all levels of variable x in the range k1 to k2
sellevidxrange(x,k1,k2) Select all level indices of variable x in the range k1 to k2
remove(x) Remove variable x from output stream

Operators

expr Evaluate expressions

The processing instructions are read from the parameter.
exprf Evaluate expressions script
Contrary to expr the processing instructions are read from a file.
aexpr Evaluate expressions and append results
Same as expr, but keep input variables and append results
aexprf Evaluate expression script and append results
Same as exprf, but keep input variables and append results

Parameter
instr STRING Processing instructions (need to be ’quoted’ in most cases)
filename STRING File with processing instructions

Note
If the input stream contains duplicate entries of the same variable name then the last one is used.

97
Arithmetic Reference manual

Example
Assume an input dataset contains at least the variables ’aprl’, ’aprc’ and ’ts’. To create a new variable
’var1’ with the sum of ’aprl’ and ’aprc’ and a variable ’var2’ which convert the temperature ’ts’ from
Kelvin to Celsius use:
cdo expr,’var1=aprl+aprc;var2=ts-273.15;’ infile outfile

The same example, but the instructions are read from a file:
cdo exprf,myexpr infile outfile

The file myexpr contains:

var1 = a p r l + a p r c ;
var2 = t s − 2 7 3 . 1 5 ;

98
Reference manual Arithmetic

2.7.2. MATH - Mathematical functions

Synopsis

< operator > infile outfile

Description
This module contains some standard mathematical functions. All trigonometric functions calculate
with radians.

Operators

abs Absolute value

o(t, x) = abs(i(t, x))
int Integer value
o(t, x) = int(i(t, x))
nint Nearest integer value
o(t, x) = nint(i(t, x))
pow Power
y
o(t, x) = i(t, x)
sqr Square
2
o(t, x) = i(t, x)
sqrt Square root
p
o(t, x) = i(t, x)
exp Exponential
o(t, x) = ei(t,x)
ln Natural logarithm
o(t, x) = ln(i(t, x))
log10 Base 10 logarithm
o(t, x) = log10 (i(t, x))
sin Sine
o(t, x) = sin(i(t, x))
cos Cosine
o(t, x) = cos(i(t, x))
tan Tangent
o(t, x) = tan(i(t, x))
asin Arc sine
o(t, x) = arcsin(i(t, x))
acos Arc cosine
o(t, x) = arccos(i(t, x))
atan Arc tangent
o(t, x) = arctan(i(t, x))
reci Reciprocal value
o(t, x) = 1/i(t, x)
not Logical NOT
o(t, x) = 1, if xequal0; else0

99
Arithmetic Reference manual

Example
To calculate the square root for all field elements use:
cdo sqrt infile outfile

100
Reference manual Arithmetic

2.7.3. ARITHC - Arithmetic with a constant

Synopsis

< operator >,c infile outfile

Description
This module performs simple arithmetic with all field elements of a dataset and a constant. The fields
in outfile inherit the meta data from infile.

Operators

addc Add a constant

o(t, x) = i(t, x) + c
subc Subtract a constant
o(t, x) = i(t, x) − c
mulc Multiply with a constant
o(t, x) = i(t, x) ∗ c
divc Divide by a constant
o(t, x) = i(t, x)/c
minc Minimum of a field and a constant
o(t, x) = min(i(t, x), c)
maxc Maximum of a field and a constant
o(t, x) = max(i(t, x), c)

Parameter
c FLOAT Constant

Example
To sum all input fields with the constant -273.15 use:
cdo addc,-273.15 infile outfile

101
Arithmetic Reference manual

2.7.4. ARITH - Arithmetic on two datasets

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of two datasets. The number of fields in infile1 should be
the same as in infile2. The fields in outfile inherit the meta data from infile1. All operators in
this module simply process one field after the other from the two input files. Neither the order of the
variables nor the date is checked. One of the input files can contain only one timestep or one variable.

Operators

add Add two fields

o(t, x) = i1 (t, x) + i2 (t, x)
sub Subtract two fields
o(t, x) = i1 (t, x) − i2 (t, x)
mul Multiply two fields
o(t, x) = i1 (t, x) ∗ i2 (t, x)
div Divide two fields
o(t, x) = i1 (t, x)/i2 (t, x)
min Minimum of two fields
o(t, x) = min(i1 (t, x), i2 (t, x))
max Maximum of two fields
o(t, x) = max(i1 (t, x), i2 (t, x))
atan2 Arc tangent of two fields
The atan2 operator calculates the arc tangent of two fields. The result is in radians, which
is between -PI and PI (inclusive).
o(t, x) = atan2(i1 (t, x), i2 (t, x))

Example
To sum all fields of the first input file with the corresponding fields of the second input file use:
cdo add infile1 infile2 outfile

102
Reference manual Arithmetic

2.7.5. DAYARITH - Daily arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same day, month
and year. For each field in infile1 the corresponding field of the timestep in infile2 with the same
day, month and year is used. The input files need to have the same structure with the same variables.
Usually infile2 is generated by an operator of the module DAYSTAT.

Operators

dayadd Add daily time series

Adds a time series and a daily time series.
daysub Subtract daily time series
Subtracts a time series and a daily time series.
daymul Multiply daily time series
Multiplies a time series and a daily time series.
daydiv Divide daily time series
Divides a time series and a daily time series.

Example
To subtract a daily time average from a time series use:
cdo daysub infile -dayavg infile outfile

103
Arithmetic Reference manual

2.7.6. MONARITH - Monthly arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same month and
year. For each field in infile1 the corresponding field of the timestep in infile2 with the same
month and year is used. The input files need to have the same structure with the same variables.
Usually infile2 is generated by an operator of the module MONSTAT.

Operators

monadd Add monthly time series

Adds a time series and a monthly time series.
monsub Subtract monthly time series
Subtracts a time series and a monthly time series.
monmul Multiply monthly time series
Multiplies a time series and a monthly time series.
mondiv Divide monthly time series
Divides a time series and a monthly time series.

Example
To subtract a monthly time average from a time series use:
cdo monsub infile -monavg infile outfile

104
Reference manual Arithmetic

2.7.7. YEARARITH - Yearly arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same year. For
each field in infile1 the corresponding field of the timestep in infile2 with the same year is used.
The header information in infile1 have to be the same as in infile2. Usually infile2 is generated
by an operator of the module YEARSTAT.

Operators

yearadd Add yearly time series

Adds a time series and a yearly time series.
yearsub Subtract yearly time series
Subtracts a time series and a yearly time series.
yearmul Multiply yearly time series
Multiplies a time series and a yearly time series.
yeardiv Divide yearly time series
Divides a time series and a yearly time series.

Example
To subtract a yearly time average from a time series use:
cdo yearsub infile -yearavg infile outfile

105
Arithmetic Reference manual

2.7.8. YHOURARITH - Multi-year hourly arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same hour and
day of year. For each field in infile1 the corresponding field of the timestep in infile2 with the
same hour and day of year is used. The input files need to have the same structure with the same
variables. Usually infile2 is generated by an operator of the module YHOURSTAT.

Operators

yhouradd Add multi-year hourly time series

Adds a time series and a multi-year hourly time series.
yhoursub Subtract multi-year hourly time series
Subtracts a time series and a multi-year hourly time series.
yhourmul Multiply multi-year hourly time series
Multiplies a time series and a multi-year hourly time series.
yhourdiv Divide multi-year hourly time series
Divides a time series and a multi-year hourly time series.

Example
To subtract a multi-year hourly time average from a time series use:
cdo yhoursub infile -yhouravg infile outfile

106
Reference manual Arithmetic

2.7.9. YDAYARITH - Multi-year daily arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same day of year.
For each field in infile1 the corresponding field of the timestep in infile2 with the same day of year
is used. The input files need to have the same structure with the same variables. Usually infile2 is
generated by an operator of the module YDAYSTAT.

Operators

ydayadd Add multi-year daily time series

Adds a time series and a multi-year daily time series.
ydaysub Subtract multi-year daily time series
Subtracts a time series and a multi-year daily time series.
ydaymul Multiply multi-year daily time series
Multiplies a time series and a multi-year daily time series.
ydaydiv Divide multi-year daily time series
Divides a time series and a multi-year daily time series.

Example
To subtract a multi-year daily time average from a time series use:
cdo ydaysub infile -ydayavg infile outfile

107
Arithmetic Reference manual

2.7.10. YMONARITH - Multi-year monthly arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same month of
year. For each field in infile1 the corresponding field of the timestep in infile2 with the same
month of year is used. The input files need to have the same structure with the same variables.
Usually infile2 is generated by an operator of the module YMONSTAT.

Operators

ymonadd Add multi-year monthly time series

Adds a time series and a multi-year monthly time series.
ymonsub Subtract multi-year monthly time series
Subtracts a time series and a multi-year monthly time series.
ymonmul Multiply multi-year monthly time series
Multiplies a time series with a multi-year monthly time series.
ymondiv Divide multi-year monthly time series
Divides a time series by a multi-year monthly time series.

Example
To subtract a multi-year monthly time average from a time series use:
cdo ymonsub infile -ymonavg infile outfile

108
Reference manual Arithmetic

2.7.11. YSEASARITH - Multi-year seasonal arithmetic

Synopsis

< operator > infile1 infile2 outfile

Description
This module performs simple arithmetic of a time series and one timestep with the same season. For
each field in infile1 the corresponding field of the timestep in infile2 with the same season is
used. The input files need to have the same structure with the same variables. Usually infile2 is
generated by an operator of the module YSEASSTAT.

Operators

yseasadd Add multi-year seasonal time series

Adds a time series and a multi-year seasonal time series.
yseassub Subtract multi-year seasonal time series
Subtracts a time series and a multi-year seasonal time series.
yseasmul Multiply multi-year seasonal time series
Multiplies a time series and a multi-year seasonal time series.
yseasdiv Divide multi-year seasonal time series
Divides a time series and a multi-year seasonal time series.

Example
To subtract a multi-year seasonal time average from a time series use:
cdo yseassub infile -yseasavg infile outfile

109
Arithmetic Reference manual

2.7.12. ARITHDAYS - Arithmetic with days

Synopsis

< operator > infile outfile

Description
This module multiplies or divides each timestep of a dataset with the corresponding days per month
or days per year. The result of these functions depends on the used calendar of the input data.

Operators

muldpm Multiply with days per month

o(t, x) = i(t, x) ∗ days_per_month
divdpm Divide by days per month
o(t, x) = i(t, x)/days_per_month
muldpy Multiply with days per year
o(t, x) = i(t, x) ∗ days_per_year
divdpy Divide by days per year
o(t, x) = i(t, x)/days_per_year

2.7.13. ARITHLAT - Arithmetic with latitude

Synopsis

< operator > infile outfile

Description
This module multiplies or divides each field element with the cosine of the latitude.

Operators

mulcoslat Multiply with the cosine of the latitude

o(t, x) = i(t, x) ∗ cos(latitude(x))
divcoslat Divide by cosine of the latitude
o(t, x) = i(t, x)/cos(latitude(x))

110
Reference manual Statistical values

2.8. Statistical values

This section contains modules to compute statistical values of datasets. In this program there is the dif-
ferent notion of "mean" and "average" to distinguish two different kinds of treatment of missing values.
While computing the mean, only the not missing values are considered to belong to the sample with the
side effect of a probably reduced sample size. Computing the average is just adding the sample members
and divide the result by the sample size. For example, the mean of 1, 2, miss and 3 is (1+2+3)/3 = 2,
whereas the average is (1+2+miss+3)/4 = miss/4 = miss. If there are no missing values in the sample,
the average and the mean are identical.
CDO is using the verification time to identify the time range for temporal statistics. The time bounds are
never used!

In this section the abbreviations as in the following table are used:

n
X
sum xi
i=1
n
mean resp. avg −1
X
n xi
x i=1
 −1
mean resp. avg Xn n
X
weighted by  wj  wi xi
{wi , i = 1, ..., n} j=1 i=1

n
Variance X
n−1 (xi − x)2
var i=1
n
X
var1 (n − 1)−1 (xi − x)2
i=1
 −1   −1 2
n n n n
var weighted by X X X X
wj  wi xi −  wj  w j xj 
  
{wi , i = 1, ..., n} j=1 i=1 j=1 j=1

v
Standard deviation u
u n
X
std tn−1 (xi − x)2
s i=1
v
u n
X
u
std1 t(n − 1)−1 (xi − x)2
i=1
v  2
u
u n −1  −1
n n n
std weighted by u X X X X
wj  wi xi −  wj  w j xj 
u  
{wi , i = 1, ..., n}
t
j=1 i=1 j=1 j=1

(
x n+1 if n is odd
median 1
2

2 x 2 + x 2 +1 if n is even
n n

111
Statistical values Reference manual

Pn
Skewness i=1 (xi − x)/n
skew s3
Pn
Kurtosis i=1 (xi − x)4 /n
kurt s4
Cumulative Ranked Z ∞
2
Probability Score [H(x1 ) − cdf ({x2 . . . xn })|r ] dr
−∞
crps

with cdf (X)|r being the cumulative distribution function of {xi , i = 2 . . . n} at r

and H(x) the Heavyside function jumping at x.

Here is a short overview of all operators in this section:

timcumsum Cumulative sum over all timesteps

consecsum Consecutive Sum

consects Consecutive Timesteps

varsmin Variables minimum

varsmax Variables maximum
varsrange Variables range
varssum Variables sum
varsmean Variables mean
varsavg Variables average
varsstd Variables standard deviation
varsstd1 Variables standard deviation (n-1)
varsvar Variables variance
varsvar1 Variables variance (n-1)

ensmin Ensemble minimum

ensmax Ensemble maximum
ensrange Ensemble range
enssum Ensemble sum
ensmean Ensemble mean
ensavg Ensemble average
ensstd Ensemble standard deviation
ensstd1 Ensemble standard deviation (n-1)
ensvar Ensemble variance
ensvar1 Ensemble variance (n-1)
ensskew Ensemble skewness
enskurt Ensemble kurtosis
ensmedian Ensemble median
enspctl Ensemble percentiles

ensrkhistspace Ranked Histogram averaged over time

ensrkhisttime Ranked Histogram averaged over space
ensroc Ensemble Receiver Operating characteristics

enscrps Ensemble CRPS and decomposition

ensbrs Ensemble Brier score

112
Reference manual Statistical values

fldmin Field minimum

fldmax Field maximum
fldrange Field range
fldsum Field sum
fldint Field integral
fldmean Field mean
fldavg Field average
fldstd Field standard deviation
fldstd1 Field standard deviation (n-1)
fldvar Field variance
fldvar1 Field variance (n-1)
fldskew Field skewness
fldkurt Field kurtosis
fldmedian Field median
fldcount Field count
fldpctl Field percentiles

zonmin Zonal minimum

zonmax Zonal maximum
zonrange Zonal range
zonsum Zonal sum
zonmean Zonal mean
zonavg Zonal average
zonstd Zonal standard deviation
zonstd1 Zonal standard deviation (n-1)
zonvar Zonal variance
zonvar1 Zonal variance (n-1)
zonskew Zonal skewness
zonkurt Zonal kurtosis
zonmedian Zonal median
zonpctl Zonal percentiles

mermin Meridional minimum

mermax Meridional maximum
merrange Meridional range
mersum Meridional sum
mermean Meridional mean
meravg Meridional average
merstd Meridional standard deviation
merstd1 Meridional standard deviation (n-1)
mervar Meridional variance
mervar1 Meridional variance (n-1)
merskew Meridional skewness
merkurt Meridional kurtosis
mermedian Meridional median
merpctl Meridional percentiles

113
Statistical values Reference manual

gridboxmin Gridbox minimum

gridboxmax Gridbox maximum
gridboxrange Gridbox range
gridboxsum Gridbox sum
gridboxmean Gridbox mean
gridboxavg Gridbox average
gridboxstd Gridbox standard deviation
gridboxstd1 Gridbox standard deviation (n-1)
gridboxvar Gridbox variance
gridboxvar1 Gridbox variance (n-1)
gridboxskew Gridbox skewness
gridboxkurt Gridbox kurtosis
gridboxmedian Gridbox median

remapmin Remap minimum

remapmax Remap maximum
remaprange Remap range
remapsum Remap sum
remapmean Remap mean
remapavg Remap average
remapstd Remap standard deviation
remapstd1 Remap standard deviation (n-1)
remapvar Remap variance
remapvar1 Remap variance (n-1)
remapskew Remap skewness
remapkurt Remap kurtosis
remapmedian Remap median

vertmin Vertical minimum

vertmax Vertical maximum
vertrange Vertical range
vertsum Vertical sum
vertmean Vertical mean
vertavg Vertical average
vertstd Vertical standard deviation
vertstd1 Vertical standard deviation (n-1)
vertvar Vertical variance
vertvar1 Vertical variance (n-1)

timselmin Time selection minimum

timselmax Time selection maximum
timselrange Time selection range
timselsum Time selection sum
timselmean Time selection mean
timselavg Time selection average
timselstd Time selection standard deviation
timselstd1 Time selection standard deviation (n-1)
timselvar Time selection variance
timselvar1 Time selection variance (n-1)

timselpctl Time range percentiles

114
Reference manual Statistical values

runmin Running minimum

runmax Running maximum
runrange Running range
runsum Running sum
runmean Running mean
runavg Running average
runstd Running standard deviation
runstd1 Running standard deviation (n-1)
runvar Running variance
runvar1 Running variance (n-1)

runpctl Running percentiles

timmin Time minimum

timmax Time maximum
timrange Time range
timsum Time sum
timmean Time mean
timavg Time average
timstd Time standard deviation
timstd1 Time standard deviation (n-1)
timvar Time variance
timvar1 Time variance (n-1)

timpctl Time percentiles

hourmin Hourly minimum

hourmax Hourly maximum
hourrange Hourly range
hoursum Hourly sum
hourmean Hourly mean
houravg Hourly average
hourstd Hourly standard deviation
hourstd1 Hourly standard deviation (n-1)
hourvar Hourly variance
hourvar1 Hourly variance (n-1)

hourpctl Hourly percentiles

daymin Daily minimum

daymax Daily maximum
dayrange Daily range
daysum Daily sum
daymean Daily mean
dayavg Daily average
daystd Daily standard deviation
daystd1 Daily standard deviation (n-1)
dayvar Daily variance
dayvar1 Daily variance (n-1)

daypctl Daily percentiles

115
Statistical values Reference manual

monmin Monthly minimum

monmax Monthly maximum
monrange Monthly range
monsum Monthly sum
monmean Monthly mean
monavg Monthly average
monstd Monthly standard deviation
monstd1 Monthly standard deviation (n-1)
monvar Monthly variance
monvar1 Monthly variance (n-1)

monpctl Monthly percentiles

yearmonmean Yearly mean from monthly data

yearmin Yearly minimum

yearmax Yearly maximum
yearminidx Yearly minimum indices
yearmaxidx Yearly maximum indices
yearrange Yearly range
yearsum Yearly sum
yearmean Yearly mean
yearavg Yearly average
yearstd Yearly standard deviation
yearstd1 Yearly standard deviation (n-1)
yearvar Yearly variance
yearvar1 Yearly variance (n-1)

yearpctl Yearly percentiles

seasmin Seasonal minimum

seasmax Seasonal maximum
seasrange Seasonal range
seassum Seasonal sum
seasmean Seasonal mean
seasavg Seasonal average
seasstd Seasonal standard deviation
seasstd1 Seasonal standard deviation (n-1)
seasvar Seasonal variance
seasvar1 Seasonal variance (n-1)

seaspctl Seasonal percentiles

yhourmin Multi-year hourly minimum

yhourmax Multi-year hourly maximum
yhourrange Multi-year hourly range
yhoursum Multi-year hourly sum
yhourmean Multi-year hourly mean
yhouravg Multi-year hourly average
yhourstd Multi-year hourly standard deviation
yhourstd1 Multi-year hourly standard deviation (n-1)
yhourvar Multi-year hourly variance
yhourvar1 Multi-year hourly variance (n-1)

116
Reference manual Statistical values

dhourmin Multi-day hourly minimum

dhourmax Multi-day hourly maximum
dhourrange Multi-day hourly range
dhoursum Multi-day hourly sum
dhourmean Multi-day hourly mean
dhouravg Multi-day hourly average
dhourstd Multi-day hourly standard deviation
dhourstd1 Multi-day hourly standard deviation (n-1)
dhourvar Multi-day hourly variance
dhourvar1 Multi-day hourly variance (n-1)

ydaymin Multi-year daily minimum

ydaymax Multi-year daily maximum
ydayrange Multi-year daily range
ydaysum Multi-year daily sum
ydaymean Multi-year daily mean
ydayavg Multi-year daily average
ydaystd Multi-year daily standard deviation
ydaystd1 Multi-year daily standard deviation (n-1)
ydayvar Multi-year daily variance
ydayvar1 Multi-year daily variance (n-1)

ydaypctl Multi-year daily percentiles

ymonmin Multi-year monthly minimum

ymonmax Multi-year monthly maximum
ymonrange Multi-year monthly range
ymonsum Multi-year monthly sum
ymonmean Multi-year monthly mean
ymonavg Multi-year monthly average
ymonstd Multi-year monthly standard deviation
ymonstd1 Multi-year monthly standard deviation (n-1)
ymonvar Multi-year monthly variance
ymonvar1 Multi-year monthly variance (n-1)

ymonpctl Multi-year monthly percentiles

yseasmin Multi-year seasonal minimum

yseasmax Multi-year seasonal maximum
yseasrange Multi-year seasonal range
yseassum Multi-year seasonal sum
yseasmean Multi-year seasonal mean
yseasavg Multi-year seasonal average
yseasstd Multi-year seasonal standard deviation
yseasstd1 Multi-year seasonal standard deviation (n-1)
yseasvar Multi-year seasonal variance
yseasvar1 Multi-year seasonal variance (n-1)

yseaspctl Multi-year seasonal percentiles

117
Statistical values Reference manual

ydrunmin Multi-year daily running minimum

ydrunmax Multi-year daily running maximum
ydrunsum Multi-year daily running sum
ydrunmean Multi-year daily running mean
ydrunavg Multi-year daily running average
ydrunstd Multi-year daily running standard deviation
ydrunstd1 Multi-year daily running standard deviation (n-1)
ydrunvar Multi-year daily running variance
ydrunvar1 Multi-year daily running variance (n-1)

ydrunpctl Multi-year daily running percentiles

118
Reference manual Statistical values

2.8.1. TIMCUMSUM - Cumulative sum over all timesteps

Synopsis

timcumsum infile outfile

Description
The timcumsum operator calculates the cumulative sum over all timesteps. Missing values are treated
as numeric zero when summing.

o(t, x) = sum{i(t′ , x), 0 < t′ ≤ t}

2.8.2. CONSECSTAT - Consecute timestep periods

Synopsis

< operator > infile outfile

Description
This module computes periods over all timesteps in infile where a certain property is valid. The
property can be chosen by creating a mask from the original data, which is the expected input format
for operators of this module. Depending on the operator full information about each period or just
its length and ending date are computed.

Operators

consecsum Consecutive Sum

This operator computes periods of consecutive timesteps similar to a runsum, but
periods are finished, when the mask value is 0. That way multiple periods can be
found. Timesteps from the input are preserved. Missing values are handled like 0,
i.e. finish periods of consecutive timesteps.
consects Consecutive Timesteps
In contrast to the operator above consects only computes the length of each period
together with its last timestep. To be able to perform statistical analysis like min,
max or mean, everything else is set to missing value.

Example
For a given time series of daily temperatures, the periods of summer days can be calculated with
inplace maskting the input field:
cdo consects -gtc,20.0 infile1 outfile

119
Statistical values Reference manual

2.8.3. VARSSTAT - Statistical values over all variables

Synopsis

< operator > infile outfile

Description
This module computes statistical values over all variables for each timestep. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation is written to
outfile. All input variables need to have the same gridsize and the same number of levels.

Operators

varsmin Variables minimum

For every timestep the minimum over all variables is computed.
varsmax Variables maximum
For every timestep the maximum over all variables is computed.
varsrange Variables range
For every timestep the range over all variables is computed.
varssum Variables sum
For every timestep the sum over all variables is computed.
varsmean Variables mean
For every timestep the mean over all variables is computed.
varsavg Variables average
For every timestep the average over all variables is computed.
varsstd Variables standard deviation
For every timestep the standard deviation over all variables is computed. Normalize
by n.
varsstd1 Variables standard deviation (n-1)
For every timestep the standard deviation over all variables is computed. Normalize
by (n-1).
varsvar Variables variance
For every timestep the variance over all variables is computed. Normalize by n.
varsvar1 Variables variance (n-1)
For every timestep the variance over all variables is computed. Normalize by (n-1).

120
Reference manual Statistical values

2.8.4. ENSSTAT - Statistical values over an ensemble

Synopsis
< operator > infiles outfile
enspctl,p infiles outfile

Description
This module computes statistical values over an ensemble of input files. Depending on the chosen
operator, the minimum, maximum, range, sum, average, standard deviation, variance, skewness,
kurtosis, median or a certain percentile over all input files is written to outfile. All input files need
to have the same structure with the same variables. The date information of a timestep in outfile
is the date of the first input file.

Operators
ensmin Ensemble minimum
o(t, x) = min{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensmax Ensemble maximum
o(t, x) = max{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensrange Ensemble range
o(t, x) = range{i1 (t, x), i2 (t, x), · · · , in (t, x)}
enssum Ensemble sum
o(t, x) = sum{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensmean Ensemble mean
o(t, x) = mean{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensavg Ensemble average
o(t, x) = avg{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensstd Ensemble standard deviation
Normalize by n.
o(t, x) = std{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensstd1 Ensemble standard deviation (n-1)
Normalize by (n-1).
o(t, x) = std1{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensvar Ensemble variance
Normalize by n.
o(t, x) = var{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensvar1 Ensemble variance (n-1)
Normalize by (n-1).
o(t, x) = var1{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensskew Ensemble skewness
o(t, x) = skew{i1 (t, x), i2 (t, x), · · · , in (t, x)}
enskurt Ensemble kurtosis
o(t, x) = kurt{i1 (t, x), i2 (t, x), · · · , in (t, x)}
ensmedian Ensemble median
o(t, x) = median{i1 (t, x), i2 (t, x), · · · , in (t, x)}
enspctl Ensemble percentiles
o(t, x) = pth percentile{i1 (t, x), i2 (t, x), · · · , in (t, x)}

121
Statistical values Reference manual

Parameter
p FLOAT Percentile number in 0, ..., 100

Note
Operators of this module need to open all input files simultaneously. The maximum number of open
files depends on the operating system!

Example
To compute the ensemble mean over 6 input files use:
cdo ensmean infile1 infile2 infile3 infile4 infile5 infile6 outfile

Or shorter with filename substitution:

cdo ensmean infile[1-6] outfile

To compute the 50th percentile (median) over 6 input files use:

cdo enspctl,50 infile1 infile2 infile3 infile4 infile5 infile6 outfile

122
Reference manual Statistical values

2.8.5. ENSSTAT2 - Statistical values over an ensemble

Synopsis

< operator > obsfile ensfiles outfile

Description
This module computes statistical values over the ensemble of ensfiles using obsfile as a reference.
Depending on the operator a ranked Histogram or a roc-curve over all Ensembles ensfiles with
reference to obsfile is written to outfile. The date and grid information of a timestep in outfile
is the date of the first input file. Thus all input files are required to have the same structure in terms
of the gridsize, variable definitions and number of timesteps.
All Operators in this module use obsfile as the reference (for instance an observation) whereas
ensfiles are understood as an ensemble consisting of n (where n is the number of ensfiles) mem-
bers.
The operators ensrkhistspace and ensrkhisttime compute Ranked Histograms. Therefor the vertical
axis is utilized as the Histogram axis, which prohibits the use of files containing more than one level.
The histogram axis has nensfiles+1 bins with level 0 containing for each grid point the number of
observations being smaller as all ensembles and level nensfiles+1 indicating the number of observations
being larger than all ensembles.
ensrkhistspace computes a ranked histogram at each timestep reducing each horizontal grid to a 1x1
grid and keeping the time axis as in obsfile. Contrary ensrkhistspace computes a histogram at
each grid point keeping the horizontal grid for each variable and reducing the time-axis. The time
information is that from the last timestep in obsfile.

Operators

ensrkhistspace Ranked Histogram averaged over time

ensrkhisttime Ranked Histogram averaged over space

ensroc Ensemble Receiver Operating characteristics

Example
To compute a rank histogram over 5 input files ensfile1-ensfile5 given an observation in obsfile
use:
cdo ensrkhisttime obsfile ensfile1 ensfile2 ensfile3 ensfile4 ensfile5 outfile

Or shorter with filename substitution:

cdo ensrkhisttime obsfile ensfile[1-5] outfile

123
Statistical values Reference manual

2.8.6. ENSVAL - Ensemble validation tools

Synopsis

enscrps rfile infiles outfilebase

ensbrs,x rfile infiles outfilebase

Description
This module computes ensemble validation scores and their decomposition such as the Brier and
cumulative ranked probability score (CRPS). The first file is used as a reference it can be a clima-
tology, observation or reanalysis against which the skill of the ensembles given in infiles is measured.
Depending on the operator a number of output files is generated each containing the skill score and
its decomposition corresponding to the operator. The output is averaged over horizontal fields using
appropriate weights for each level and timestep in rfile.
All input files need to have the same structure with the same variables. The date information of a
timestep in outfile is the date of the first input file. The output files are named as <outfilebase>.<type>.<files
where <type> depends on the operator and <filesuffix> is determined from the output file type.
There are three output files for operator enscrps and four output files for operator ensbrs.
The CRPS and its decomposition into Reliability and the potential CRPS are calculated by an
appropriate averaging over the field members (note, that the CRPS does *not* average linearly). In
the three output files <type> has the following meaning: crps for the CRPS, reli for the reliability
and crpspot for the potential crps. The relation CRP S = CRP Spot + RELI
holds.
The Brier score of the Ensemble given by infiles with respect to the reference given in rfile and
the threshold x is calculated. In the four output files <type> has the following meaning: brs for the
Brier score wrt threshold x; brsreli for the Brier score reliability wrt threshold x; brsreso for the
Brier score resolution wrt threshold x; brsunct for the Brier score uncertainty wrt threshold x. In
analogy to the CRPS the following relation holds: BRS(x) = RELI(x) − RESO(x) + U N CT (x).
The implementation of the decomposition of the CRPS and Brier Score follows Hans Hersbach (2000):
Decomposition of the Continuous Ranked Probability Score for Ensemble Prediction Systems, in:
Weather and Forecasting (15) pp. 559-570.
The CRPS code decomposition has been verified against the CRAN - ensemble validation package
from R. Differences occur when grid-cell area is not uniform as the implementation in R does not
account for that.

Operators

enscrps Ensemble CRPS and decomposition

ensbrs Ensemble Brier score

Ensemble Brier Score and Decomposition

Example
To compute the field averaged Brier score at x=5 over an ensemble with 5 members ensfile1-5
w.r.t. the reference rfile and write the results to files obase.brs.<suff>, obase.brsreli<suff>,
obase.brsreso<suff>, obase.brsunct<suff> where <suff> is determined from the output file
type, use
cdo ensbrs,5 rfile ensfile1 ensfile2 ensfile3 ensfile4 ensfile5 obase

124
Reference manual Statistical values

or shorter using file name substitution:

cdo ensbrs,5 rfile ensfile[1-5] obase

125
Statistical values Reference manual

2.8.7. FLDSTAT - Statistical values over a field

Synopsis

< operator > infile outfile

fldint,weights infile outfile
fldmean,weights infile outfile
fldavg,weights infile outfile
fldstd,weights infile outfile
fldstd1,weights infile outfile
fldvar,weights infile outfile
fldvar1,weights infile outfile
fldpctl,p infile outfile

Description
This module computes statistical values of all input fields. A field is a horizontal layer of a data
variable. Depending on the chosen operator, the minimum, maximum, range, sum, integral, average,
standard deviation, variance, skewness, kurtosis, median or a certain percentile of the field is written
to outfile.

Operators

fldmin Field minimum

For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = min{i(t, x′ ), x1 < x′ ≤ xn }
fldmax Field maximum
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = max{i(t, x′ ), x1 < x′ ≤ xn }
fldrange Field range
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = range{i(t, x′ ), x1 < x′ ≤ xn }
fldsum Field sum
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = sum{i(t, x′ ), x1 < x′ ≤ xn }
fldint Field integral
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = sum{i(t, x′ ) ∗ cellarea(x′ ), x1 < x′ ≤ xn }
fldmean Field mean
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = mean{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.
fldavg Field average
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = avg{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.

126
Reference manual Statistical values

fldstd Field standard deviation

Normalize by n. For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = std{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.
fldstd1 Field standard deviation (n-1)
Normalize by (n-1). For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = std1{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.
fldvar Field variance
Normalize by n. For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = var{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.
fldvar1 Field variance (n-1)
Normalize by (n-1). For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = var1{i(t, x′ ), x1 < x′ ≤ xn }
weighted by area weights obtained by the input field.
fldskew Field skewness
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = skew{i(t, x′ ), x1 < x′ ≤ xn }
fldkurt Field kurtosis
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = kurt{i(t, x′ ), x1 < x′ ≤ xn }
fldmedian Field median
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = median{i(t, x′ ), x1 < x′ ≤ xn }
fldcount Field count
Number of non-missing values of the field.
fldpctl Field percentiles
For every gridpoint x_1, ..., x_n of the same field it is:
o(t, 1) = pth percentile{i(t, x′ ), x1 < x′ ≤ xn }

Parameter
weights BOOL weights=FALSE disables weighting by grid cell area [default: weights=TRUE]
p FLOAT Percentile number in 0, ..., 100

Example
To compute the field mean of all input fields use:
cdo fldmean infile outfile

To compute the 90th percentile of all input fields use:

cdo fldpctl,90 infile outfile

127
Statistical values Reference manual

2.8.8. ZONSTAT - Zonal statistical values

Synopsis

< operator > infile outfile

zonmean[,zonaldes] infile outfile
zonpctl,p infile outfile

Description
This module computes zonal statistical values of the input fields. Depending on the chosen operator,
the zonal minimum, maximum, range, sum, average, standard deviation, variance, skewness, kurtosis,
median or a certain percentile of the field is written to outfile. Operators of this module require
all variables on the same regular lon/lat grid. Only the zonal mean (zonmean) can be calculated for
data on an unstructured grid if the latitude bins are defined with the optional parameter zonaldes.

Operators

zonmin Zonal minimum

For every latitude the minimum over all longitudes is computed.
zonmax Zonal maximum
For every latitude the maximum over all longitudes is computed.
zonrange Zonal range
For every latitude the range over all longitudes is computed.
zonsum Zonal sum
For every latitude the sum over all longitudes is computed.
zonmean Zonal mean
For every latitude the mean over all longitudes is computed. Use the optional pa-
rameter zonaldes for data on an unstructured grid.
zonavg Zonal average
For every latitude the average over all longitudes is computed.
zonstd Zonal standard deviation
For every latitude the standard deviation over all longitudes is computed. Normalize
by n.
zonstd1 Zonal standard deviation (n-1)
For every latitude the standard deviation over all longitudes is computed. Normalize
by (n-1).
zonvar Zonal variance
For every latitude the variance over all longitudes is computed. Normalize by n.
zonvar1 Zonal variance (n-1)
For every latitude the variance over all longitudes is computed. Normalize by (n-1).
zonskew Zonal skewness
For every latitude the skewness over all longitudes is computed.
zonkurt Zonal kurtosis
For every latitude the kurtosis over all longitudes is computed.
zonmedian Zonal median
For every latitude the median over all longitudes is computed.
zonpctl Zonal percentiles
For every latitude the pth percentile over all longitudes is computed.

128
Reference manual Statistical values

Parameter
p FLOAT Percentile number in 0, ..., 100
zonaldes STRING Description of the zonal latitude bins needed for data on an unstructured
grid. A predefined zonal description is zonal_<DY>. DY is the increment of the lati-
tudes in degrees.

Example
To compute the zonal mean of all input fields use:
cdo zonmean infile outfile

To compute the 50th meridional percentile (median) of all input fields use:
cdo zonpctl,50 infile outfile

129
Statistical values Reference manual

2.8.9. MERSTAT - Meridional statistical values

Synopsis
< operator > infile outfile
merpctl,p infile outfile

Description
This module computes meridional statistical values of the input fields. Depending on the chosen
operator, the meridional minimum, maximum, range, sum, average, standard deviation, variance,
skewness, kurtosis, median or a certain percentile of the field is written to outfile. Operators of this
module require all variables on the same regular lon/lat grid.

Operators
mermin Meridional minimum
For every longitude the minimum over all latitudes is computed.
mermax Meridional maximum
For every longitude the maximum over all latitudes is computed.
merrange Meridional range
For every longitude the range over all latitudes is computed.
mersum Meridional sum
For every longitude the sum over all latitudes is computed.
mermean Meridional mean
For every longitude the area weighted mean over all latitudes is computed.
meravg Meridional average
For every longitude the area weighted average over all latitudes is computed.
merstd Meridional standard deviation
For every longitude the standard deviation over all latitudes is computed. Normalize
by n.
merstd1 Meridional standard deviation (n-1)
For every longitude the standard deviation over all latitudes is computed. Normalize
by (n-1).
mervar Meridional variance
For every longitude the variance over all latitudes is computed. Normalize by n.
mervar1 Meridional variance (n-1)
For every longitude the variance over all latitudes is computed. Normalize by (n-1).
merskew Meridional skewness
For every longitude the skewness over all latitudes is computed.
merkurt Meridional kurtosis
For every longitude the kurtosis over all latitudes is computed.
mermedian Meridional median
For every longitude the median over all latitudes is computed.
merpctl Meridional percentiles
For every longitude the pth percentile over all latitudes is computed.

Parameter
p FLOAT Percentile number in 0, ..., 100

130
Reference manual Statistical values

Example
To compute the meridional mean of all input fields use:
cdo mermean infile outfile

To compute the 50th meridional percentile (median) of all input fields use:
cdo merpctl,50 infile outfile

131
Statistical values Reference manual

2.8.10. GRIDBOXSTAT - Statistical values over grid boxes

Synopsis
< operator >,nx,ny infile outfile

Description
This module computes statistical values over surrounding grid boxes. Depending on the chosen
operator, the minimum, maximum, range, sum, average, standard deviation, variance, skewness,
kurtosis or median of the neighboring grid boxes is written to outfile. All gridbox operators only
work on quadrilateral curvilinear grids.

Operators
gridboxmin Gridbox minimum
Minimum value of the selected grid boxes.
gridboxmax Gridbox maximum
Maximum value of the selected grid boxes.
gridboxrange Gridbox range
Range (max-min value) of the selected grid boxes.
gridboxsum Gridbox sum
Sum of the selected grid boxes.
gridboxmean Gridbox mean
Mean of the selected grid boxes.
gridboxavg Gridbox average
Average of the selected grid boxes.
gridboxstd Gridbox standard deviation
Standard deviation of the selected grid boxes. Normalize by n.
gridboxstd1 Gridbox standard deviation (n-1)
Standard deviation of the selected grid boxes. Normalize by (n-1).
gridboxvar Gridbox variance
Variance of the selected grid boxes. Normalize by n.
gridboxvar1 Gridbox variance (n-1)
Variance of the selected grid boxes. Normalize by (n-1).
gridboxskew Gridbox skewness
Skewness of the selected grid boxes.
gridboxkurt Gridbox kurtosis
Kurtosis of the selected grid boxes.
gridboxmedian Gridbox median
Median of the selected grid boxes.

Parameter
nx INTEGER Number of grid boxes in x direction
ny INTEGER Number of grid boxes in y direction

Example
To compute the mean over 10x10 grid boxes of the input field use:
cdo gridboxmean,10,10 infile outfile

132
Reference manual Statistical values

2.8.11. REMAPSTAT - Remaps source points to target cells

Synopsis

< operator >,grid infile outfile

Description
This module maps source points to target cells by calculating a statistical value from the source
points. Each target cell contains the statistical value from all source points within that target cell.
If there are no source points within a target cell, it gets a missing value. The target grid must be
regular lon/lat or Gaussian. Depending on the chosen operator the minimum, maximum, range, sum,
average, variance, standard deviation, skewness, kurtosis or median of source points is computed.

Operators

remapmin Remap minimum

Minimum value of the source points.
remapmax Remap maximum
Maximum value of the source points.
remaprange Remap range
Range (max-min value) of the source points.
remapsum Remap sum
Sum of the source points.
remapmean Remap mean
Mean of the source points.
remapavg Remap average
Average of the source points.
remapstd Remap standard deviation
Standard deviation of the source points. Normalize by n.
remapstd1 Remap standard deviation (n-1)
Standard deviation of the source points. Normalize by (n-1).
remapvar Remap variance
Variance of the source points. Normalize by n.
remapvar1 Remap variance (n-1)
Variance of the source points. Normalize by (n-1).
remapskew Remap skewness
Skewness of the source points.
remapkurt Remap kurtosis
Kurtosis of the source points.
remapmedian Remap median
Median of the source points.

Parameter
grid STRING Target grid description file or name

133
Statistical values Reference manual

Example
To compute the mean over source points within the taget cells, use:
cdo remapmean,<targetgrid> infile outfile

If some of the target cells contain missing values, use the Operator setmisstonn to fill these missing
values with the nearest neighbor cell:
cdo setmisstonn -remapmean,<targetgrid> infile outfile

134
Reference manual Statistical values

2.8.12. VERTSTAT - Vertical statistical values

Synopsis

< operator >,weights infile outfile

Description
This module computes statistical values over all levels of the input variables. According to chosen
operator the vertical minimum, maximum, range, sum, average, variance or standard deviation is
written to outfile.

Operators

vertmin Vertical minimum

For every gridpoint the minimum over all levels is computed.
vertmax Vertical maximum
For every gridpoint the maximum over all levels is computed.
vertrange Vertical range
For every gridpoint the range over all levels is computed.
vertsum Vertical sum
For every gridpoint the sum over all levels is computed.
vertmean Vertical mean
For every gridpoint the layer weighted mean over all levels is computed.
vertavg Vertical average
For every gridpoint the layer weighted average over all levels is computed.
vertstd Vertical standard deviation
For every gridpoint the standard deviation over all levels is computed. Normalize by
n.
vertstd1 Vertical standard deviation (n-1)
For every gridpoint the standard deviation over all levels is computed. Normalize by
(n-1).
vertvar Vertical variance
For every gridpoint the variance over all levels is computed. Normalize by n.
vertvar1 Vertical variance (n-1)
For every gridpoint the variance over all levels is computed. Normalize by (n-1).

Parameter
weights BOOL weights=FALSE disables weighting by layer thickness [default: weights=TRUE]

Example
To compute the vertical sum of all input variables use:
cdo vertsum infile outfile

135
Statistical values Reference manual

2.8.13. TIMSELSTAT - Time range statistical values

Synopsis

< operator >,nsets[,noffset[,nskip]] infile outfile

Description
This module computes statistical values for a selected number of timesteps. According to the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of the selected
timesteps is written to outfile. The time of outfile is determined by the time in the middle of
all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

timselmin Time selection minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
timselmax Time selection maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
timselrange Time selection range
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
timselsum Time selection sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
timselmean Time selection mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
timselavg Time selection average
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time
range it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
timselstd Time selection standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
selected time range it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
timselstd1 Time selection standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the
same selected time range it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
timselvar Time selection variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
selected time range it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }

136
Reference manual Statistical values

timselvar1 Time selection variance (n-1)

Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the
same selected time range it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

Parameter
nsets INTEGER Number of input timesteps for each output timestep
noffset INTEGER Number of input timesteps skipped before the first timestep range (optional)
nskip INTEGER Number of input timesteps skipped between timestep ranges (optional)

Example
Assume an input dataset has monthly means over several years. To compute seasonal means from
monthly means the first two month have to be skipped:
cdo timselmean,3,2 infile outfile

2.8.14. TIMSELPCTL - Time range percentile values

Synopsis

timselpctl,p,nsets[,noffset[,nskip]] infile1 infile2 infile3 outfile

Description
This operator computes percentile values over a selected number of timesteps in infile1. The
algorithm uses histograms with minimum and maximum bounds given in infile2 and infile3,
respectively. The default number of histogram bins is 101. The default can be overridden by setting
the environment variable CDO_PCTL_NBINS to a different value. The files infile2 and infile3 should
be the result of corresponding timselmin and timselmax operations, respectively. The time of outfile
is determined by the time in the middle of all contributing timesteps of infile1. This can be change
with the CDO option --timestat_date <first|middle|last>.
For every adjacent sequence t_1, ..., t_n of timesteps of the same selected time range it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100
nsets INTEGER Number of input timesteps for each output timestep
noffset INTEGER Number of input timesteps skipped before the first timestep range (optional)
nskip INTEGER Number of input timesteps skipped between timestep ranges (optional)

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

137
Statistical values Reference manual

2.8.15. RUNSTAT - Running statistical values

Synopsis

< operator >,nts infile outfile

Description
This module computes running statistical values over a selected number of timesteps. Depending on
the chosen operator the minimum, maximum, range, sum, average, variance or standard deviation
of a selected number of consecutive timesteps read from infile is written to outfile. The time of
outfile is determined by the time in the middle of all contributing timesteps of infile. This can
be change with the CDO option --timestat_date <first|middle|last>.

Operators

runmin Running minimum

o(t + (nts − 1)/2, x) = min{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runmax Running maximum
o(t + (nts − 1)/2, x) = max{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runrange Running range
o(t + (nts − 1)/2, x) = range{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runsum Running sum
o(t + (nts − 1)/2, x) = sum{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runmean Running mean
o(t + (nts − 1)/2, x) = mean{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runavg Running average
o(t + (nts − 1)/2, x) = avg{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runstd Running standard deviation
Normalize by n.
o(t + (nts − 1)/2, x) = std{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runstd1 Running standard deviation (n-1)
Normalize by (n-1).
o(t + (nts − 1)/2, x) = std1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runvar Running variance
Normalize by n.
o(t + (nts − 1)/2, x) = var{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}
runvar1 Running variance (n-1)
Normalize by (n-1).
o(t + (nts − 1)/2, x) = var1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}

Parameter
nts INTEGER Number of timesteps

Environment
CDO_TIMESTAT_DATE Sets the time stamp in outfile to the "first", "middle" or "last" contributing
timestep of infile.

138
Reference manual Statistical values

Example
To compute the running mean over 9 timesteps use:
cdo runmean,9 infile outfile

2.8.16. RUNPCTL - Running percentile values

Synopsis

runpctl,p,nts infile outfile

Description
This module computes running percentiles over a selected number of timesteps in infile. The time
of outfile is determined by the time in the middle of all contributing timesteps of infile. This can
be change with the CDO option --timestat_date <first|middle|last>.

o(t + (nts − 1)/2, x) = pth percentile{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x)}

Parameter
p FLOAT Percentile number in 0, ..., 100
nts INTEGER Number of timesteps

Example
To compute the running 50th percentile (median) over 9 timesteps use:
cdo runpctl,50,9 infile outfile

139
Statistical values Reference manual

2.8.17. TIMSTAT - Statistical values over all timesteps

Synopsis

< operator > infile outfile

Description
This module computes statistical values over all timesteps in infile. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of all timesteps
read from infile is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

timmin Time minimum

o(1, x) = min{i(t′ , x), t1 < t′ ≤ tn }
timmax Time maximum
o(1, x) = max{i(t′ , x), t1 < t′ ≤ tn }
timrange Time range
o(1, x) = range{i(t′ , x), t1 < t′ ≤ tn }
timsum Time sum
o(1, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
timmean Time mean
o(1, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
timavg Time average
o(1, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
timstd Time standard deviation
Normalize by n.
o(1, x) = std{i(t′ , x), t1 < t′ ≤ tn }
timstd1 Time standard deviation (n-1)
Normalize by (n-1).
o(1, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
timvar Time variance
Normalize by n.
o(1, x) = var{i(t′ , x), t1 < t′ ≤ tn }
timvar1 Time variance (n-1)
Normalize by (n-1).
o(1, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

Example
To compute the mean over all input timesteps use:
cdo timmean infile outfile

140
Reference manual Statistical values

2.8.18. TIMPCTL - Percentile values over all timesteps

Synopsis

timpctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps in infile1. The algorithm uses histograms
with minimum and maximum bounds given in infile2 and infile3, respectively. The default
number of histogram bins is 101. The default can be overridden by defining the environment variable
CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding timmin and
timmax operations, respectively. The time of outfile is determined by the time in the middle of
all contributing timesteps of infile1. This can be change with the CDO option --timestat_date
<first|middle|last>.

o(1, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the 90th percentile over all input timesteps use:
cdo timmin infile minfile
cdo timmax infile maxfile
cdo timpctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo timpctl,90 infile -timmin infile -timmax infile outfile

141
Statistical values Reference manual

2.8.19. HOURSTAT - Hourly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values over timesteps of the same hour. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of timesteps
of the same hour is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

hourmin Hourly minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
hourmax Hourly maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
hourrange Hourly range
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
hoursum Hourly sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
hourmean Hourly mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
houravg Hourly average
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
hourstd Hourly standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
hour it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
hourstd1 Hourly standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
hour it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
hourvar Hourly variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
hour it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }
hourvar1 Hourly variance (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
hour it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

142
Reference manual Statistical values

Example
To compute the hourly mean of a time series use:
cdo hourmean infile outfile

2.8.20. HOURPCTL - Hourly percentile values

Synopsis

hourpctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps of the same hour in infile1. The algorithm uses
histograms with minimum and maximum bounds given in infile2 and infile3, respectively. The
default number of histogram bins is 101. The default can be overridden by defining the environment
variable CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding
hourmin and hourmax operations, respectively. The time of outfile is determined by the time
in the middle of all contributing timesteps of infile1. This can be change with the CDO option
--timestat_date <first|middle|last>.
For every adjacent sequence t_1, ..., t_n of timesteps of the same hour it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the hourly 90th percentile of a time series use:
cdo hourmin infile minfile
cdo hourmax infile maxfile
cdo hourpctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo hourpctl,90 infile -hourmin infile -hourmax infile outfile

143
Statistical values Reference manual

2.8.21. DAYSTAT - Daily statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values over timesteps of the same day. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of timesteps
of the same day is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

daymin Daily minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
daymax Daily maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
dayrange Daily range
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
daysum Daily sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
daymean Daily mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
dayavg Daily average
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
daystd Daily standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same day
it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
daystd1 Daily standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
day it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
dayvar Daily variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same day
it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }
dayvar1 Daily variance (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
day it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

144
Reference manual Statistical values

Example
To compute the daily mean of a time series use:
cdo daymean infile outfile

2.8.22. DAYPCTL - Daily percentile values

Synopsis

daypctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps of the same day in infile1. The algorithm uses
histograms with minimum and maximum bounds given in infile2 and infile3, respectively. The
default number of histogram bins is 101. The default can be overridden by defining the environment
variable CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding
daymin and daymax operations, respectively. The time of outfile is determined by the time in
the middle of all contributing timesteps of infile1. This can be change with the CDO option
--timestat_date <first|middle|last>.
For every adjacent sequence t_1, ..., t_n of timesteps of the same day it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the daily 90th percentile of a time series use:
cdo daymin infile minfile
cdo daymax infile maxfile
cdo daypctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo daypctl,90 infile -daymin infile -daymax infile outfile

145
Statistical values Reference manual

2.8.23. MONSTAT - Monthly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values over timesteps of the same month. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of timesteps of
the same month is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

monmin Monthly minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
monmax Monthly maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
monrange Monthly range
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
monsum Monthly sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
monmean Monthly mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
monavg Monthly average
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
monstd Monthly standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
month it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
monstd1 Monthly standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
month it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
monvar Monthly variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
month it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }
monvar1 Monthly variance (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
month it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

146
Reference manual Statistical values

Example
To compute the monthly mean of a time series use:
cdo monmean infile outfile

2.8.24. MONPCTL - Monthly percentile values

Synopsis

monpctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps of the same month in infile1. The algorithm
uses histograms with minimum and maximum bounds given in infile2 and infile3, respectively.
The default number of histogram bins is 101. The default can be overridden by defining the environ-
ment variable CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding
monmin and monmax operations, respectively. The time of outfile is determined by the time in
the middle of all contributing timesteps of infile1. This can be change with the CDO option
--timestat_date <first|middle|last>.
For every adjacent sequence t_1, ..., t_n of timesteps of the same month it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the monthly 90th percentile of a time series use:
cdo monmin infile minfile
cdo monmax infile maxfile
cdo monpctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo monpctl,90 infile -monmin infile -monmax infile outfile

147
Statistical values Reference manual

2.8.25. YEARMONSTAT - Yearly mean from monthly data

Synopsis

yearmonmean infile outfile

Description
This operator computes the yearly mean of a monthly time series. Each month is weighted with the
number of days per month. The time of outfile is determined by the time in the middle of all
contributing timesteps of infile.
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }

Environment
CDO_TIMESTAT_DATE Sets the date information in outfile to the "first", "middle" or "last" contribut-
ing timestep of infile.

Example
To compute the yearly mean of a monthly time series use:
cdo yearmonmean infile outfile

148
Reference manual Statistical values

2.8.26. YEARSTAT - Yearly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values over timesteps of the same year. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of timesteps
of the same year is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>.

Operators

yearmin Yearly minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
yearmax Yearly maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
yearminidx Yearly minimum indices
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = minidx{i(t′ , x), t1 < t′ ≤ tn }
yearmaxidx Yearly maximum indices
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = maxidx{i(t′ , x), t1 < t′ ≤ tn }
yearrange Yearly range
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
yearsum Yearly sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
yearmean Yearly mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
yearavg Yearly average
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
yearstd Yearly standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
year it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
yearstd1 Yearly standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the
same year it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
yearvar Yearly variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
year it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }

149
Statistical values Reference manual

yearvar1 Yearly variance (n-1)

Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the
same year it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

Note
The operators yearmean and yearavg compute only arithmetical means!

Example
To compute the yearly mean of a time series use:
cdo yearmean infile outfile

To compute the yearly mean from the correct weighted monthly mean use:
cdo yearmonmean infile outfile

2.8.27. YEARPCTL - Yearly percentile values

Synopsis
yearpctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps of the same year in infile1. The algorithm uses
histograms with minimum and maximum bounds given in infile2 and infile3, respectively. The
default number of histogram bins is 101. The default can be overridden by defining the environment
variable CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding
yearmin and yearmax operations, respectively. The time of outfile is determined by the time in
the middle of all contributing timesteps of infile1. This can be change with the CDO option
--timestat_date <first|middle|last>.
For every adjacent sequence t_1, ..., t_n of timesteps of the same year it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the yearly 90th percentile of a time series use:
cdo yearmin infile minfile
cdo yearmax infile maxfile
cdo yearpctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo yearpctl,90 infile -yearmin infile -yearmax infile outfile

150
Reference manual Statistical values

2.8.28. SEASSTAT - Seasonal statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values over timesteps of the same season. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of timesteps of
the same season is written to outfile. The time of outfile is determined by the time in the middle
of all contributing timesteps of infile. This can be change with the CDO option --timestat_date
<first|middle|last>. Be careful about the first and the last output timestep, they may be incorrect
values if the seasons have incomplete timesteps.

Operators

seasmin Seasonal minimum

For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = min{i(t′ , x), t1 < t′ ≤ tn }
seasmax Seasonal maximum
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = max{i(t′ , x), t1 < t′ ≤ tn }
seasrange Seasonal range
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = range{i(t′ , x), t1 < t′ ≤ tn }
seassum Seasonal sum
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = sum{i(t′ , x), t1 < t′ ≤ tn }
seasmean Seasonal mean
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = mean{i(t′ , x), t1 < t′ ≤ tn }
seasavg Seasonal average
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:
o(t, x) = avg{i(t′ , x), t1 < t′ ≤ tn }
seasstd Seasonal standard deviation
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
season it is:
o(t, x) = std{i(t′ , x), t1 < t′ ≤ tn }
seasstd1 Seasonal standard deviation (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
season it is:
o(t, x) = std1{i(t′ , x), t1 < t′ ≤ tn }
seasvar Seasonal variance
Normalize by n. For every adjacent sequence t_1, ..., t_n of timesteps of the same
season it is:
o(t, x) = var{i(t′ , x), t1 < t′ ≤ tn }
seasvar1 Seasonal variance (n-1)
Normalize by (n-1). For every adjacent sequence t_1, ..., t_n of timesteps of the same
season it is:
o(t, x) = var1{i(t′ , x), t1 < t′ ≤ tn }

151
Statistical values Reference manual

Example
To compute the seasonal mean of a time series use:
cdo seasmean infile outfile

2.8.29. SEASPCTL - Seasonal percentile values

Synopsis

seaspctl,p infile1 infile2 infile3 outfile

Description
This operator computes percentiles over all timesteps in infile1 of the same season. The algorithm
uses histograms with minimum and maximum bounds given in infile2 and infile3, respectively.
The default number of histogram bins is 101. The default can be overridden by defining the environ-
ment variable CDO_PCTL_NBINS. The files infile2 and infile3 should be the result of corresponding
seasmin and seasmax operations, respectively. The time of outfile is determined by the time in
the middle of all contributing timesteps of infile1. This can be change with the CDO option --
timestat_date <first|middle|last>. Be careful about the first and the last output timestep, they may
be incorrect values if the seasons have incomplete timesteps.
For every adjacent sequence t_1, ..., t_n of timesteps of the same season it is:

o(t, x) = pth percentile{i(t′ , x), t1 < t′ ≤ tn }

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the seasonal 90th percentile of a time series use:
cdo seasmin infile minfile
cdo seasmax infile maxfile
cdo seaspctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo seaspctl,90 infile -seasmin infile -seasmax infile outfile

152
Reference manual Statistical values

2.8.30. YHOURSTAT - Multi-year hourly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values of each hour and day of year. Depending on the chosen
operator the minimum, maximum, range, sum, average, variance or standard deviation of each hour
and day of year in infile is written to outfile. The date information in an output field is the date
of the last contributing input field.

Operators

yhourmin Multi-year hourly minimum

o(0001, x) = min{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = min{i(t, x), day(i(t)) = 8784}
yhourmax Multi-year hourly maximum
o(0001, x) = max{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = max{i(t, x), day(i(t)) = 8784}
yhourrange Multi-year hourly range
o(0001, x) = range{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = range{i(t, x), day(i(t)) = 8784}
yhoursum Multi-year hourly sum
o(0001, x) = sum{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = sum{i(t, x), day(i(t)) = 8784}
yhourmean Multi-year hourly mean
o(0001, x) = mean{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = mean{i(t, x), day(i(t)) = 8784}
yhouravg Multi-year hourly average
o(0001, x) = avg{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = avg{i(t, x), day(i(t)) = 8784}
yhourstd Multi-year hourly standard deviation
Normalize by n.
o(0001, x) = std{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = std{i(t, x), day(i(t)) = 8784}
yhourstd1 Multi-year hourly standard deviation (n-1)
Normalize by (n-1).

153
Statistical values Reference manual

o(0001, x) = std1{i(t, x), day(i(t)) = 0001}

..
.
o(8784, x) = std1{i(t, x), day(i(t)) = 8784}
yhourvar Multi-year hourly variance
Normalize by n.
o(0001, x) = var{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = var{i(t, x), day(i(t)) = 8784}
yhourvar1 Multi-year hourly variance (n-1)
Normalize by (n-1).
o(0001, x) = var1{i(t, x), day(i(t)) = 0001}
..
.
o(8784, x) = var1{i(t, x), day(i(t)) = 8784}

154
Reference manual Statistical values

2.8.31. DHOURSTAT - Multi-day hourly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values of each hour of day. Depending on the chosen operator
the minimum, maximum, range, sum, average, variance or standard deviation of each hour of day
in infile is written to outfile. The date information in an output field is the date of the last
contributing input field.

Operators

dhourmin Multi-day hourly minimum

o(01, x) = min{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = min{i(t, x), day(i(t)) = 24}
dhourmax Multi-day hourly maximum
o(01, x) = max{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = max{i(t, x), day(i(t)) = 24}
dhourrange Multi-day hourly range
o(01, x) = range{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = range{i(t, x), day(i(t)) = 24}
dhoursum Multi-day hourly sum
o(01, x) = sum{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = sum{i(t, x), day(i(t)) = 24}
dhourmean Multi-day hourly mean
o(01, x) = mean{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = mean{i(t, x), day(i(t)) = 24}
dhouravg Multi-day hourly average
o(01, x) = avg{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = avg{i(t, x), day(i(t)) = 24}
dhourstd Multi-day hourly standard deviation
Normalize by n.
o(01, x) = std{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = std{i(t, x), day(i(t)) = 24}
dhourstd1 Multi-day hourly standard deviation (n-1)
Normalize by (n-1).

155
Statistical values Reference manual

o(01, x) = std1{i(t, x), day(i(t)) = 01}

..
.
o(24, x) = std1{i(t, x), day(i(t)) = 24}
dhourvar Multi-day hourly variance
Normalize by n.
o(01, x) = var{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = var{i(t, x), day(i(t)) = 24}
dhourvar1 Multi-day hourly variance (n-1)
Normalize by (n-1).
o(01, x) = var1{i(t, x), day(i(t)) = 01}
..
.
o(24, x) = var1{i(t, x), day(i(t)) = 24}

156
Reference manual Statistical values

2.8.32. YDAYSTAT - Multi-year daily statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values of each day of year. Depending on the chosen operator
the minimum, maximum, range, sum, average, variance or standard deviation of each day of year
in infile is written to outfile. The date information in an output field is the date of the last
contributing input field.

Operators

ydaymin Multi-year daily minimum

o(001, x) = min{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = min{i(t, x), day(i(t)) = 366}
ydaymax Multi-year daily maximum
o(001, x) = max{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = max{i(t, x), day(i(t)) = 366}
ydayrange Multi-year daily range
o(001, x) = range{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = range{i(t, x), day(i(t)) = 366}
ydaysum Multi-year daily sum
o(001, x) = sum{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = sum{i(t, x), day(i(t)) = 366}
ydaymean Multi-year daily mean
o(001, x) = mean{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = mean{i(t, x), day(i(t)) = 366}
ydayavg Multi-year daily average
o(001, x) = avg{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = avg{i(t, x), day(i(t)) = 366}
ydaystd Multi-year daily standard deviation
Normalize by n.
o(001, x) = std{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = std{i(t, x), day(i(t)) = 366}
ydaystd1 Multi-year daily standard deviation (n-1)
Normalize by (n-1).

157
Statistical values Reference manual

o(001, x) = std1{i(t, x), day(i(t)) = 001}

..
.
o(366, x) = std1{i(t, x), day(i(t)) = 366}
ydayvar Multi-year daily variance
Normalize by n.
o(001, x) = var{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = var{i(t, x), day(i(t)) = 366}
ydayvar1 Multi-year daily variance (n-1)
Normalize by (n-1).
o(001, x) = var1{i(t, x), day(i(t)) = 001}
..
.
o(366, x) = var1{i(t, x), day(i(t)) = 366}

Example
To compute the daily mean over all input years use:
cdo ydaymean infile outfile

158
Reference manual Statistical values

2.8.33. YDAYPCTL - Multi-year daily percentile values

Synopsis

ydaypctl,p infile1 infile2 infile3 outfile

Description
This operator writes a certain percentile of each day of year in infile1 to outfile. The algorithm
uses histograms with minimum and maximum bounds given in infile2 and infile3, respectively.
The default number of histogram bins is 101. The default can be overridden by setting the environment
variable CDO_PCTL_NBINS to a different value. The files infile2 and infile3 should be the result
of corresponding ydaymin and ydaymax operations, respectively. The date information in an output
field is the date of the last contributing input field.

o(001, x) = pth percentile{i(t, x), day(i(t)) = 001}

..
.
o(366, x) = pth percentile{i(t, x), day(i(t)) = 366}

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the daily 90th percentile over all input years use:
cdo ydaymin infile minfile
cdo ydaymax infile maxfile
cdo ydaypctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo ydaypctl,90 infile -ydaymin infile -ydaymax infile outfile

159
Statistical values Reference manual

2.8.34. YMONSTAT - Multi-year monthly statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values of each month of year. Depending on the chosen operator
the minimum, maximum, range, sum, average, variance or standard deviation of each month of year
in infile is written to outfile. The date information in an output field is the date of the last con-
tributing input field. This can be change with the CDO option --timestat_date <first|middle|last>.

Operators

ymonmin Multi-year monthly minimum

o(01, x) = min{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = min{i(t, x), month(i(t)) = 12}
ymonmax Multi-year monthly maximum
o(01, x) = max{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = max{i(t, x), month(i(t)) = 12}
ymonrange Multi-year monthly range
o(01, x) = range{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = range{i(t, x), month(i(t)) = 12}
ymonsum Multi-year monthly sum
o(01, x) = sum{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = sum{i(t, x), month(i(t)) = 12}
ymonmean Multi-year monthly mean
o(01, x) = mean{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = mean{i(t, x), month(i(t)) = 12}
ymonavg Multi-year monthly average
o(01, x) = avg{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = avg{i(t, x), month(i(t)) = 12}
ymonstd Multi-year monthly standard deviation
Normalize by n.
o(01, x) = std{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = std{i(t, x), month(i(t)) = 12}
ymonstd1 Multi-year monthly standard deviation (n-1)
Normalize by (n-1).

160
Reference manual Statistical values

o(01, x) = std1{i(t, x), month(i(t)) = 01}

..
.
o(12, x) = std1{i(t, x), month(i(t)) = 12}
ymonvar Multi-year monthly variance
Normalize by n.
o(01, x) = var{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = var{i(t, x), month(i(t)) = 12}
ymonvar1 Multi-year monthly variance (n-1)
Normalize by (n-1).
o(01, x) = var1{i(t, x), month(i(t)) = 01}
..
.
o(12, x) = var1{i(t, x), month(i(t)) = 12}

Example
To compute the monthly mean over all input years use:
cdo ymonmean infile outfile

161
Statistical values Reference manual

2.8.35. YMONPCTL - Multi-year monthly percentile values

Synopsis

ymonpctl,p infile1 infile2 infile3 outfile

Description
This operator writes a certain percentile of each month of year in infile1 to outfile. The algorithm
uses histograms with minimum and maximum bounds given in infile2 and infile3, respectively.
The default number of histogram bins is 101. The default can be overridden by setting the environment
variable CDO_PCTL_NBINS to a different value. The files infile2 and infile3 should be the result of
corresponding ymonmin and ymonmax operations, respectively. The date information in an output
field is the date of the last contributing input field.

o(01, x) = pth percentile{i(t, x), month(i(t)) = 01}

..
.
o(12, x) = pth percentile{i(t, x), month(i(t)) = 12}

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the monthly 90th percentile over all input years use:
cdo ymonmin infile minfile
cdo ymonmax infile maxfile
cdo ymonpctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo ymonpctl,90 infile -ymonmin infile -ymonmax infile outfile

162
Reference manual Statistical values

2.8.36. YSEASSTAT - Multi-year seasonal statistical values

Synopsis

< operator > infile outfile

Description
This module computes statistical values of each season. Depending on the chosen operator the
minimum, maximum, range, sum, average, variance or standard deviation of each season in infile
is written to outfile. The date information in an output field is the date of the last contributing
input field.

Operators

yseasmin Multi-year seasonal minimum

o(1, x) = min{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = min{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = min{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = min{i(t, x), month(i(t)) = 09, 10, 11}
yseasmax Multi-year seasonal maximum
o(1, x) = max{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = max{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = max{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = max{i(t, x), month(i(t)) = 09, 10, 11}
yseasrange Multi-year seasonal range
o(1, x) = range{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = range{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = range{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = range{i(t, x), month(i(t)) = 09, 10, 11}
yseassum Multi-year seasonal sum
o(1, x) = sum{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = sum{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = sum{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = sum{i(t, x), month(i(t)) = 09, 10, 11}
yseasmean Multi-year seasonal mean
o(1, x) = mean{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = mean{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = mean{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = mean{i(t, x), month(i(t)) = 09, 10, 11}
yseasavg Multi-year seasonal average
o(1, x) = avg{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = avg{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = avg{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = avg{i(t, x), month(i(t)) = 09, 10, 11}
yseasstd Multi-year seasonal standard deviation
o(1, x) = std{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = std{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = std{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = std{i(t, x), month(i(t)) = 09, 10, 11}

163
Statistical values Reference manual

yseasstd1 Multi-year seasonal standard deviation (n-1)

o(1, x) = std1{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = std1{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = std1{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = std1{i(t, x), month(i(t)) = 09, 10, 11}
yseasvar Multi-year seasonal variance
o(1, x) = var{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = var{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = var{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = var{i(t, x), month(i(t)) = 09, 10, 11}
yseasvar1 Multi-year seasonal variance (n-1)
o(1, x) = var1{i(t, x), month(i(t)) = 12, 01, 02}
o(2, x) = var1{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = var1{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = var1{i(t, x), month(i(t)) = 09, 10, 11}

Example
To compute the seasonal mean over all input years use:
cdo yseasmean infile outfile

164
Reference manual Statistical values

2.8.37. YSEASPCTL - Multi-year seasonal percentile values

Synopsis

yseaspctl,p infile1 infile2 infile3 outfile

Description
This operator writes a certain percentile of each season in infile1 to outfile. The algorithm uses
histograms with minimum and maximum bounds given in infile2 and infile3, respectively. The
default number of histogram bins is 101. The default can be overridden by setting the environment
variable CDO_PCTL_NBINS to a different value. The files infile2 and infile3 should be the result of
corresponding yseasmin and yseasmax operations, respectively. The date information in an output
field is the date of the last contributing input field.

o(1, x) = pth percentile{i(t, x), month(i(t)) = 12, 01, 02}

o(2, x) = pth percentile{i(t, x), month(i(t)) = 03, 04, 05}
o(3, x) = pth percentile{i(t, x), month(i(t)) = 06, 07, 08}
o(4, x) = pth percentile{i(t, x), month(i(t)) = 09, 10, 11}

Parameter
p FLOAT Percentile number in 0, ..., 100

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
To compute the seasonal 90th percentile over all input years use:
cdo yseasmin infile minfile
cdo yseasmax infile maxfile
cdo yseaspctl,90 infile minfile maxfile outfile

Or shorter using operator piping:

cdo yseaspctl,90 infile -yseasmin infile -yseasmax infile outfile

165
Statistical values Reference manual

2.8.38. YDRUNSTAT - Multi-year daily running statistical values

Synopsis

< operator >,nts infile outfile

Description
This module writes running statistical values for each day of year in infile to outfile. Depending
on the chosen operator, the minimum, maximum, sum, average, variance or standard deviation of all
timesteps in running windows of which the medium timestep corresponds to a certain day of year is
computed. The date information in an output field is the date of the timestep in the middle of the last
contributing running window. Note that the operator have to be applied to a continuous time series
of daily measurements in order to yield physically meaningful results. Also note that the output time
series begins (nts-1)/2 timesteps after the first timestep of the input time series and ends (nts-1)/2
timesteps before the last one. For input data which are complete but not continuous, such as time
series of daily measurements for the same month or season within different years, the operator yields
physically meaningful results only if the input time series does include the (nts-1)/2 days before and
after each period of interest.

Operators

ydrunmin Multi-year daily running minimum

o(001, x) = min{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = min{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunmax Multi-year daily running maximum
o(001, x) = max{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = max{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunsum Multi-year daily running sum
o(001, x) = sum{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = sum{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunmean Multi-year daily running mean
o(001, x) = mean{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = mean{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunavg Multi-year daily running average
o(001, x) = avg{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = avg{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunstd Multi-year daily running standard deviation
Normalize by n.

o(001, x) = std{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = std{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}

166
Reference manual Statistical values

ydrunstd1 Multi-year daily running standard deviation (n-1)

Normalize by (n-1).
o(001, x) = std1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = std1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunvar Multi-year daily running variance
Normalize by n.
o(001, x) = var{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = var{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}
ydrunvar1 Multi-year daily running variance (n-1)
Normalize by (n-1).
o(001, x) = var1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = var1{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}

Parameter
nts INTEGER Number of timesteps

Example
Assume the input data provide a continuous time series of daily measurements. To compute the
running multi-year daily mean over all input timesteps for a running window of five days use:
cdo ydrunmean,5 infile outfile

Note that except for the standard deviation the results of the operators in this module are equivalent
to a composition of corresponding operators from the YDAYSTAT and RUNSTAT modules. For
instance, the above command yields the same result as:
cdo ydaymean -runmean,5 infile outfile

167
Statistical values Reference manual

2.8.39. YDRUNPCTL - Multi-year daily running percentile values

Synopsis

ydrunpctl,p,nts infile1 infile2 infile3 outfile

Description
This operator writes running percentile values for each day of year in infile1 to outfile. A certain
percentile is computed for all timesteps in running windows of which the medium timestep corresponds
to a certain day of year. The algorithm uses histograms with minimum and maximum bounds given
in infile2 and infile3, respectively. The default number of histogram bins is 101. The default
can be overridden by setting the environment variable CDO_PCTL_NBINS to a different value. The files
infile2 and infile3 should be the result of corresponding ydrunmin and ydrunmax operations,
respectively. The date information in an output field is the date of the timestep in the middle of
the last contributing running window. Note that the operator have to be applied to a continuous
time series of daily measurements in order to yield physically meaningful results. Also note that the
output time series begins (nts-1)/2 timesteps after the first timestep of the input time series and ends
(nts-1)/2 timesteps before the last. For input data which are complete but not continuous, such as
time series of daily measurements for the same month or season within different years, the operator
only yields physically meaningful results if the input time series does include the (nts-1)/2 days before
and after each period of interest.

o(001, x) = pth percentile{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 001}
..
.
o(366, x) = pth percentile{i(t, x), i(t + 1, x), ..., i(t + nts − 1, x); day[(i(t + (nts − 1)/2)] = 366}

Parameter
p FLOAT Percentile number in 0, ..., 100
nts INTEGER Number of timesteps

Environment
CDO_PCTL_NBINS Sets the number of histogram bins. The default number is 101.

Example
Assume the input data provide a continuous time series of daily measurements. To compute the
running multi-year daily 90th percentile over all input timesteps for a running window of five days
use:
cdo ydrunmin,5 infile minfile
cdo ydrunmax,5 infile maxfile
cdo ydrunpctl,90,5 infile minfile maxfile outfile

Or shorter using operator piping:

cdo ydrunpctl,90,5 infile -ydrunmin infile -ydrunmax infile outfile

168
Reference manual Correlation and co.

2.9. Correlation and co.

This sections contains modules for correlation and co. in grid space and over time.
In this section the abbreviations as in the following table are used:

n
Covariance X
n−1 (xi − x)(yi − y)
covar i=1
 −1   −1   −1 
n n n n n n
covar weighted by X X X X X X
wj  wi xi −  wj  wj xj  yi −  wj  w j yj 
   
{wi , i = 1, ..., n} j=1 i=1 j=1 j=1 j=1 j=1

Here is a short overview of all operators in this section:

fldcor Correlation in grid space

timcor Correlation over time

fldcovar Covariance in grid space

timcovar Covariance over time

169
Correlation and co. Reference manual

2.9.1. FLDCOR - Correlation in grid space

Synopsis

fldcor infile1 infile2 outfile

Description
The correlation coefficient is a quantity that gives the quality of a least squares fitting to the original
data. This operator correlates all gridpoints of two fields for each timestep. With
S(t) = {x, i1 (t, x) ̸= missval ∧ i2 (t, x) ̸= missval}
it is
P P
i1 (t, x)i2 (t, x)w(x) − i1 (t, x) i2 (t, x) w(x)
x∈S(t) x∈S(t)
o(t, 1) = v ! !
u
u P 2 P P 2 P
t i1 (t, x)2 w(x) − i1 (t, x) w(x) i2 (t, x)2 w(x) − i2 (t, x) w(x)
x∈S(t) x∈S(t) x∈S(t) x∈S(t)

where w(x) are the area weights obtained by the input streams. For every timestep t only those field
elements x belong to the sample, which have i1 (t, x) ̸= missval and i2 (t, x) =
̸ missval.

2.9.2. TIMCOR - Correlation over time

Synopsis

timcor infile1 infile2 outfile

Description
The correlation coefficient is a quantity that gives the quality of a least squares fitting to the original
data. This operator correlates each gridpoint of two fields over all timesteps. With
S(x) = {t, i1 (t, x) ̸= missval ∧ i2 (t, x) ̸= missval}
it is
P
i1 (t, x)i2 (t, x) − n i1 (t, x) i2 (t, x)
t∈S(x)
o(1, x) = v ! !
u
u P 2 P 2
t i1 (t, x)2 − n i1 (t, x) i2 (t, x)2 − n i2 (t, x)
t∈S(x) t∈S(x)

For every gridpoint x only those timesteps t belong to the sample, which have i1 (t, x) ̸= missval and
i2 (t, x) ̸= missval.

170
Reference manual Correlation and co.

2.9.3. FLDCOVAR - Covariance in grid space

Synopsis

fldcovar infile1 infile2 outfile

Description
This operator calculates the covariance of two fields over all gridpoints for each timestep. With
S(t) = {x, i1 (t, x) ̸= missval ∧ i2 (t, x) ̸= missval}
it is
−1  P  P 
 w(x) i1 (t, x) w(x) i2 (t, x)
X X x∈S(t) x∈S(t)
o(t, 1) =  w(x) w(x) i1 (t, x) −  i2 (t, x) −
 P  P 
w(x) w(x)

x∈S(t) x∈S(t)
x∈S(t) x∈S(t)

where w(x) are the area weights obtained by the input streams. For every timestep t only those field
elements x belong to the sample, which have i1 (t, x) ̸= missval and i2 (t, x) =
̸ missval.

2.9.4. TIMCOVAR - Covariance over time

Synopsis

timcovar infile1 infile2 outfile

Description
This operator calculates the covariance of two fields at each gridpoint over all timesteps. With
S(x) = {t, i1 (t, x) ̸= missval ∧ i2 (t, x) ̸= missval}
it is
X   
o(1, x) = n−1 i1 (t, x) − i1 (t, x) i2 (t, x) − i2 (t, x)
t∈S(x)

For every gridpoint x only those timesteps t belong to the sample, which have i1 (t, x) ̸= missval and
i2 (t, x) ̸= missval.

171
Regression Reference manual

2.10. Regression
This sections contains modules for linear regression of time series.
Here is a short overview of all operators in this section:

regres Regression

detrend Detrend

trend Trend

addtrend Add trend

subtrend Subtract trend

172
Reference manual Regression

2.10.1. REGRES - Regression

Synopsis

regres[,equal] infile outfile

Description
The values of the input file infile are assumed to be distributed as N (a + bt, σ 2 ) with unknown a,
b and σ 2 . This operator estimates the parameter b. For every field element x only those timesteps t
belong to the sample S(x), which have i(t, x) ̸= miss. It is
! !
′
P 1
P 1
P ′
i(t, x) − #S(x) i(t , x) t − #S(x) t
t∈S(x) t′ ∈S(x) t′ ∈S(x)
o(1, x) = !2
1
t′
P P
t− #S(x)
t∈S(x) t′ ∈S(x)

It is assumed that all timesteps are equidistant, if this is not the case set the parameter equal=false.

Parameter
equal BOOL Set to false for unequal distributed timesteps (default: true)

2.10.2. DETREND - Detrend time series

Synopsis

detrend[,equal] infile outfile

Description
Every time series in infile is linearly detrended. For every field element x only those timesteps t
belong to the sample S(x), which have i(t, x) ̸= miss. It is assumed that all timesteps are equidistant,
if this is not the case set the parameter equal=false. With
 
1 X 1 X
a(x) = i(t, x) − b(x)  t
#S(x) #S(x)
t∈S(x) t∈S(x)

and
! !
1
i(t′ , x) 1
t′
P P P
i(t, x) − #S(x) t− #S(x)
t∈S(x) t′ ∈S(x) t′ ∈S(x)
b(x) = !2
1
t′
P P
t− #S(x)
t∈S(x) t′ ∈S(x)

it is
o(t, x) = i(t, x) − (a(x) + b(x)t)

Parameter
equal BOOL Set to false for unequal distributed timesteps (default: true)

173
Regression Reference manual

Note
This operator has to keep the fields of all timesteps concurrently in the memory. If not enough
memory is available use the operators trend and subtrend.

Example
To detrend the data in infile and to store the detrended data in outfile use:
cdo detrend infile outfile

2.10.3. TREND - Trend of time series

Synopsis

trend[,equal] infile outfile1 outfile2

Description
The values of the input file infile are assumed to be distributed as N (a + bt, σ 2 ) with unknown
a, b and σ 2 . This operator estimates the parameter a and b. For every field element x only those
timesteps t belong to the sample S(x), which have i(t, x) ̸= miss. It is
 
1 X 1 X
o1 (1, x) = i(t, x) − b(x)  t
#S(x) #S(x)
t∈S(x) t∈S(x)

and
! !
1 ′ 1 ′
P P P
i(t, x) − #S(x) i(t , x) t− #S(x) t
t∈S(x) t′ ∈S(x) t′ ∈S(x)
o2 (1, x) = !2
1
t′
P P
t− #S(x)
t∈S(x) t′ ∈S(x)

Thus the estimation for a is stored in outfile1 and that for b is stored in outfile2. To subtract the
trend from the data see operator subtrend. It is assumed that all timesteps are equidistant, if this is
not the case set the parameter equal=false.

Parameter
equal BOOL Set to false for unequal distributed timesteps (default: true)

174
Reference manual Regression

2.10.4. TRENDARITH - Add or subtract a trend

Synopsis

< operator >[,equal] infile1 infile2 infile3 outfile

Description
This module is for adding or subtracting a trend computed by the operator trend.

Operators

addtrend Add trend

It is
o(t, x) = i1 (t, x) + (i2 (1, x) + i3 (1, x) · t)
where t is the timesteps.
subtrend Subtract trend
It is
o(t, x) = i1 (t, x) − (i2 (1, x) + i3 (1, x) · t)
where t is the timesteps.

Parameter
equal BOOL Set to false for unequal distributed timesteps (default: true)

Example
The typical call for detrending the data in infile and storing the detrended data in outfile is:
cdo trend infile afile bfile
cdo subtrend infile afile bfile outfile

The result is identical to a call of the operator detrend:

cdo detrend infile outfile

175
EOFs Reference manual

2.11. EOFs
This section contains modules to compute Empirical Orthogonal Functions and - once they are computed
- their principal coefficients.
An introduction to the theory of principal component analysis as applied here can be found in:
Principal Component Analysis in Meteorology and Oceanography [Preisendorfer]
Details about calculation in the time- and spatial spaces are found in:
Statistical Analysis in Climate Research [vonStorch]
EOFs are defined as the eigen values of the scatter matrix (covariance matrix) of the data. For the sake of
simplicity, samples are regarded as time series of anomalies

(z(t)) , t ∈ {1, . . . , n}

of (column-) vectors z(t) with p entries (where p is the gridsize). Thus, using the fact, that zj (t) are
anomalies, i.e.
n
X
⟨zj ⟩ = n−1 zj (i) = 0 ∀ 1 ≤ j ≤ p
i=1

the scatter matrix S can be written as

n h
X √ i h√ iT
S= Wz(t) Wz(t)
t=1

where W is the diagonal matrix containing the area weight of cell p0 in z at W(x, x).
The matrix S has a set of orthonormal eigenvectors ej , j = 1, . . . p, which are called empirical orthogonal
functions (EOFs) of the sample z. (Please note, that ej is the eigenvector of S and not the weighted
eigen-vector which would be Wej .) Let the corresponding eigenvalues be denoted λj . The vectors ej are
spatial patterns which explain a certain amount of variance of the time series z(t) that is related linearly
to λj . Thus, the spatial pattern defined by the first eigenvector (the one with the largest eigenvalue ) is the
pattern which explains a maximum possible amount of variance of the sample z(t). The orthonormality of
eigenvectors reads as
p h p 
̸ k
0 if j =
X p i hp i X
W(x, x)ej (x) W(x, x)ek (x) = W(x, x)ej (x)ek (x) =
x=1 x=1
1 if j = k

If all EOFs ej with λj ̸= 0 are calculated, the data can be reconstructed from
p
X
z(t, x) = W(x, x)aj (t)ej (x)
j=1

where aj are called the principal components or principal coefficients or EOF coefficients of z. These
coefficients - as readily seen from above - are calculated as the projection of an EOF ej onto a time step
of the data sample z(t0 ) as
p h
X i hp i h√ iT h√ i
p
aj (t0 ) = W(x, x)ej (x) W(x, x)z(t0 , x) = Wz(t0 ) Wej .
x=1

Here is a short overview of all operators in this section:

eof Calculate EOFs in spatial or time space

eoftime Calculate EOFs in time space
eofspatial Calculate EOFs in spatial space
eof3d Calculate 3-Dimensional EOFs in time space

eofcoeff Calculate principal coefficients of EOFs

176
Reference manual EOFs

2.11.1. EOFS - Empirical Orthogonal Functions

Synopsis
< operator >,neof infile outfile1 outfile2

Description
This module calculates empirical orthogonal functions of the data in infile as the eigen values of
the scatter matrix (covariance matrix) S of the data sample z(t). A more detailed description can be
found above.
Please note, that the input data are assumed to be anomalies.
If operator eof is chosen, the EOFs are computed in either time or spatial space, whichever is the
fastest. If the user already knows, which computation is faster, the module can be forced to perform
a computation in time- or gridspace by using the operators eoftime or eofspatial, respectively. This
can enhance performance, especially for very long time series, where the number of timesteps is larger
than the number of grid-points. Data in infile are assumed to be anomalies. If they are not, the
behavior of this module is not well defined. After execution outfile1 will contain all eigen-values
and outfile2 the eigenvectors e_j. All EOFs and eigen-values are computed. However, only the
first neof EOFs are written to outfile2. Nonetheless, outfile1 contains all eigen-values.
Missing values are not fully supported. Support is only checked for non-changing masks of missing
values in time. Although there still will be results, they are not trustworthy, and a warning will occur.
In the latter case we suggest to replace missing values by 0 in infile.

Operators
eof Calculate EOFs in spatial or time space

eoftime Calculate EOFs in time space

eofspatial Calculate EOFs in spatial space

eof3d Calculate 3-Dimensional EOFs in time space

Parameter
neof INTEGER Number of eigen functions

Environment
CDO_SVD_MODE Is used to choose the algorithm for eigenvalue calculation. Options are ’jacobi’
for a one-sided parallel jacobi-algorithm (only executed in parallel if -P flag
is set) and ’danielson_lanczos’ for a non-parallel d/l algorithm. The default
setting is ’jacobi’.
CDO_WEIGHT_MODE It is used to set the weight mode. The default is ’off’. Set it to ’on’ for a
weighted version.
MAX_JACOBI_ITER Is the maximum integer number of annihilation sweeps that is executed if the
jacobi-algorithm is used to compute the eigen values. The default value is 12.
FNORM_PRECISION Is the Frobenius norm of the matrix consisting of an annihilation pair of eigen-
vectors that is used to determine if the eigenvectors have reached a sufficient
level of convergence. If all annihilation-pairs of vectors have a norm below this
value, the computation is considered to have converged properly. Otherwise, a
warning will occur. The default value 1e-12.

177
EOFs Reference manual

Example
To calculate the first 40 EOFs of a data-set containing anomalies use:
cdo eof,40 infile outfile1 outfile2

If the dataset does not containt anomalies, process them first, and use:
cdo sub infile1 -timmean infile1 anom_file
cdo eof,40 anom_file outfile1 outfile2

178
Reference manual EOFs

2.11.2. EOFCOEFF - Principal coefficients of EOFs

Synopsis

eofcoeff infile1 infile2 obase

Description
This module calculates the time series of the principal coefficients for given EOF (empirical orthogonal
functions) and data. Time steps in infile1 are assumed to be the EOFs, time steps in infile2 are
assumed to be the time series. Note, that this operator calculates a non weighted dot product of the
fields in infile1 and infile2. For consistency set the environment variable CDO_WEIGHT_MODE=off
when using eof or eof3d. Given a set of EOFs e_j and a time series of data z(t) with p entries for
each timestep from which e_j have been calculated, this operator calculates the time series of the
projections of data onto each EOF
p
X
oj (t) = z(t, x)ej (x)
x=1

There will be a seperate file o_j for the principal coefficients of each EOF.
As the EOFs e_j are uncorrelated, so are their principal coefficients, i.e.
n  n
X 0 if j ̸= k X
oj (t)ok (t) = with oj (t) = 0∀j ∈ {1, . . . , p}.
t=1
λj if j = k t=1

There will be a separate file containing a time series of principal coefficients with time information from
infile2 for each EOF in infile1. Output files will be numbered as <obase><neof><suffix>
where neof+1 is the number of the EOF (timestep) in infile1 and suffix is the filename extension
derived from the file format.

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

Example
To calculate principal coefficients of the first 40 EOFs of anom_file, and write them to files beginning
with obase, use:
export CDO_WEIGHT_MODE=off
cdo eof,40 anom_file eval_file eof_file
cdo eofcoeff eof_file anom_file obase

The principal coefficients of the first EOF will be in the file obase000000.nc (and so forth for higher
EOFs, nth EOF will be in obase<n-1>).
If the dataset infile does not containt anomalies, process them first, and use:
export CDO_WEIGHT_MODE=off
cdo sub infile -timmean infile anom_file
cdo eof,40 anom_file eval_file eof_file
cdo eofcoeff eof_file anom_file obase

179
Interpolation Reference manual

2.12. Interpolation
This section contains modules to interpolate datasets. There are several operators to interpolate horizontal
fields to a new grid. Some of those operators can handle only 2D fields on a regular rectangular grid. Vertical
interpolation of 3D variables is possible from hybrid model levels to height or pressure levels. Interpolation
in time is possible between time steps and years.
Here is a short overview of all operators in this section:

remapbil Bilinear interpolation

genbil Generate bilinear interpolation weights

remapbic Bicubic interpolation

genbic Generate bicubic interpolation weights

remapnn Nearest neighbor remapping

gennn Generate nearest neighbor remap weights

remapdis Distance weighted average remapping

gendis Generate distance weighted average remap weights

remapcon First order conservative remapping

gencon Generate 1st order conservative remap weights

remapcon2 Second order conservative remapping

gencon2 Generate 2nd order conservative remap weights

remaplaf Largest area fraction remapping

genlaf Generate largest area fraction remap weights

remap Grid remapping

remapeta Remap vertical hybrid level

ml2pl Model to pressure level interpolation

ml2hl Model to height level interpolation

ap2pl Air pressure to pressure level interpolation

gh2hl Geometric height to height level interpolation

intlevel Linear level interpolation

intlevel3d Linear level interpolation onto a 3D vertical coordinate

intlevelx3d like intlevel3d but with extrapolation

inttime Interpolation between timesteps

intntime Interpolation between timesteps

intyear Interpolation between two years

180
Reference manual Interpolation

2.12.1. REMAPBIL - Bilinear interpolation

Synopsis
< operator >,grid infile outfile

Description
This module contains operators for a bilinear remapping of fields between grids in spherical coordi-
nates. The interpolation is based on an adapted SCRIP library version. For a detailed description
of the interpolation method see [SCRIP]. This interpolation method only works on quadrilateral
curvilinear source grids. Below is a schematic illustration of the bilinear remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators
remapbil Bilinear interpolation
Performs a bilinear interpolation on all input fields.
genbil Generate bilinear interpolation weights
Generates bilinear interpolation weights for the first input field and writes the result
to a file. The format of this file is NetCDF following the SCRIP convention. Use the
operator remap to apply this remapping weights to a data file with the same source
grid.

Parameter
grid STRING Target grid description file or name

Environment
REMAP_EXTRAPOLATE This variable is used to switch the extrapolation feature ’on’ or ’off’. By
default the extrapolation is enabled for circular grids.

Example
Say infile contains fields on a quadrilateral curvilinear grid. To remap all fields bilinear to a Gaussian
N32 grid, type:
cdo remapbil,n32 infile outfile

181
Interpolation Reference manual

2.12.2. REMAPBIC - Bicubic interpolation

Synopsis
< operator >,grid infile outfile

Description
This module contains operators for a bicubic remapping of fields between grids in spherical coordi-
nates. The interpolation is based on an adapted SCRIP library version. For a detailed description
of the interpolation method see [SCRIP]. This interpolation method only works on quadrilateral
curvilinear source grids. Below is a schematic illustration of the bicubic remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators
remapbic Bicubic interpolation
Performs a bicubic interpolation on all input fields.
genbic Generate bicubic interpolation weights
Generates bicubic interpolation weights for the first input field and writes the result
to a file. The format of this file is NetCDF following the SCRIP convention. Use the
operator remap to apply this remapping weights to a data file with the same source
grid.

Parameter
grid STRING Target grid description file or name

Environment
REMAP_EXTRAPOLATE This variable is used to switch the extrapolation feature ’on’ or ’off’. By
default the extrapolation is enabled for circular grids.

Example
Say infile contains fields on a quadrilateral curvilinear grid. To remap all fields bicubic to a Gaussian
N32 grid, type:
cdo remapbic,n32 infile outfile

182
Reference manual Interpolation

2.12.3. REMAPNN - Nearest neighbor remapping

Synopsis

< operator >,grid infile outfile

Description
This module contains operators for a nearest neighbor remapping of fields between grids in spherical
coordinates. Below is a schematic illustration of the nearest neighbor remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators

remapnn Nearest neighbor remapping

Performs a nearest neighbor remapping on all input fields.
gennn Generate nearest neighbor remap weights
Generates nearest neighbor remapping weights for the first input field and writes the
result to a file. The format of this file is NetCDF following the SCRIP convention.
Use the operator remap to apply this remapping weights to a data file with the same
source grid.

Parameter
grid STRING Target grid description file or name

Environment
REMAP_EXTRAPOLATE This variable is used to switch the extrapolation feature ’on’ or ’off’. By
default the extrapolation is enabled for this remapping method.
CDO_GRIDSEARCH_RADIUS Grid search radius in degree, default 180 degree.

183
Interpolation Reference manual

2.12.4. REMAPDIS - Distance weighted average remapping

Synopsis

< operator >,grid[,neighbors] infile outfile

Description
This module contains operators for an inverse distance weighted average remapping of the four nearest
neighbor values of fields between grids in spherical coordinates. The default number of 4 neighbors can
be changed with the neighbors parameter. Below is a schematic illustration of the distance weighted
average remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators

remapdis Distance weighted average remapping

Performs an inverse distance weighted averaged remapping of the nearest neighbor
values on all input fields.
gendis Generate distance weighted average remap weights
Generates distance weighted averaged remapping weights of the nearest neighbor values
for the first input field and writes the result to a file. The format of this file is NetCDF
following the SCRIP convention. Use the operator remap to apply this remapping
weights to a data file with the same source grid.

Parameter
grid STRING Target grid description file or name
neighbors INTEGER Number of nearest neighbors [default: 4]

Environment
REMAP_EXTRAPOLATE This variable is used to switch the extrapolation feature ’on’ or ’off’. By
default the extrapolation is enabled for this remapping method.
CDO_GRIDSEARCH_RADIUS Grid search radius in degree, default 180 degree.

184
Reference manual Interpolation

2.12.5. REMAPCON - First order conservative remapping

Synopsis
< operator >,grid infile outfile

Description
This module contains operators for a first order conservative remapping of fields between grids in
spherical coordinates. The operators in this module uses code from the YAC software package to
compute the conservative remapping weights. For a detailed description of the interpolation method
see [YAC]. The interpolation method is completely general and can be used for any grid on a sphere.
The search algorithm for the conservative remapping requires that no grid cell occurs more than once.
Below is a schematic illustration of the 1st order conservative remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators
remapcon First order conservative remapping
Performs a first order conservative remapping on all input fields.
gencon Generate 1st order conservative remap weights
Generates first order conservative remapping weights for the first input field and writes
the result to a file. The format of this file is NetCDF following the SCRIP convention.
Use the operator remap to apply this remapping weights to a data file with the same
source grid.

Parameter
grid STRING Target grid description file or name

Environment
CDO_REMAP_NORM This variable is used to choose the normalization of the conservative interpola-
tion. By default CDO_REMAP_NORM is set to ’fracarea’. ’fracarea’ uses the sum of
the non-masked source cell intersected areas to normalize each target cell field
value. This results in a reasonable flux value but the flux is not locally con-
served. The option ’destarea’ uses the total target cell area to normalize each
target cell field value. Local flux conservation is ensured, but unreasonable flux
values may result.

185
Interpolation Reference manual

REMAP_AREA_MIN This variable is used to set the minimum destination area fraction. The default
of this variable is 0.0.

Example
Say infile contains fields on a quadrilateral curvilinear grid. To remap all fields conservative to a
Gaussian N32 grid, type:
cdo remapcon,n32 infile outfile

186
Reference manual Interpolation

2.12.6. REMAPCON2 - Second order conservative remapping

Synopsis
< operator >,grid infile outfile

Description
This module contains operators for a second order conservative remapping of fields between grids in
spherical coordinates. The interpolation is based on an adapted SCRIP library version. For a detailed
description of the interpolation method see [SCRIP]. The second order conservative remapping is not
available for unstructured source grids. The search algorithm for the conservative remapping requires
that no grid cell occurs more than once. Below is a schematic illustration of the 2nd order conservative
remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators
remapcon2 Second order conservative remapping
Performs a second order conservative remapping on all input fields.
gencon2 Generate 2nd order conservative remap weights
Generates second order conservative remapping weights for the first input field and
writes the result to a file. The format of this file is NetCDF following the SCRIP
convention. Use the operator remap to apply this remapping weights to a data file
with the same source grid.

Parameter
grid STRING Target grid description file or name

Environment
CDO_REMAP_NORM This variable is used to choose the normalization of the conservative interpola-
tion. By default CDO_REMAP_NORM is set to ’fracarea’. ’fracarea’ uses the sum of
the non-masked source cell intersected areas to normalize each target cell field
value. This results in a reasonable flux value but the flux is not locally con-
served. The option ’destarea’ uses the total target cell area to normalize each
target cell field value. Local flux conservation is ensured, but unreasonable flux
values may result.

187
Interpolation Reference manual

REMAP_AREA_MIN This variable is used to set the minimum destination area fraction. The default
of this variable is 0.0.

Note
The SCRIP conservative remapping method doesn’t work correctly for some grid combinations.

Example
Say infile contains fields on a quadrilateral curvilinear grid. To remap all fields conservative (2nd
order) to a Gaussian N32 grid, type:
cdo remapcon2,n32 infile outfile

188
Reference manual Interpolation

2.12.7. REMAPLAF - Largest area fraction remapping

Synopsis

< operator >,grid infile outfile

Description
This module contains operators for a largest area fraction remapping of fields between grids in spherical
coordinates. The operators in this module uses code from the YAC software package to compute
the largest area fraction. For a detailed description of the interpolation method see [YAC]. The
interpolation method is completely general and can be used for any grid on a sphere. The search
algorithm for this remapping method requires that no grid cell occurs more than once. Below is a
schematic illustration of the largest area fraction conservative remapping:

The figure on the left side shows the input data on a regular lon/lat source grid and on the right side
the remapped result on an unstructured triangular target grid. The figure in the middle shows the
input data with the target grid. Grid cells with missing value are grey colored.

Operators

remaplaf Largest area fraction remapping

Performs a largest area fraction remapping on all input fields.
genlaf Generate largest area fraction remap weights
Generates largest area fraction remapping weights for the first input field and writes
the result to a file. The format of this file is NetCDF following the SCRIP convention.
Use the operator remap to apply this remapping weights to a data file with the same
source grid.

Parameter
grid STRING Target grid description file or name

Environment
REMAP_AREA_MIN This variable is used to set the minimum destination area fraction. The default
of this variable is 0.0.

189
Interpolation Reference manual

2.12.8. REMAP - Grid remapping

Synopsis

remap,grid,weights infile outfile

Description
Interpolation between different horizontal grids can be a very time-consuming process. Especially if
the data are on an unstructured and/or a large grid. In this case the interpolation process can be split
into two parts. Firstly the generation of the interpolation weights, which is the most time-consuming
part. These interpolation weights can be reused for every remapping process with the operator remap.
This operator remaps all input fields to a new horizontal grid. The remap type and the interpolation
weights of one input grid are read from a NetCDF file. More weights are computed if the input
fields are on different grids. The NetCDF file with the weights should follow the [SCRIP] convention.
Normally these weights come from a previous call to one of the genXXX operators (e.g. genbil) or
were created by the original SCRIP package.

Parameter
grid STRING Target grid description file or name
weights STRING Interpolation weights (SCRIP NetCDF file)

Environment
CDO_REMAP_NORM This variable is used to choose the normalization of the conservative
interpolation. By default CDO_REMAP_NORM is set to ’fracarea’. ’fracarea’
uses the sum of the non-masked source cell intersected areas to normalize
each target cell field value. This results in a reasonable flux value but
the flux is not locally conserved. The option ’destarea’ uses the total
target cell area to normalize each target cell field value. Local flux
conservation is ensured, but unreasonable flux values may result.
REMAP_EXTRAPOLATE This variable is used to switch the extrapolation feature ’on’ or ’off’.
By default the extrapolation is enabled for remapdis, remapnn and for
circular grids.
REMAP_AREA_MIN This variable is used to set the minimum destination area fraction. The
default of this variable is 0.0.
CDO_GRIDSEARCH_RADIUS Grid search radius in degree, default 180 degree.

Example
Say infile contains fields on a quadrilateral curvilinear grid. To remap all fields bilinear to a Gaussian
N32 grid use:
cdo genbil,n32 infile remapweights.nc
cdo remap,n32,remapweights.nc infile outfile

The result will be the same as:

cdo remapbil,n32 infile outfile

190
Reference manual Interpolation

2.12.9. REMAPETA - Remap vertical hybrid level

Synopsis

remapeta,vct[,oro] infile outfile

Description
This operator interpolates between different vertical hybrid levels. This include the preparation of
consistent data for the free atmosphere. The procedure for the vertical interpolation is based on the
HIRLAM scheme and was adapted from [INTERA]. The vertical interpolation is based on the vertical
integration of the hydrostatic equation with few adjustments. The basic tasks are the following one:
• at first integration of hydrostatic equation
• extrapolation of surface pressure
• Planetary Boundary-Layer (PBL) proutfile interpolation
• interpolation in free atmosphere
• merging of both proutfiles
• final surface pressure correction
The vertical interpolation corrects the surface pressure. This is simply a cut-off or an addition of
air mass. This mass correction should not influence the geostrophic velocity field in the middle
troposhere. Therefore the total mass above a given reference level is conserved. As reference level the
geopotential height of the 400 hPa level is used. Near the surface the correction can affect the vertical
structure of the PBL. Therefore the interpolation is done using the potential temperature. But in
the free atmosphere above a certain n (n=0.8 defining the top of the PBL) the interpolation is done
linearly. After the interpolation both proutfiles are merged. With the resulting temperature/pressure
correction the hydrostatic equation is integrated again and adjusted to the reference level finding the
final surface pressure correction. A more detailed description of the interpolation can be found in
[INTERA]. This operator requires all variables on the same horizontal grid.

Parameter
vct STRING File name of an ASCII dataset with the vertical coordinate table
oro STRING File name with the orography (surf. geopotential) of the target dataset (optional)

Environment
REMAPETA_PTOP Sets the minimum pressure level for condensation. Above this level the humidity
is set to the constant 1.E-6. The default value is 0 Pa.

Note
The code numbers or the variable names of the required parameter have to follow the [ECHAM]
convention.
Use the sinfo command to test if your vertical coordinate system is recognized as hybrid system.
In case remapeta complains about not finding any data on hybrid model levels you may wish to use
the setzaxis command to generate a zaxis description which conforms to the ECHAM convention. See
section "1.4 Z-axis description" for an example how to define a hybrid Z-axis.

191
Interpolation Reference manual

Example
To remap between different hybrid model level data use:
cdo remapeta,vct infile outfile

Here is an example vct file with 19 hybrid model level:

0 0.00000000000000000 0.00000000000000000
1 2000.00000000000000000 0.00000000000000000
2 4000.00000000000000000 0.00000000000000000
3 6046.10937500000000000 0.00033899326808751
4 8267.92968750000000000 0.00335718691349030
5 10609.51171875000000000 0.01307003945112228
6 12851.10156250000000000 0.03407714888453484
7 14698.50000000000000000 0.07064980268478394
8 15861.12890625000000000 0.12591671943664551
9 16116.23828125000000000 0.20119541883468628
10 15356.92187500000000000 0.29551959037780762
11 13621.46093750000000000 0.40540921688079834
12 11101.55859375000000000 0.52493220567703247
13 8127.14453125000000000 0.64610791206359863
14 5125.14062500000000000 0.75969839096069336
15 2549.96899414062500000 0.85643762350082397
16 783.19506835937500000 0.92874687910079956
17 0.00000000000000000 0.97298520803451538
18 0.00000000000000000 0.99228149652481079
19 0.00000000000000000 1.00000000000000000

192
Reference manual Interpolation

2.12.10. VERTINTML - Vertical interpolation

Synopsis
ml2pl,plevels infile outfile
ml2hl,hlevels infile outfile

Description
Interpolates 3D variables on hybrid sigma pressure level to pressure or height levels. The input file
should contain the log. surface pressure or the surface pressure. To extrapolate the temperature,
the surface geopotential is also needed. It is assumed that the geopotential heights are located at
the hybrid layer interfaces. For the lowest layer of geopotential heights the surface geopotential is
required. The pressure, temperature, geopotential height, and surface geopotential are identified by
their GRIB1 code number or NetCDF CF standard name. Supported parameter tables are: WMO
standard table number 2 and ECMWF local table number 128.

CF standard name Units GRIB 1 code

surface_air_pressure Pa 134
air_temperature K 130
surface_geopotential m2 s-2 129
geopotential_height m 156

Use the alias ml2plx/ml2hlx or the environment variable EXTRAPOLATE to extrapolate missing values.
This operator requires all variables on the same horizontal grid. Missing values in the input data are
not supported.

Operators
ml2pl Model to pressure level interpolation
Interpolates 3D variables on hybrid sigma pressure level to pressure level.
ml2hl Model to height level interpolation
Interpolates 3D variables on hybrid sigma pressure level to height level. The procedure is
the same as for the operator ml2pl except for the pressure levels being calculated from the
heights by: plevel = 101325 ∗ exp(hlevel/ − 7000)

Parameter
plevels FLOAT Pressure levels in pascal
hlevels FLOAT Height levels in meter

Environment
EXTRAPOLATE If set to 1 extrapolate missing values.

Note
The components of the hybrid coordinate must always be avaiable at the hybrid layer interfaces even
if the data is defined at the hybrid layer midpoints.

Example
To interpolate hybrid model level data to pressure levels of 925, 850, 500 and 200 hPa use:
cdo ml2pl,92500,85000,50000,20000 infile outfile

193
Interpolation Reference manual

2.12.11. VERTINTAP - Vertical pressure interpolation

Synopsis

ap2pl,plevels infile outfile

Description
Interpolate 3D variables on hybrid sigma height coordinates to pressure levels. The input file must
contain the 3D air pressure in pascal. The air pressure is identified by the NetCDF CF standard
name air_pressure. Use the alias ap2plx or the environment variable EXTRAPOLATE to extrapolate
missing values. This operator requires all variables on the same horizontal grid.

Parameter
plevels FLOAT Comma-separated list of pressure levels in pascal

Environment
EXTRAPOLATE If set to 1 extrapolate missing values.

Note
This is a specific implementation for NetCDF files from the ICON model, it may not work with data
from other sources.

Example
To interpolate 3D variables on hybrid sigma height level to pressure levels of 925, 850, 500 and 200
hPa use:
cdo ap2pl,92500,85000,50000,20000 infile outfile

194
Reference manual Interpolation

2.12.12. VERTINTGH - Vertical height interpolation

Synopsis

gh2hl,hlevels infile outfile

Description
Interpolate 3D variables on hybrid sigma height coordinates to height levels. The input file must
contain the 3D geometric height in meter. The geometric height is identified by the NetCDF CF
standard name geometric_height_at_full_level_center. Use the alias gh2hlx or the environ-
ment variable EXTRAPOLATE to extrapolate missing values. This operator requires all variables on the
same horizontal grid.

Parameter
hlevels FLOAT Comma-separated list of height levels in meter

Environment
EXTRAPOLATE If set to 1 extrapolate missing values.

Note
This is a specific implementation for NetCDF files from the ICON model, it may not work with data
from other sources.

Example
To interpolate 3D variables on hybrid sigma height level to height levels of 20, 100, 500, 1000, 5000,
10000 and 20000 meter use:
cdo gh2hl,20,100,500,1000,5000,10000,20000 infile outfile

195
Interpolation Reference manual

2.12.13. INTLEVEL - Linear level interpolation

Synopsis
intlevel,parameter infile outfile

Description
This operator performs a linear vertical interpolation of 3D variables. The target levels can be specified
with the level parameter or read in via a Z-axis description file.

Parameter
level FLOAT Comma-separated list of target levels
file STRING Path to a file containing a description of the Z-axis

Example
To interpolate 3D variables on height levels to a new set of height levels use:
cdo intlevel,level=10,50,100,500,1000 infile outfile

2.12.14. INTLEVEL3D - Linear level interpolation from/to 3D vertical coordinates

Synopsis
< operator >,tgtcoordinate infile1 infile2 outfile

Description
This operator performs a linear vertical interpolation of 3D variables fields with given 3D vertical
coordinates. infile1 contains the 3D data variables and infile2 the 3D vertical source coordinate.
The parameter tgtcoordinate is a datafile with the 3D vertical target coordinate.

Operators
intlevel3d Linear level interpolation onto a 3D vertical coordinate

intlevelx3d like intlevel3d but with extrapolation

Parameter
tgtcoordinate STRING filename for 3D vertical target coordinates

Example
To interpolate 3D variables from one set of 3D height levels into another one where
• infile2 contains a single 3D variable, which represents the source 3D vertical coordinate
• infile1 contains the source data, which the vertical coordinate from infile2 belongs to
• tgtcoordinate only contains the target 3D height levels
cdo intlevel3d,tgtcoordinate infile1 infile2 outfile

196
Reference manual Interpolation

2.12.15. INTTIME - Time interpolation

Synopsis

inttime,date,time[,inc] infile outfile

intntime,n infile outfile

Description
This module performs linear interpolation between timesteps. Interpolation is only performed if both
values exist. If both values are missing values, the result is also a missing value. If only one value
exists, it is taken if the time weighting is greater than or equal to 0.5. So no new value will be created
at existing time steps, if the value is missing there.

Operators

inttime Interpolation between timesteps

This operator creates a new dataset by linear interpolation between timesteps. The
user has to define the start date/time with an optional increment.
intntime Interpolation between timesteps
This operator performs linear interpolation between timesteps. The user has to define
the number of timesteps from one timestep to the next.

Parameter
date STRING Start date (format YYYY-MM-DD)
time STRING Start time (format hh:mm:ss)
inc STRING Optional increment (seconds, minutes, hours, days, months, years) [default:
0hour]
n INTEGER Number of timesteps from one timestep to the next

Example
Assumed a 6 hourly dataset starts at 1987-01-01 12:00:00. To interpolate this time series to a one
hourly dataset use:
cdo inttime,1987-01-01,12:00:00,1hour infile outfile

197
Interpolation Reference manual

2.12.16. INTYEAR - Year interpolation

Synopsis

intyear,years infile1 infile2 obase

Description
This operator performs linear interpolation between two years, timestep by timestep. The input
files need to have the same structure with the same variables. The output files will be named
<obase><yyyy><suffix> where yyyy will be the year and suffix is the filename extension derived
from the file format.

Parameter
years INTEGER Comma-separated list or first/last[/inc] range of years

Environment
CDO_FILE_SUFFIX Set the default file suffix. This suffix will be added to the output file names
instead of the filename extension derived from the file format. Set this variable
to NULL to disable the adding of a file suffix.

Note
This operator needs to open all output files simultaneously. The maximum number of open files
depends on the operating system!

Example
Assume there are two monthly mean datasets over a year. The first dataset has 12 timesteps for the
year 1985 and the second one for the year 1990. To interpolate the years between 1985 and 1990
month by month use:
cdo intyear,1986,1987,1988,1989 infile1 infile2 year

Example result of ’dir year*’ for NetCDF datasets:

y e a r 1 9 8 6 . nc y e a r 1 9 8 7 . nc y e a r 1 9 8 8 . nc y e a r 1 9 8 9 . nc

198
Reference manual Transformation

2.13. Transformation
This section contains modules to perform spectral transformations.
Here is a short overview of all operators in this section:

sp2gp Spectral to gridpoint

gp2sp Gridpoint to spectral

sp2sp Spectral to spectral

dv2ps D and V to velocity potential and stream function

dv2uv Divergence and vorticity to U and V wind

uv2dv U and V wind to divergence and vorticity

fourier Fourier transformation

199
Transformation Reference manual

2.13.1. SPECTRAL - Spectral transformation

Synopsis

< operator >[,type|trunc] infile outfile

Description
This module transforms fields on a global regular Gaussian grid to spectral coefficients and vice
versa. The transformation is achieved by applying Fast Fourier Transformation (FFT) first and
direct Legendre Transformation afterwards in gp2sp. In sp2gp the inverse Legendre Transformation
and inverse FFT are used. Missing values are not supported.
The relationship between the spectral resolution, governed by the truncation number T, and the grid
resolution depends on the number of grid points at which the shortest wavelength field is represented.
For a grid with 2N points between the poles (so 4N grid points in total around the globe) the
relationship is:
linear grid: the shortest wavelength is represented by 2 grid points → 4N ≃ 2(TL + 1)
quadratic grid: the shortest wavelength is represented by 3 grid points → 4N ≃ 3(TQ + 1)
cubic grid: the shortest wavelength is represented by 4 grid points → 4N ≃ 4(TC + 1)
The quadratic grid is used by ECHAM and ERA15. ERA40 is using a linear Gaussian grid reflected
by the TL notation.
The following table shows the calculation of the number of latitudes and the triangular truncation
for the different grid types:

Gridtype Number of latitudes: nlat Triangular truncation: ntr

linear NINT((ntr*2 + 1)/2) (nlat*2 - 1) / 2
quadratic NINT((ntr*3 + 1)/2) (nlat*2 - 1) / 3
cubic NINT((ntr*4 + 1)/2) (nlat*2 - 1) / 4

Operators

sp2gp Spectral to gridpoint

Convert all spectral fields to a global regular Gaussian grid. The optional parameter trunc
must be greater than the input truncation.
gp2sp Gridpoint to spectral
Convert all Gaussian gridpoint fields to spectral fields. The optional parameter trunc
must be lower than the input truncation.

Parameter
type STRING Type of the grid: quadratic, linear, cubic (default: type=quadratic)
trunc STRING Triangular truncation

Note
To speed up the calculations, the Legendre polynoms are kept in memory. This requires a relatively
large amount of memory. This is for example 12GB for T1279 data.

200
Reference manual Transformation

Example
To transform spectral coefficients from T106 to N80 Gaussian grid use:
cdo sp2gp infile outfile

To transform spectral coefficients from TL159 to N80 Gaussian grid use:

cdo sp2gp,type=linear infile outfile

201
Transformation Reference manual

2.13.2. SPECCONV - Spectral conversion

Synopsis

sp2sp,trunc infile outfile

Description
Changed the triangular truncation of all spectral fields. This operator performs downward conversion
by cutting the resolution. Upward conversions are achieved by filling in zeros.

Parameter
trunc INTEGER New spectral resolution

2.13.3. WIND2 - D and V to velocity potential and stream function

Synopsis

dv2ps infile outfile

Description
Calculate spherical harmonic coefficients of velocity potential and stream function from spherical
harmonic coefficients of relative divergence and vorticity. The divergence and vorticity need to have
the names sd and svo or code numbers 155 and 138.

202
Reference manual Transformation

2.13.4. WIND - Wind transformation

Synopsis

< operator >[,gridtype] infile outfile

Description
This module converts relative divergence and vorticity to U and V wind and vice versa. Divergence
and vorticity are spherical harmonic coefficients in spectral space and U and V are on a global regular
Gaussian grid. The Gaussian latitudes need to be ordered from north to south. Missing values are
not supported.
The relationship between the spectral resolution, governed by the truncation number T, and the grid
resolution depends on the number of grid points at which the shortest wavelength field is represented.
For a grid with 2N points between the poles (so 4N grid points in total around the globe) the
relationship is:
linear grid: the shortest wavelength is represented by 2 grid points → 4N ≃ 2(TL + 1)
quadratic grid: the shortest wavelength is represented by 3 grid points → 4N ≃ 3(TQ + 1)
cubic grid: the shortest wavelength is represented by 4 grid points → 4N ≃ 4(TC + 1)
The quadratic grid is used by ECHAM and ERA15. ERA40 is using a linear Gaussian grid reflected
by the TL notation.
The following table shows the calculation of the number of latitudes and the triangular truncation
for the different grid types:

Gridtype Number of latitudes: nlat Triangular truncation: ntr

linear NINT((ntr*2 + 1)/2) (nlat*2 - 1) / 2
quadratic NINT((ntr*3 + 1)/2) (nlat*2 - 1) / 3
cubic NINT((ntr*4 + 1)/2) (nlat*2 - 1) / 4

Operators

dv2uv Divergence and vorticity to U and V wind

Calculate U and V wind on a Gaussian grid from spherical harmonic coefficients of relative
divergence and vorticity. The divergence and vorticity need to have the names sd and svo
or code numbers 155 and 138.
uv2dv U and V wind to divergence and vorticity
Calculate spherical harmonic coefficients of relative divergence and vorticity from U and
V wind. The U and V wind need to be on a Gaussian grid and need to have the names u
and v or the code numbers 131 and 132.

Parameter
gridtype STRING Type of the grid: quadratic, linear (default: quadratic)

Note
To speed up the calculations, the Legendre polynoms are kept in memory. This requires a relatively
large amount of memory. This is for example 12GB for T1279 data.

203
Transformation Reference manual

Example
Assume a dataset has at least spherical harmonic coefficients of divergence and vorticity. To transform
the spectral divergence and vorticity to U and V wind on a Gaussian grid use:
cdo dv2uv infile outfile

204
Reference manual Transformation

2.13.5. FOURIER - Fourier transformation

Synopsis

fourier,epsilon infile outfile

Description
The fourier operator performs the fourier transformation or the inverse fourier transformation of
all input fields. If the number of timesteps is a power of 2 then the algorithm of the Fast Fourier
Transformation (FFT) is used.
It is
n−1
1 X
o(t, x) = √ i(t, x)eϵ2πij
n j=0

where a user given epsilon = −1 leads to the forward transformation and a user given epsilon = 1
leads to the backward transformation.
If the input stream infile consists only of complex fields, then the fields of outfile, computed by
cdo -f ext fourier,1 -fourier,-1 infile outfile

are the same than that of infile. For real input files see function retocomplex.

Parameter
epsilon INTEGER -1: forward transformation; 1: backward transformation

Note
Complex numbers can only be stored in NetCDF4 and EXTRA format.

205
Import/Export Reference manual

2.14. Import/Export
This section contains modules to import and export data files which can not read or write directly with
CDO.
Here is a short overview of all operators in this section:

import_binary Import binary data sets

import_cmsaf Import CM-SAF HDF5 files

import_amsr Import AMSR binary files

input ASCII input

inputsrv SERVICE ASCII input
inputext EXTRA ASCII input

output ASCII output

outputf Formatted output
outputint Integer output
outputsrv SERVICE ASCII output
outputext EXTRA ASCII output

outputtab Table output

gmtxyz GMT xyz format

gmtcells GMT multiple segment format

206
Reference manual Import/Export

2.14.1. IMPORTBINARY - Import binary data sets

Synopsis

import_binary infile outfile

Description
This operator imports gridded binary data sets via a GrADS data descriptor file. The GrADS data
descriptor file contains a complete description of the binary data as well as instructions on where to
find the data and how to read it. The descriptor file is an ASCII file that can be created easily with
a text editor. The general contents of a gridded data descriptor file are as follows:
• Filename for the binary data
• Missing or undefined data value
• Mapping between grid coordinates and world coordinates
• Description of variables in the binary data set
A detailed description of the components of a GrADS data descriptor file can be found in [GrADS].
Here is a list of the supported components: BYTESWAPPED, CHSUB, DSET, ENDVARS, FILE-
HEADER, HEADERBYTES, OPTIONS, TDEF, TITLE, TRAILERBYTES, UNDEF, VARS, XDEF,
XYHEADER, YDEF, ZDEF

Note
Only 32-bit IEEE floats are supported for standard binary files!

Example
To convert a binary data file to NetCDF use:
cdo -f nc import_binary infile.ctl outfile.nc

Here is an example of a GrADS data descriptor file:

DSET ^ i n f i l e . b i n
OPTIONS s e q u e n t i a l
UNDEF −9e+33
XDEF 360 LINEAR −179.5 1
YDEF 180 LINEAR −89.5 1
ZDEF 1 LINEAR 1 1
TDEF 1 LINEAR 0 0 : 0 0 Z15jun1989 12 hr
VARS 1
param 1 99 d e s c r i p t i o n o f t h e v a r i a b l e
ENDVARS

The binary data file infile.bin contains one parameter on a global 1 degree lon/lat grid written with
FORTRAN record length headers (sequential).

207
Import/Export Reference manual

2.14.2. IMPORTCMSAF - Import CM-SAF HDF5 files

Synopsis

import_cmsaf infile outfile

Description
This operator imports gridded CM-SAF (Satellite Application Facility on Climate Monitoring) HDF5
files. CM-SAF exploits data from polar-orbiting and geostationary satellites in order to provide
climate monitoring products of the following parameters:

Cloud parameters: cloud fraction (CFC), cloud type (CTY), cloud phase (CPH), cloud top height,
pressure and temperature (CTH,CTP,CTT), cloud optical thickness (COT), cloud water
path (CWP).
Surface radiation components: Surface albedo (SAL); surface incoming (SIS) and net (SNS) shortwave
radiation; surface downward (SDL) and outgoing (SOL) longwave radiation, surface net
longwave radiation (SNL) and surface radiation budget (SRB).
Top-of-atmosphere radiation components: Incoming (TIS) and reflected (TRS) solar radiative flux
at top-of-atmosphere. Emitted thermal radiative flux at top-of-atmosphere (TET).
Water vapour: Vertically integrated water vapour (HTW), layered vertically integrated water vapour
and layer mean temperature and relative humidity for 5 layers (HLW), temperature and
mixing ratio at 6 pressure levels.

Daily and monthly mean products can be ordered via the CM-SAF web page (www.cmsaf.eu). Prod-
ucts with higher spatial and temporal resolution, i.e. instantaneous swath-based products, are avail-
able on request ([email protected]). All products are distributed free-of-charge. More informa-
tion on the data is available on the CM-SAF homepage (www.cmsaf.eu).
Daily and monthly mean products are provided in equal-area projections. CDO reads the projection
parameters from the metadata in the HDF5-headers in order to allow spatial operations like remap-
ping. For spatial operations with instantaneous products on original satellite projection, additional
files with arrays of latitudes and longitudes are needed. These can be obtained from CM-SAF together
with the data.

Note
To use this operator, it is necessary to build CDO with HDF5 support (version 1.6 or higher). The
PROJ library (version 5.0 or higher) is needed for full support of the remapping functionality.

Example
A typical sequence of commands with this operator could look like this:
cdo -f nc remapbil,r360x180 -import_cmsaf cmsaf_product.hdf output.nc

(bilinear remapping to a predefined global grid with 1 deg resolution and conversion to NetCDF).

If you work with CM-SAF data on original satellite project, an additional file with information on
geolocation is required, to perform such spatial operations:
cdo -f nc remapbil,r720x360 -setgrid,cmsaf_latlon.h5 -import_cmsaf cmsaf.hdf out.nc

Some CM-SAF data are stored as scaled integer values. For some operations, it could be desirable
(or necessary) to increase the accuracy of the converted products:

208
Reference manual Import/Export

cdo -b f32 -f nc fldmean -sellonlatbox,0,10,0,10 -remapbil,r720x360 \

-import_cmsaf cmsaf_product.hdf output.nc

2.14.3. IMPORTAMSR - Import AMSR binary files

Synopsis

import_amsr infile outfile

Description
This operator imports gridded binary AMSR (Advanced Microwave Scanning Radiometer) data. The
binary data files are available from the AMSR ftp site (ftp://ftp.ssmi.com/amsre). Each file consists
of twelve (daily) or five (averaged) 0.25 x 0.25 degree grid (1440,720) byte maps. For daily files,
six daytime maps in the following order, Time (UTC), Sea Surface Temperature (SST), 10 meter
Surface Wind Speed (WSPD), Atmospheric Water Vapor (VAPOR), Cloud Liquid Water (CLOUD),
and Rain Rate (RAIN), are followed by six nighttime maps in the same order. Time-Averaged files
contain just the geophysical layers in the same order [SST, WSPD, VAPOR, CLOUD, RAIN]. More
information to the data is available on the AMSR homepage http://www.remss.com/amsr.

Example
To convert monthly binary AMSR files to NetCDF use:
cdo -f nc amsre_yyyymmv5 amsre_yyyymmv5.nc

209
Import/Export Reference manual

2.14.4. INPUT - Formatted input

Synopsis

input,grid[,zaxis] outfile
inputsrv outfile
inputext outfile

Description
This module reads time series of one 2D variable from standard input. All input fields need to have
the same horizontal grid. The format of the input depends on the chosen operator.

Operators

input ASCII input

Reads fields with ASCII numbers from standard input and stores them in outfile.
The numbers read are exactly that ones which are written out by the output operator.
inputsrv SERVICE ASCII input
Reads fields with ASCII numbers from standard input and stores them in outfile.
Each field should have a header of 8 integers (SERVICE likely). The numbers that are
read are exactly that ones which are written out by the outputsrv operator.
inputext EXTRA ASCII input
Read fields with ASCII numbers from standard input and stores them in outfile.
Each field should have header of 4 integers (EXTRA likely). The numbers read are
exactly that ones which are written out by the outputext operator.

Parameter
grid STRING Grid description file or name
zaxis STRING Z-axis description file

Example
Assume an ASCII dataset contains a field on a global regular grid with 32 longitudes and 16 latitudes
(512 elements). To create a GRIB1 dataset from the ASCII dataset use:
cdo -f grb input,r32x16 outfile.grb < my_ascii_data

210
Reference manual Import/Export

2.14.5. OUTPUT - Formatted output

Synopsis
output infiles
outputf ,format[,nelem] infiles
outputint infiles
outputsrv infiles
outputext infiles

Description
This module prints all values of all input datasets to standard output. All input fields need to have
the same horizontal grid. All input files need to have the same structure with the same variables.
The format of the output depends on the chosen operator.

Operators
output ASCII output
Prints all values to standard output. Each row has 6 elements with the C-style format
"%13.6g".
outputf Formatted output
Prints all values to standard output. The format and number of elements for each row
have to be specified by the parameters format and nelem. The default for nelem is 1.
outputint Integer output
Prints all values rounded to the nearest integer to standard output.
outputsrv SERVICE ASCII output
Prints all values to standard output. Each field with a header of 8 integers (SERVICE
likely).
outputext EXTRA ASCII output
Prints all values to standard output. Each field with a header of 4 integers (EXTRA
likely).

Parameter
format STRING C-style format for one element (e.g. %13.6g)
nelem INTEGER Number of elements for each row (default: nelem = 1)

Example
To print all field elements of a dataset formatted with "%8.4g" and 8 values per line use:
cdo outputf,%8.4g,8 infile

Example result of a dataset with one field on 64 grid points:

261.7 262 257.8 252.5 248.8 247.7 246.3 246.1
250.6 252.6 253.9 254.8 252 246.6 249.7 257.9
273.4 266.2 259.8 261.6 257.2 253.4 251 263.7
267.5 267.4 272.2 266.7 259.6 255.2 272.9 277.1
275.3 275.5 276.4 278.4 282 269.6 278.7 279.5
282.3 284.5 280.3 280.3 280 281.5 284.7 283.6
292.9 290.5 293.9 292.6 292.7 292.8 294.1 293.6
293.8 292.6 291.2 292.6 293.2 292.8 291 291.2

211
Import/Export Reference manual

2.14.6. OUTPUTTAB - Table output

Synopsis
outputtab,parameter infiles outfile

Description
This operator prints a table of all input datasets to standard output. infiles is an arbitrary number
of input files. All input files need to have the same structure with the same variables on different
timesteps. All input fields need to have the same horizontal grid.
The contents of the table depends on the chosen parameters. The format of each table parameter is
keyname[:len]. len is the optional length of a table entry. The number of significant digits of floating
point parameters can be set with the CDO option --precision, the default is 7. Here is a list of all
valid keynames:

Keyname Type Description

value FLOAT Value of the variable [len:8]
name STRING Name of the variable [len:8]
param STRING Parameter ID (GRIB1: code[.tabnum]; GRIB2: num[.cat[.dis]]) [len:11]
code INTEGER Code number [len:4]
x FLOAT X coordinate of the original grid [len:6]
y FLOAT Y coordinate of the original grid [len:6]
lon FLOAT Longitude coordinate in degrees [len:6]
lat FLOAT Latitude coordinate in degrees [len:6]
lev FLOAT Vertical level [len:6]
xind INTEGER Grid x index [len:4]
yind INTEGER Grid y index [len:4]
timestep INTEGER Timestep number [len:6]
date STRING Date (format YYYY-MM-DD) [len:10]
time STRING Time (format hh:mm:ss) [len:8]
year INTEGER Year [len:5]
month INTEGER Month [len:2]
day INTEGER Day [len:2]
nohead INTEGER Disable output of header line

Parameter
parameter STRING Comma-separated list of keynames, one for each column of the table

Example
To print a table with name, date, lon, lat and value information use:
cdo outputtab,name,date,lon,lat,value infile

Here is an example output of a time series with the yearly mean temperatur at lon=10/lat=53.5:
# name date lon lat value
tsurf 1991−12−31 10 53.5 8.83903
tsurf 1992−12−31 10 53.5 8.17439
tsurf 1993−12−31 10 53.5 7.90489
tsurf 1994−12−31 10 53.5 10.0216
tsurf 1995−12−31 10 53.5 9.07798

212
Reference manual Import/Export

2.14.7. OUTPUTGMT - GMT output

Synopsis

< operator > infile

Description
This module prints the first field of the input dataset to standard output. The output can be used to
generate 2D Lon/Lat plots with [GMT]. The format of the output depends on the chosen operator.

Operators

gmtxyz GMT xyz format

The operator exports the first field to the GMT xyz ASCII format. The output can be
used to create contour plots with the GMT module pscontour.
gmtcells GMT multiple segment format
The operator exports the first field to the GMT multiple segment ASCII format. The
output can be used to create shaded gridfill plots with the GMT module psxy.

Example
1) GMT shaded contour plot of a global temperature field with a resolution of 4 degree. The contour
interval is 3 with a rainbow color table.

cdo gmtxyz temp > data.gmt

makecpt -T213/318/3 -Crainbow > gmt.cpt
pscontour -K -JQ0/10i -Rd -I -Cgmt.cpt data.gmt > gmtplot.ps
pscoast -O -J -R -Dc -W -B40g20 >> gmtplot.ps

−160˚ −120˚ −80˚ −40˚ 0˚ 40˚ 80˚ 120˚ 160˚

80˚ 80˚

40˚ 40˚

0˚ 0˚

−40˚ −40˚

−80˚ −80˚

−160˚ −120˚ −80˚ −40˚ 0˚ 40˚ 80˚ 120˚ 160˚

2) GMT shaded gridfill plot of a global temperature field with a resolution of 4 degree. The contour
interval is 3 with a rainbow color table.

213
Import/Export Reference manual

cdo gmtcells temp > data.gmt

makecpt -T213/318/3 -Crainbow > gmt.cpt
psxy -K -JQ0/10i -Rd -L -Cgmt.cpt -m data.gmt > gmtplot.ps
pscoast -O -J -R -Dc -W -B40g20 >> gmtplot.ps

−160˚ −120˚ −80˚ −40˚ 0˚ 40˚ 80˚ 120˚ 160˚

80˚ 80˚

40˚ 40˚

0˚ 0˚

−40˚ −40˚

−80˚ −80˚

−160˚ −120˚ −80˚ −40˚ 0˚ 40˚ 80˚ 120˚ 160˚

214
Reference manual Miscellaneous

2.15. Miscellaneous
This section contains miscellaneous modules which do not fit to the other sections before.
Here is a short overview of all operators in this section:

gradsdes GrADS data descriptor file

after ECHAM standard post processor

bandpass Bandpass filtering

lowpass Lowpass filtering
highpass Highpass filtering

gridarea Grid cell area

gridweights Grid cell weights

smooth Smooth grid points

smooth9 9 point smoothing

setvals Set list of old values to new values

setrtoc Set range to constant
setrtoc2 Set range to constant others to constant2

gridcellindex Get grid cell index from lon/lat point

const Create a constant field

random Create a field with random numbers
topo Create a field with topography
seq Create a time series
stdatm Create values for pressure and temperature for hydrostatic atmosphere

timsort Sort over the time

uvDestag Destaggering of u/v wind components

rotuvNorth Rotate u/v wind to North pole.
projuvLatLon Cylindrical Equidistant projection

rotuvb Backward rotation

mrotuvb Backward rotation of MPIOM data

mastrfu Mass stream function

sealevelpressure Sea level pressure

gheight Geopotential height

adisit Potential temperature to in-situ temperature

adipot In-situ temperature to potential temperature

rhopot Calculates potential density

histcount Histogram count

histsum Histogram sum
histmean Histogram mean
histfreq Histogram frequency

215
Miscellaneous Reference manual

sethalo Set the bounds of a field

wct Windchill temperature

fdns Frost days where no snow index per time period

strwin Strong wind days index per time period

strbre Strong breeze days index per time period

strgal Strong gale days index per time period

hurr Hurricane days index per time period

cmorlite CMOR lite

verifygrid Verify grid coordinates

hpdegrade Degrade healpix

hpupgrade Upgrade healpix

216
Reference manual Miscellaneous

2.15.1. GRADSDES - GrADS data descriptor file

Synopsis

gradsdes[,mapversion] infile

Description
Creates a [GrADS] data descriptor file. Supported file formats are GRIB1, NetCDF, SERVICE,
EXTRA and IEG. For GRIB1 files the GrADS map file is also generated. For SERVICE and EXTRA
files the grid have to be specified with the CDO option ’-g <grid>’. This module takes infile in
order to create filenames for the descriptor (infile.ctl) and the map (infile.gmp) file.

Parameter
mapversion INTEGER Format version of the GrADS map file for GRIB1 datasets. Use 1 for
a machine specific version 1 GrADS map file, 2 for a machine independent version 2
GrADS map file and 4 to support GRIB files >2GB. A version 2 map file can be used
only with GrADS version 1.8 or newer. A version 4 map file can be used only with
GrADS version 2.0 or newer. The default is 4 for files >2GB, otherwise 2.

Example
To create a GrADS data descriptor file from a GRIB1 dataset use:
cdo gradsdes infile.grb

This will create a descriptor file with the name infile.ctl and the map file infile.gmp.
Assumed the input GRIB1 dataset has 3 variables over 12 timesteps on a Gaussian N16 grid. The
contents of the resulting GrADS data description file is approximately:
DSET ^ i n f i l e . grb
DTYPE GRIB
INDEX ^ i n f i l e . gmp
XDEF 64 LINEAR 0 . 0 0 0 0 0 0 5 . 6 2 5 0 0 0
YDEF 32 LEVELS −85.761 −80.269 −74.745 −69.213 −63.679 −58.143
−52.607 −47.070 −41.532 −35.995 −30.458 −24.920
−19.382 −13.844 −8.307 −2.769 2.769 8.307
13.844 19.382 24.920 30.458 35.995 41.532
47.070 52.607 58.143 63.679 69.213 74.745
80.269 85.761
ZDEF 4 LEVELS 925 850 500 200
TDEF 12 LINEAR 1 2 : 0 0 Z1jan1987 1mo
TITLE i n f i l e . grb T21 g r i d
OPTIONS y r e v
UNDEF −9e+33
VARS 3
geosp 0 129 ,1 ,0 s u r f a c e g e o p o t e n t i a l ( o ro g ra ph y ) [m^2/ s ^ 2 ]
t 4 130 ,99 ,0 temperature [K]
tslm1 0 139 ,1 ,0 s u r f a c e temperature of land [K]
ENDVARS

217
Miscellaneous Reference manual

2.15.2. AFTERBURNER - ECHAM standard post processor

Synopsis

after[,vct] infiles outfile

Description
The "afterburner" is the standard post processor for [ECHAM] GRIB and NetCDF data which provides
the following operations:

• Extract specified variables and levels

• Compute derived variables

• Transform spectral data to Gaussian grid representation

• Vertical interpolation to pressure levels

• Compute temporal means

This operator reads selection parameters as namelist from stdin. Use the UNIX redirection "<namelistfile"
to read the namelist from file.

The input files can’t be combined with other CDO operators because of an optimized reader for this
operator.

Namelist
Namelist parameter and there defaults:

TYPE=0 , CODE=−1, LEVEL=−1, INTERVAL=0 , MEAN=0, EXTRAPOLATE=1

TYPE controls the transformation and vertical interpolation. Transforming spectral data to Gaussian
grid representation and vertical interpolation to pressure levels are performed in a chain of steps. The
TYPE parameter may be used to stop the chain at a certain step. Valid values are:

TYPE = 0 : Hybrid level spectral coefficients

TYPE = 10 : Hybrid level fourier coefficients
TYPE = 11 : Hybrid level z o n a l mean s e c t i o n s
TYPE = 20 : Hybrid level gauss g r i d s
TYPE = 30 : Pressure level gauss g r i d s
TYPE = 40 : Pressure level fourier coefficients
TYPE = 41 : Pressure level z o n a l mean s e c t i o n s
TYPE = 50 : Pressure level spectral coefficients
TYPE = 60 : Pressure level fourier coefficients
TYPE = 61 : Pressure level z o n a l mean s e c t i o n s
TYPE = 70 : Pressure level gauss g r i d s

Vorticity, divergence, streamfunction and velocity potential need special treatment in the vertical
transformation. They are not available as types 30, 40 and 41. If you select one of these combinations,
type is automatically switched to the equivalent types 70, 60 and 61. The type of all other variables
will be switched too, because the type is a global parameter.

CODE selects the variables by the ECHAM GRIB1 code number (1-255). The default value -1 processes
all detected codes. Derived variables computed by the afterburner:

218
Reference manual Miscellaneous

Code Name Longname Units Level Needed Codes

34 low_cld low cloud single 223 on modellevel
35 mid_cld mid cloud single 223 on modellevel
36 hih_cld high cloud single 223 on modellevel
131 u u-velocity m/s atm (ml+pl) 138, 155
132 v v-velocity m/s atm (ml+pl) 138, 155
135 omega vertical velocity Pa/s atm (ml+pl) 138, 152, 155
148 stream streamfunction mˆ 2/s atm (ml+pl) 131, 132
149 velopot velocity potential mˆ 2/s atm (ml+pl) 131, 132
151 slp mean sea level pressure Pa surface 129, 130, 152
156 geopoth geopotential height m atm (ml+pl) 129, 130, 133, 152
157 rhumidity relative humidity atm (ml+pl) 130, 133, 152
189 sclfs surface solar cloud forcing surface 176-185
190 tclfs surface thermal cloud forcing surface 177-186
191 sclf0 top solar cloud forcing surface 178-187
192 tclf0 top thermal cloud forcing surface 179-188
259 windspeed windspeed m/s atm (ml+pl) sqrt(u*u+v*v)
260 precip total precipitation surface 142+143

LEVEL selects the hybrid or pressure levels. The allowed values depends on the parameter TYPE. The
default value -1 processes all detected levels.
INTERVAL selects the processing interval. The default value 0 process data on monthly intervals.
INTERVAL=1 sets the interval to daily.
MEAN=1 compute and write monthly or daily mean fields. The default value 0 writes out all timesteps.
EXTRAPOLATE=0 switch of the extrapolation of missing values during the interpolation from model to
pressure level (only available with MEAN=0 and TYPE=30). The default value 1 extrapolate missing
values.
Possible combinations of TYPE, CODE and MEAN:

TYPE CODE MEAN

0/10/11 130 temperature 0
0/10/11 131 u-velocity 0
0/10/11 132 v-velocity 0
0/10/11 133 specific humidity 0
0/10/11 138 vorticity 0
0/10/11 148 streamfunction 0
0/10/11 149 velocity potential 0
0/10/11 152 LnPs 0
0/10/11 155 divergence 0
>11 all codes 0/1

Parameter
vct STRING File with VCT in ASCII format

Example
To interpolate ECHAM hybrid model level data to pressure levels of 925, 850, 500 and 200 hPa, use:
cdo after infile outfile << EON
TYPE=30 LEVEL=92500,85000,50000,20000
EON

219
Miscellaneous Reference manual

2.15.3. FILTER - Time series filtering

Synopsis

bandpass,fmin,fmax infile outfile

lowpass,fmax infile outfile
highpass,fmin infile outfile

Description
This module takes the time series for each gridpoint in infile and (fast fourier) transforms it into the
frequency domain. According to the particular operator and its parameters certain frequencies are
filtered (set to zero) in the frequency domain and the spectrum is (inverse fast fourier) transformed
back into the time domain. To determine the frequency the time-axis of infile is used. (Data should
have a constant time increment since this assumption applies for transformation. However, the time
increment has to be different from zero.) All frequencies given as parameter are interpreted per year.
This is done by the assumption of a 365-day calendar. Consequently if you want to perform multiyear-
filtering accurately you have to delete the 29th of February. If your infile has a 360 year calendar
the frequency parameters fmin respectively fmax should be multiplied with a factor of 360/365 in
order to obtain accurate results. For the set up of a frequency filter the frequency parameters have to
be adjusted to a frequency in the data. Here fmin is rounded down and fmax is always rounded up.
Consequently it is possible to use bandpass with fmin=fmax without getting a zero-field for outfile.
Hints for efficient usage:
• to get reliable results the time-series has to be detrended (cdo detrend)
• the lowest frequency greater zero that can be contained in infile is 1/(N*dT),
• the greatest frequency is 1/(2dT) (Nyquist frequency),
with N the number of timesteps and dT the time increment of infile in years.
Missing value support for operators in this module is not implemented, yet!

Operators

bandpass Bandpass filtering

Bandpass filtering (pass for frequencies between fmin and fmax). Suppresses all vari-
ability outside the frequency range specified by [fmin,fmax].
lowpass Lowpass filtering
Lowpass filtering (pass for frequencies lower than fmax). Suppresses all variability
with frequencies greater than fmax.
highpass Highpass filtering
Highpass filtering (pass for frequencies greater than fmin). Suppresses all variabilty
with frequencies lower than fmin.

Parameter
fmin FLOAT Minimum frequency per year that passes the filter.
fmax FLOAT Maximum frequency per year that passes the filter.

Note
For better performace of these operators use the CDO configure option --with-fftw3.

220
Reference manual Miscellaneous

Example
Now assume your data are still hourly for a time period of 5 years but with a 365/366-day- calendar
and you want to suppress the variability on timescales greater or equal to one year (we suggest here
to use a number x bigger than one (e.g. x=1.5) since there will be dominant frequencies around the
peak (if there is one) as well due to the issue that the time series is not of infinite length). Therefor
you can use the following:
cdo highpass,x -del29feb infile outfile

Accordingly you might use the following to suppress variability on timescales shorter than one year:
cdo lowpass,1 -del29feb infile outfile

Finally you might be interested in 2-year variability. If you want to suppress the seasonal cycle as
well as say the longer cycles in climate system you might use
cdo bandpass,x,y -del29feb infile outfile

with x<=0.5 and y >=0.5.

2.15.4. GRIDCELL - Grid cell quantities

Synopsis

< operator > infile outfile

Description
This module reads the grid cell area of the first grid from the input stream. If the grid cell area
is missing it will be computed from the grid coordinates. The area of a grid cell is calculated using
spherical triangles from the coordinates of the center and the vertices. The base is a unit sphere which
is scaled with the radius of the earth. The default earth radius is 6371000 meter. This value can be
changed with the environment variable PLANET_RADIUS. Depending on the chosen operator the
grid cell area or weights are written to the output stream.

Operators

gridarea Grid cell area

Writes the grid cell area to the output stream. If the grid cell area have to be
computed it is scaled with the earth radius to square meters.
gridweights Grid cell weights
Writes the grid cell area weights to the output stream.

Environment
PLANET_RADIUS This variable is used to scale the computed grid cell areas to square meters. By
default PLANET_RADIUS is set to an earth radius of 6371000 meter.

221
Miscellaneous Reference manual

2.15.5. SMOOTH - Smooth grid points

Synopsis

smooth[,options] infile outfile

smooth9 infile outfile

Description
Smooth all grid points of a horizontal grid. Options is a comma-separated list of "key=value" pairs
with optional parameters.

Operators

smooth Smooth grid points

Performs a N point smoothing on all input fields. The number of points used depend
on the search radius (radius) and the maximum number of points (maxpoints). Per
default all points within the search radius of 1degree are used. The weights for the
points depend on the form of the curve and the distance. The implemented form of the
curve is linear with constant default weights of 0.25 at distance 0 (weight0) and at the
search radius (weightR).
smooth9 9 point smoothing
Performs a 9 point smoothing on all fields with a quadrilateral curvilinear grid. The
result at each grid point is a weighted average of the grid point plus the 8 surrounding
points. The center point receives a weight of 1.0, the points at each side and above and
below receive a weight of 0.5, and corner points receive a weight of 0.3. All 9 points are
multiplied by their weights and summed, then divided by the total weight to obtain the
smoothed value. Any missing data points are not included in the sum; points beyond
the grid boundary are considered to be missing. Thus the final result may be the result
of an averaging with less than 9 points.

Parameter
nsmooth INTEGER Number of times to smooth, default nsmooth=1
radius STRING Search radius, default radius=1deg (units: deg, rad, km, m)
maxpoints INTEGER Maximum number of points, default maxpoints=<gridsize>
form STRING Form of the curve, default form=linear
weight0 FLOAT Weight at distance 0, default weight0=0.25
weightR FLOAT Weight at the search radius, default weightR=0.25

2.15.6. DELTAT - Difference between timesteps

Synopsis

deltat infile outfile

Description
This operator computes the difference between each timestep.

222
Reference manual Miscellaneous

2.15.7. REPLACEVALUES - Replace variable values

Synopsis

setvals,oldval,newval[,...] infile outfile

setrtoc,rmin,rmax,c infile outfile
setrtoc2,rmin,rmax,c,c2 infile outfile

Description
This module replaces old variable values with new values, depending on the operator.

Operators

setvals Set list of old values to new values

Supply a list of n pairs of old and new values.
setrtoc Set rangeto constant
c if i(t, x) ≥ rmin ∧ i(t, x) ≤ rmax
o(t, x) =
i(t, x) if i(t, x) < rmin ∨ i(t, x) > rmax
setrtoc2 Set rangeto constant others to constant2
c if i(t, x) ≥ rmin ∧ i(t, x) ≤ rmax
o(t, x) =
c2 if i(t, x) < rmin ∨ i(t, x) > rmax

Parameter
oldval,newval,... FLOAT Pairs of old and new values
rmin FLOAT Lower bound
rmax FLOAT Upper bound
c FLOAT New value - inside range
c2 FLOAT New value - outside range

2.15.8. GETGRIDCELL - Get grid cell index

Synopsis

gridcellindex[,parameter] infile

Description
Get the grid cell index of one grid point selected by the parameter lon and lat.

Parameter
lon INTEGER Longitude of the grid cell in degree
lat INTEGER Latitude of the grid cell in degree

223
Miscellaneous Reference manual

2.15.9. VARGEN - Generate a field

Synopsis

const,const,grid outfile
random,grid[,seed] outfile
topo[,grid] outfile
seq,start,end[,inc] outfile
stdatm,levels outfile

Description
Generates a dataset with one or more fields

Operators

const Create a constant field

Creates a constant field. All field elements of the grid have the same value.
random Create a field with random numbers
Creates a field with rectangularly distrubuted random numbers in the interval [0,1].
topo Create a field with topography
Creates a field with topography data, per default on a global half degree grid.
seq Create a time series
Creates a time series with field size 1 and field elements beginning with a start value in
time step 1 which is increased from one time step to the next.
stdatm Create values for pressure and temperature for hydrostatic atmosphere
Creates pressure and temperature values for the given list of vertical levels. The formu-
lars are:
  
z
g H exp( H )T0 +∆T
P (z) = P0 exp − R T0 log T0 +∆T

z


T (z) = T0 + ∆T exp − H
with the following constants

T0 = 213K : offset to get a surface temperature of 288K

∆T = 75K : Temperature lapse rate for 10Km
P0 = 1013.25hPa : surface pressure
H= 10000.0m : scale height
g= 9.80665 m
s2 : earth gravity
R= J
287.05 kgK : gas constant for air

This is the solution for the hydrostatic equations and is only valid for the troposphere
(constant positive lapse rate). The temperature increase in the stratosphere and other
effects of the upper atmosphere are not taken into account.

224
Reference manual Miscellaneous

Parameter
const FLOAT Constant
seed INTEGER The seed for a new sequence of pseudo-random numbers [default: 1]
grid STRING Target grid description file or name
start FLOAT Start value of the loop
end FLOAT End value of the loop
inc FLOAT Increment of the loop [default: 1]
levels FLOAT Target levels in metre above surface

Example
To create a standard atmosphere dataset on a given horizontal grid:
cdo enlarge,gridfile -stdatm,10000,8000,5000,3000,2000,1000,500,200,0 outfile

2.15.10. TIMSORT - Timsort

Synopsis

timsort infile outfile

Description
Sorts the elements in ascending order over all timesteps for every field position. After sorting it is:

o(t1 , x) <= o(t2 , x) ∀(t1 < t2 ), x

Example
To sort all field elements of a dataset over all timesteps use:
cdo timsort infile outfile

225
Miscellaneous Reference manual

2.15.11. WINDTRANS - Wind Transformation

Synopsis

uvDestag,u,v[,-/+0.5[,-/+0.5]] infile outfile

rotuvNorth,u,v infile outfile
projuvLatLon,u,v infile outfile

Description
This module contains special operators for datsets with wind components on a rotated lon/lat grid,
e.g. data from the regional model HIRLAM or REMO.

Operators

uvDestag Destaggering of u/v wind components

This is a special operator for destaggering of wind components. If the file contains
a grid with temperature (name=’t’ or code=11) then grid_temp will be used for
destaggered wind.
rotuvNorth Rotate u/v wind to North pole.
This is an operator for transformation of wind-vectors from grid-relative to north-
pole relative for the whole file. (FAST implementation with JACOBIANS)
projuvLatLon Cylindrical Equidistant projection
Thus is an operator for transformation of wind-vectors from the globe-spherical
coordinate system into a flat Cylindrical Equidistant (lat-lon) projection. (FAST
JACOBIAN implementation)

Parameter
u,v STRING Pair of u,v wind components (use variable names or code numbers)
-/+0.5,-/+0.5 STRING Destaggered grid offsets are optional (default -0.5,-0.5)

Example
Typical operator sequence on HIRLAM NWP model output (LAMH_D11 files):
cdo uvDestag,33,34 inputfile inputfile_destag
cdo rotuvNorth,33,34 inputfile_destag inputfile_rotuvN

226
Reference manual Miscellaneous

2.15.12. ROTUVB - Rotation

Synopsis

rotuvb,u,v,... infile outfile

Description
This is a special operator for datsets with wind components on a rotated grid, e.g. data from the
regional model REMO. It performs a backward transformation of velocity components U and V from
a rotated spherical system to a geographical system.

Parameter
u,v,... STRING Pairs of zonal and meridional velocity components (use variable names or
code numbers)

Note
This is a specific implementation for data from the REMO model, it may not work with data from
other sources.

Example
To transform the u and v velocity of a dataset from a rotated spherical system to a geographical
system use:
cdo rotuvb,u,v infile outfile

2.15.13. MROTUVB - Backward rotation of MPIOM data

Synopsis

mrotuvb infile1 infile2 outfile

Description
MPIOM data are on a rotated Arakawa C grid. The velocity components U and V are located
on the edges of the cells and point in the direction of the grid lines and rows. With mrotuvb the
velocity vector is rotated in latitudinal and longitudinal direction. Before the rotation, U and V
are interpolated to the scalar points (cell center). U is located with the coordinates for U in infile1
and V in infile2. mrotuvb assumes a positive meridional flow for a flow from grid point(i,j) to grid
point(i,j+1) and positive zonal flow for a flow from grid point(i+1,j) to point(i,j).

Note
This is a specific implementation for data from the MPIOM model, it may not work with data from
other sources.

227
Miscellaneous Reference manual

2.15.14. MASTRFU - Mass stream function

Synopsis

mastrfu infile outfile

Description
This is a special operator for the post processing of the atmospheric general circulation model
[ECHAM]. It computes the mass stream function (code=272). The input dataset have to be a
zonal mean of v-velocity [m/s] (code=132) on pressure levels.

Example
To compute the mass stream function from a zonal mean v-velocity dataset use:
cdo mastrfu infile outfile

2.15.15. DERIVEPAR - Derived model parameters

Synopsis

< operator > infile outfile

Description
This module contains operators that calculate derived model parameters. These are currently the
parameters sea level pressure and geopotential height. All necessary input parameters are identified
by their GRIB1 code number or the NetCDF CF standard name. Supported GRIB1 parameter tables
are: WMO standard table number 2 and ECMWF local table number 128.

CF standard name Units GRIB 1 code

surface_air_pressure Pa 134
air_temperature K 130
specific_humidity kg/kg 133
surface_geopotential m2 s-2 129
geopotential_height m 156

Operators

sealevelpressure Sea level pressure

This operator computes the sea level pressure (air_pressure_at_sea_level).
Required input fields are surface_air_pressure, surface_geopotential and air_temperature
on full hybrid sigma pressure levels.
gheight Geopotential height
This operator computes the geopotential height (geopotential_height) on full
model levels in metres. Required input fields are surface_air_pressure, sur-
face_geopotential, specific_humidity and air_temperature on full hybrid sigma
pressure levels. Note, this procedure is an approximation, which doesn’t take
into account the effects of e.g. cloud ice and water, rain and snow.

228
Reference manual Miscellaneous

2.15.16. ADISIT - Potential temperature to in-situ temperature and vice versa

Synopsis

< operator >[,pressure] infile outfile

Description

Operators

adisit Potential temperature to in-situ temperature

This is a special operator for the post processing of the ocean and sea ice model [MPIOM].
It converts potential temperature adiabatically to in-situ temperature to(t, s, p). Required
input fields are sea water potential temperature (name=tho; code=2) and sea water salinity
(name=sao; code=5). Pressure is calculated from the level information or can be specified
by the optional parameter. Output fields are sea water temperature (name=to; code=20)
and sea water salinity (name=s; code=5).
adipot In-situ temperature to potential temperature
This is a special operator for the post processing of the ocean and sea ice model [MPIOM].
It converts in-situ temperature to potential temperature tho(to, s, p). Required input fields
are sea water in-situ temperature (name=t; code=2) and sea water salinity (name=sao,s;
code=5). Pressure is calculated from the level information or can be specified by the
optional parameter. Output fields are sea water temperature (name=tho; code=2) and
sea water salinity (name=s; code=5).

Parameter
pressure FLOAT Pressure in bar (constant value assigned to all levels)

2.15.17. RHOPOT - Calculates potential density

Synopsis

rhopot[,pressure] infile outfile

Description
This is a special operator for the post processing of the ocean and sea ice model [MPIOM]. It calculates
the sea water potential density (name=rhopoto; code=18). Required input fields are sea water in-situ
temperature (name=to; code=20) and sea water salinity (name=sao; code=5). Pressure is calculated
from the level information or can be specified by the optional parameter.

Parameter
pressure FLOAT Pressure in bar (constant value assigned to all levels)

Example
To compute the sea water potential density from the potential temperature use this operator in
combination with adisit:
cdo rhopot -adisit infile outfile

229
Miscellaneous Reference manual

2.15.18. HISTOGRAM - Histogram

Synopsis

< operator >,bounds infile outfile

Description
This module creates bins for a histogram of the input data. The bins have to be adjacent and have
non-overlapping intervals. The user has to define the bounds of the bins. The first value is the lower
bound and the second value the upper bound of the first bin. The bounds of the second bin are
defined by the second and third value, aso. Only 2-dimensional input fields are allowed. The output
file contains one vertical level for each of the bins requested.

Operators

histcount Histogram count

Number of elements in the bin range.
histsum Histogram sum
Sum of elements in the bin range.
histmean Histogram mean
Mean of elements in the bin range.
histfreq Histogram frequency
Relative frequency of elements in the bin range.

Parameter
bounds FLOAT Comma-separated list of the bin bounds (-inf and inf valid)

2.15.19. SETHALO - Set the bounds of a field

Synopsis

sethalo[,parameter] infile outfile

Description
This operator sets the boundary in the east, west, south and north of the rectangular understood
fields. Positive values of the parameters increase the boundary in the selected direction. Negative
values decrease the field at the selected boundary. The new rows and columns are filled with the
missing value. With the optional parameter value a different fill value can be used. Global cyclic
fields are filled cyclically at the east and west borders, if the fill value is not set by the user.

Parameter
east INTEGER East halo
west INTEGER West halo
south INTEGER South halo
north INTEGER North halo
value FLOAT Fill value (default is the missing value)

230
Reference manual Miscellaneous

2.15.20. WCT - Windchill temperature

Synopsis

wct infile1 infile2 outfile

Description
Let infile1 and infile2 be time series of temperature and wind speed records, then a corresponding
time series of resulting windchill temperatures is written to outfile. The wind chill temperature cal-
culation is only valid for a temperature of T <= 33 ℃ and a wind speed of v >= 1.39 m/s. Whenever
these conditions are not satisfied, a missing value is written to outfile. Note that temperature and
wind speed records have to be given in units of ℃ and m/s, respectively.

2.15.21. FDNS - Frost days where no snow index per time period

Synopsis

fdns infile1 infile2 outfile

Description
Let infile1 be a time series of the daily minimum temperature TN and infile2 be a corresponding
series of daily surface snow amounts. Then the number of days where TN < 0 ℃ and the surface snow
amount is less than 1 cm is counted. The temperature TN have to be given in units of Kelvin. The
date information of a timestep in outfile is the date of the last contributing timestep in infile.

2.15.22. STRWIN - Strong wind days index per time period

Synopsis

strwin[,v] infile outfile

Description
Let infile be a time series of the daily maximum horizontal wind speed VX, then the number of
days where VX > v is counted. The horizontal wind speed v is an optional parameter with default v
= 10.5 m/s. A further output variable is the maximum number of consecutive days with maximum
wind speed greater than or equal to v. Note that both VX and v have to be given in units of m/s.
Also note that the horizontal wind speed is defined as the square root of the sum of squares of the
zonal and meridional wind speeds. The date information of a timestep in outfile is the date of the
last contributing timestep in infile.

Parameter
v FLOAT Horizontal wind speed threshold (m/s, default v = 10.5 m/s)

231
Miscellaneous Reference manual

2.15.23. STRBRE - Strong breeze days index per time period

Synopsis

strbre infile outfile

Description
Let infile be a time series of the daily maximum horizontal wind speed VX, then the number of
days where VX is greater than or equal to 10.5 m/s is counted. A further output variable is the
maximum number of consecutive days with maximum wind speed greater than or equal to 10.5 m/s.
Note that VX is defined as the square root of the sum of squares of the zonal and meridional wind
speeds and have to be given in units of m/s. The date information of a timestep in outfile is the
date of the last contributing timestep in infile.

2.15.24. STRGAL - Strong gale days index per time period

Synopsis

strgal infile outfile

Description
Let infile be a time series of the daily maximum horizontal wind speed VX, then the number of
days where VX is greater than or equal to 20.5 m/s is counted. A further output variable is the
maximum number of consecutive days with maximum wind speed greater than or equal to 20.5 m/s.
Note that VX is defined as the square root of the sum of square of the zonal and meridional wind
speeds and have to be given in units of m/s. The date information of a timestep in outfile is the
date of the last contributing timestep in infile.

2.15.25. HURR - Hurricane days index per time period

Synopsis

hurr infile outfile

Description
Let infile be a time series of the daily maximum horizontal wind speed VX, then the number of
days where VX is greater than or equal to 32.5 m/s is counted. A further output variable is the
maximum number of consecutive days with maximum wind speed greater than or equal to 32.5 m/s.
Note that VX is defined as the square root of the sum of squares of the zonal and meridional wind
speeds and have to be given in units of m/s. The date information of a timestep in outfile is the
date of the last contributing timestep in infile.

232
Reference manual Miscellaneous

2.15.26. CMORLITE - CMOR lite

Synopsis

cmorlite,table[,convert] infile outfile

Description
The [CMOR] (Climate Model Output Rewriter) library comprises a set of functions, that can be
used to produce CF-compliant NetCDF files that fulfill the requirements of many of the climate
community’s standard model experiments. These experiments are collectively referred to as MIP’s.
Much of the metadata written to the output files is defined in MIP-specific tables, typically made
available from each MIP’s web site.
The CDO operator cmorlite process the header and variable section of such MIP tables and writes
the result with the internal IO library [CDI]. In addition to the CMOR 2 and 3 table format, the
CDO parameter table format is also supported. The following parameter table entries are available:

Entry Type Description

name WORD Name of the variable
out_name WORD New name of the variable
type WORD Data type (real or double)
standard_name WORD As defined in the CF standard name table
long_name STRING Describing the variable
units STRING Specifying the units for the variable
comment STRING Information concerning the variable
cell_methods STRING Information concerning calculation of means or climatologies
cell_measures STRING Indicates the names of the variables containing cell areas and volumes
missing_value FLOAT Specifying how missing data will be identified
valid_min FLOAT Minimum valid value
valid_max FLOAT Maximum valid value
ok_min_mean_abs FLOAT Minimum absolute mean
ok_max_mean_abs FLOAT Maximum absolute mean
factor FLOAT Scale factor
delete INTEGER Set to 1 to delete variable
convert INTEGER Set to 1 to convert the unit if necessary

Most of the above entries are stored as variables attributes, some of them are handled differently. The
variable name is used as a search key for the parameter table. valid_min, valid_max, ok_min_mean_abs
and ok_max_mean_abs are used to check the range of the data.

Parameter
table STRING Name of the CMOR table as specified from PCMDI
convert STRING Converts the units if necessary

Example
Here is an example of a parameter table for one variable:
prompt> cat mypartab
&parameter
name = t

233
Miscellaneous Reference manual

out_name = ta
standard_name = air_temperature
units = "K"
missing_value = 1.0e+20
valid_min = 157.1
valid_max = 336.3
/

To apply this parameter table to a dataset use:

cdo -f nc cmorlite,mypartab,convert infile outfile

This command renames the variable t to ta. The standard name of this variable is set to air_temperature
and the unit is set to [K] (converts the unit if necessary). The missing value will be set to 1.0e+20.
In addition it will be checked whether the values of the variable are in the range of 157.1 to 336.3.
The result will be stored in NetCDF.

2.15.27. VERIFYGRID - Verify grid coordinates

Synopsis

verifygrid infile

Description
This operator verifies the coordinates of all horizontal grids found in infile. Among other things,
it searches for duplicate cells, non-convex cells, and whether the center is located outside the cell
bounds. Use the CDO option -v to output the position of these cells. This information can be useful
to avoid problems when interpolating the data.

234
Reference manual Miscellaneous

2.15.28. HEALPIX - Change healpix resolution

Synopsis

< operator >,parameter infile outfile

Description
Degrade or upgrade the resolution of a healpix grid.

Operators

hpdegrade Degrade healpix

Degrade the resolution of a healpix grid. The value of the target pixel is the mean
of the source pixels.
hpupgrade Upgrade healpix
Upgrade the resolution of a healpix grid. The values of the target pixels is the value
of the source pixel.

Parameter
nside INTEGER The nside of the target healpix, must be a power of two [default: same as
input].
order STRING Pixel ordering of the target healpix (’nested’ or ’ring’).
power FLOAT If non-zero, divide the result by (nside[in]/nside[out])**power. power=-2 keeps
the sum of the map invariant.

235
3. Contributors

3.1. History
CDO was originally developed by Uwe Schulzweida at the Max Planck Institute for Meteorology (MPI-M).
The first public release is available since 2003. The MPI-M, together with the DKRZ, has a long history
in the development of tools for processing climate data. CDO was inspired by some of these tools, such as
the PINGO package and the GRIB-Modules.
PINGO1 was developed by Jürgen Waszkewitz, Peter Lenzen, and Nathan Gillet in 1995 at the DKRZ,
Hamburg (Germany). CDO has a similar user interface and uses some of the PINGO routines.
The GRIB-Modules was developed by Heiko Borgert and Wolfgang Welke in 1991 at the MPI-M. CDO is
using a similar module structure and also some of the routines.

3.2. External sources

CDO has incorporated code from several sources:

afterburner is a postprocessing application for ECHAM data and ECMWF analysis data, originally de-
veloped by Edilbert Kirk, Michael Ponater and Arno Hellbach. The afterburner code was modified
for the CDO operators after, ml2pl, ml2hl, sp2gp, gp2sp.
SCRIP is a software package used to generate interpolation weights for remapping fields from one grid to
another in spherical geometry [SCRIP]. It was developed at the Los Alamos National Laboratory by
Philip W. Jones. The SCRIP library was converted from Fortran to ANSI C and is used as the base
for the remapping operators in CDO.
YAC (Yet Another Coupler) was jointly developed by DKRZ and MPI-M by Moritz Hanke and Rene
Redler [YAC]. CDO is using the clipping and cell search routines for the conservative remapping
with remapcon.
libkdtree a C99 implementation of the kd-tree algorithm developed by Jörg Dietrich.

CDO uses tools from the GNU project, including automake, and libtool.

3.3. Contributors
The primary contributors to the CDO development have been:

Uwe Schulweida : Concept, design and implementation of CDO, project coordination, and releases.
Luis Kornblueh : He supports CDO from the beginning. His main contributions are GRIB performance
and compression, GME and unstructured grid support. Luis also helps with design and planning.
Ralf Müller : He is working on CDO since 2009. His main contributions are the implementation of the
User Portal, the ruby and python interface for all CDO operators, the building process and the
Windows support. The CDO User Portal was funded by the European Commission infracstructure
project IS-ENES. Ralf also helps a lot with the user support. Implemented operators: intlevel3d,
consecsum, consects, ngrids, ngridpoints, reducegrid
1 Procedural INterface for GRIB formatted Objects

236
Contributors Contributors

Cedrick Ansorge : He worked on the software package CDO as a student assistant at MPI-M from 2007-
2011. Implemented operators: eof, eof3d, enscrps, ensbrs, maskregion, bandpass, lowpass, highpass,
smooth9
Oliver Heidmann : He worked on the software package CDO as a student assistant at MPI-M from 2015-
2018.
Karin Meier-Fleischer : She is working in the CDO user support since 2017.
Fabian Wachsmann : He is working on CDO for the CMIP6 project since 2016. His main task is the
implementation and support of the cmor operator. He has also implemented the ETCCDI Indices of
Daily Temperature and Precipitation Extremes.
Ralf Quast : He worked on CDO on behalf of the Service Gruppe Anpassung (SGA), DKRZ in 2006.
He implemented all ECA Indices of Daily Temperature and Precipitation Extremes, all percentile
operators, module YDRUNSTAT and wct.
Kameswarrao Modali : He worked on CDO from 2012-2013.
Implemented operators: contour, shaded, grfill, vector, graph.
Michal Koutek : Implemented operators: selmulti delmulti, changemulti, samplegrid, uvDestag, rotu-
vNorth, projuvLatLon.
Etienne Tourigny : Implemented operators: setclonlatbox, setcindexbox, setvals, splitsel, histfreq, setrtoc,
setrtoc2.
Karl-Hermann Wieners : Implemented operators: aexpr, aexprf, selzaxisname.
Asela Rajapakse : He worked on CDO from 2016-2017 as part of the EUDAT project.
Implemented operator: verifygrid
Estanislao Gavilan : Improved the CDO documentation for the installation section.

Many users have contributed to CDO by sending bug reports, patches and suggestions over time. Very
helpful is also the active participation in the user forum of some users. Here is an incomplete list:
Jaison-Thomas Ambadan, Harald Anlauf, Andy Aschwanden, Stefan Bauer, Simon Blessing,
Renate Brokopf, Michael Boettinger, Tim Brücher, Reinhard Budich, Martin Claus,
Traute Crüger, Brendan de Tracey, Irene Fischer-Bruns, Chris Fletscher, Helmut Frank,
Kristina Fröhlich, Oliver Fuhrer, Monika Esch, Pier Giuseppe Fogli, Beate Gayer,
Veronika Gayler, Marco Giorgetta, David Gobbett, Holger Goettel, Helmut Haak,
Stefan Hagemann, Angelika Heil, Barbara Hennemuth, Daniel Hernandez, Nathanael Huebbe,
Thomas Jahns, Frank Kaspar, Daniel Klocke, Edi Kirk, Yvonne Küstermann,
Stefanie Legutke, Leonidas Linardakis, Stephan Lorenz, Frank Lunkeit, Uwe Mikolajewicz,
Laura Niederdrenk, Dirk Notz, Hans-Jürgen Panitz, Ronny Petrik, Swantje Preuschmann,
Florian Prill, Asela Rajapakse, Daniel Reinert, Hannes Reuter, Mathis Rosenhauer,
Reiner Schnur, Martin Schultz, Dennis Shea, Kevin Sieck, Martin Stendel,
Bjorn Stevens, Martina Stockhaus, Claas Teichmann, Adrian Tompkins, Jörg Trentmann,
Álvaro M. Valdebenito, Geert Jan van Oldenborgh, Jin-Song von Storch, David Wang,
Joerg Wegner, Heiner Widmann, Claudia Wunram, Klaus Wyser

Please let me know if your name was omitted!

237
Bibliography
[BitInformation.jl]
M Klöwer, M Razinger, JJ Dominguez, PD Düben and TN Palmer, 2021. Compressing atmospheric
data into its real information content. Nature Computational Science 1, 713–724. 10.1038/s43588-021-
00156-2
[CDI]
Climate Data Interface, from the Max Planck Institute for Meteorologie
[CM-SAF]
Satellite Application Facility on Climate Monitoring, from the German Weather Service (Deutscher
Wetterdienst, DWD)
[CMOR]
Climate Model Output Rewriter, from the Program For Climate Model Diagnosis and Intercomparison
(PCMDI)
[ecCodes]
API for GRIB decoding/encoding, from the European Centre for Medium-Range Weather Forecasts
(ECMWF)
[ECHAM]
The atmospheric general circulation model ECHAM5, from the Max Planck Institute for Meteorologie
[GMT]
The Generic Mapping Tool, from the School of Ocean and Earth Science and Technology (SOEST)
[GrADS]
Grid Analysis and Display System, from the Center for Ocean-Land-Atmosphere Studies (COLA)
[GRIB]
GRIB version 1, from the World Meteorological Organisation (WMO)
[HDF5]
HDF version 5, from the HDF Group
[INTERA]
INTERA Software Package, from the Max Planck Institute for Meteorologie
[Magics]
Magics Software Package, from the European Centre for Medium-Range Weather Forecasts (ECMWF)
[MPIOM]
Ocean and sea ice model, from the Max Planck Institute for Meteorologie
[NetCDF]
NetCDF Software Package, from the UNIDATA Program Center of the University Corporation for
Atmospheric Research
[PINGO]
The PINGO package, from the Model & Data group at the Max Planck Institute for Meteorologie
[REMO]
Regional Model, from the Max Planck Institute for Meteorologie
[Preisendorfer]
Rudolph W. Preisendorfer: Principal Component Analysis in Meteorology and Oceanography, Elsevier
(1988)

238
Bibliography Bibliography

[PROJ]
Cartographic Projections Library, originally written by Gerald Evenden then of the USGS.
[SCRIP]
SCRIP Software Package, from the Los Alamos National Laboratory
[szip]
Szip compression software, developed at University of New Mexico.
[vonStorch]
Hans von Storch, Walter Zwiers: Statistical Analysis in Climate Research, Cambridge University Press
(1999)
[YAC]
YAC - Yet Another Coupler Software Package, from DKRZ and MPI for Meteorologie

239
A. Environment Variables
The following table describes the environment variables that affect CDO.

Variable name Default Description

CDO_DOWNLOAD_PATH None Path where CDO can store downloads.
CDO_FILE_SUFFIX None Default filename suffix. This suffix will be added to the output file
name instead of the filename extension derived from the file
format. NULL will disable the adding of a file suffix.
CDO_GRIDSEARCH_RADIUS 180 Grid search radius in degree. Used by the operators
setmisstonn, remapdis and remapnn.
CDO_HISTORY_INFO true ’false’ don’t write information to the global history attribute.
CDO_ICON_GRIDS None Root directory of the installed ICON grids (e.g. /pool/data/ICON).
CDO_PCTL_NBINS 101 Number of histogram bins.
CDO_RESET_HISTORY false ’true’ resets the global history attribute.
CDO_REMAP_NORM fracarea Choose the normalization for the conservative interpolation
CDO_TIMESTAT_DATE None Set target timestamp of a temporal statistic operator to the "first",
"middle", "midhigh" or "last" contributing source timestep.
CDO_USE_FFTW 1 Set to 0 to switch off usage of FFTW. Used in the Filter module.
CDO_VERSION_INFO true ’false’ disables the global NetCDF attribute CDO.

240
B. Parallelized operators
Some of the CDO operators are parallelized with OpenMP. To use CDO with multiple OpenMP threads,
you have to set the number of threads with the option ’-P’. Here is an example to distribute the bilinear
interpolation on 8 OpenMP threads:
cdo -P 8 remapbil,targetgrid infile outfile

The following CDO operators are parallelized with OpenMP:

Module Operator Description

Afterburner after ECHAM standard post processor
Detrend detrend Detrend
EcaEtccdi etccdi_tx90p % of days when daily max temperature is > the 90th percentile
EcaEtccdi etccdi_tx10p % of days when daily max temperature is < the 10th percentile
EcaEtccdi etccdi_tn90p % of days when daily min temperature is > the 90th percentile
EcaEtccdi etccdi_tn10p % of days when daily min temperature is < the 10th percentile
EcaEtccdi etccdi_r95p Annual tot precip when daily precip exceeds the 95th percentile of ...
EcaEtccdi etccdi_r99p Annual tot precip when daily precip exceeds the 99th percentile of ...
Ensstat ens<STAT> Statistical values over an ensemble
EOF eof Empirical Orthogonal Functions
Fillmiss setmisstonn Set missing value to nearest neighbor
Fillmiss setmisstodis Set missing value to distance-weighted average
Filter bandpass Bandpass filtering
Filter lowpass Lowpass filtering
Filter highpass Highpass filtering
Fourier fourier Fourier transformation
Genweights genbil Generate bilinear interpolation weights
Genweights genbic Generate bicubic interpolation weights
Genweights gendis Generate distance-weighted average remap weights
Genweights gennn Generate nearest neighbor remap weights
Genweights gencon Generate 1st order conservative remap weights
Genweights gencon2 Generate 2nd order conservative remap weights
Genweights genlaf Generate largest area fraction remap weights
Gridboxstat gridbox<STAT> Statistical values over grid boxes
Intlevel intlevel Linear level interpolation
Intlevel3d intlevel3d Linear level interpolation from/to 3D vertical coordinates
Remapeta remapeta Remap vertical hybrid level
Remap remapbil Bilinear interpolation
Remap remapbic Bicubic interpolation
Remap remapdis Distance-weighted average remapping
Remap remapnn Nearest neighbor remapping
Remap remapcon First order conservative remapping
Remap remapcon2 Second order conservative remapping
Remap remaplaf Largest area fraction remapping
Smooth smooth Smooth grid points
Spectral sp2gp, gp2sp Spectral transformation

241
 Parallelized operators

Module Operator Description

Vertintap ap2pl, ap2hl Vertical interpolation on hybrid sigma height coordinates
Vertintgh gh2hl Vertical height interpolation
Vertintml ml2pl, ml2hl Vertical interpolation on hybrid sigma pressure coordinates

242
C. Standard name table
The following CF standard names are supported by CDO.

CF standard name Units GRIB 1 code variable name

surface_geopotential m2 s-2 129 geosp
air_temperature K 130 ta
specific_humidity 1 133 hus
surface_air_pressure Pa 134 aps
air_pressure_at_sea_level Pa 151 psl
geopotential_height m 156 zg

243
D. Grid description examples

D.1. Example of a curvilinear grid description

Here is an example for the CDO description of a curvilinear grid. xvals/yvals describe the positions of the
6x5 quadrilateral grid cells. The first 4 values of xbounds/ybounds are the corners of the first grid cell.
gridtype = curvilinear
gridsize = 30
xsize = 6
ysize = 5
xvals = −21 −11 0 11 21 30 −25 −13 0 13
25 36 −31 −16 0 16 31 43 −38 −21
0 21 38 52 −51 −30 0 30 51 64
xbounds = −23 −14 −17 −28 −14 −5 −6 −17 −5 5 6 −6
5 14 17 6 14 23 28 17 23 32 38 28
−28 −17 −21 −34 −17 −6 −7 −21 −6 6 7 −7
6 17 21 7 17 28 34 21 28 38 44 34
−34 −21 −27 −41 −21 −7 −9 −27 −7 7 9 −9
7 21 27 9 21 34 41 27 34 44 52 41
−41 −27 −35 −51 −27 −9 −13 −35 −9 9 13 −13
9 27 35 13 27 41 51 35 41 52 63 51
−51 −35 −51 −67 −35 −13 −21 −51 −13 13 21 −21
13 35 51 21 35 51 67 51 51 63 77 67
yvals = 29 32 32 32 29 26 39 42 42 42
39 35 48 51 52 51 48 43 57 61
62 61 57 51 65 70 72 70 65 58
ybounds = 23 26 36 32 26 27 37 36 27 27 37 37
27 26 36 37 26 23 32 36 23 19 28 32
32 36 45 41 36 37 47 45 37 37 47 47
37 36 45 47 36 32 41 45 32 28 36 41
41 45 55 50 45 47 57 55 47 47 57 57
47 45 55 57 45 41 50 55 41 36 44 50
50 55 64 58 55 57 67 64 57 57 67 67
57 55 64 67 55 50 58 64 50 44 51 58
58 64 72 64 64 67 77 72 67 67 77 77
67 64 72 77 64 58 64 72 58 51 56 64

Figure D.1.: Orthographic and Robinson projection of the curvilinear grid, the first grid cell is colored red

244
Grid description examples Example description for an unstructured grid

D.2. Example description for an unstructured grid

Here is an example of the CDO description for an unstructured grid. xvals/yvals describe the positions
of 30 independent hexagonal grid cells. The first 6 values of xbounds/ybounds are the corners of the first
grid cell. The grid cell corners have to rotate counterclockwise. The first grid cell is colored red.
gridtype = unstructured
gridsize = 30
nvertex = 6
xvals = −36 36 0 −18 18 108 72 54 90 180 144 126 162 −108 −144
−162 −126 −72 −90 −54 0 72 36 144 108 −144 180 −72 −108 −36
xbounds = 339 0 0 288 288 309 21 51 72 72 0 0
0 16 21 0 339 344 340 0 −0 344 324 324
20 36 36 16 0 0 93 123 144 144 72 72
72 88 93 72 51 56 52 72 72 56 36 36
92 108 108 88 72 72 165 195 216 216 144 144
144 160 165 144 123 128 124 144 144 128 108 108
164 180 180 160 144 144 237 267 288 288 216 216
216 232 237 216 195 200 196 216 216 200 180 180
236 252 252 232 216 216 288 304 309 288 267 272
268 288 288 272 252 252 308 324 324 304 288 288
345 324 324 36 36 15 36 36 108 108 87 57
20 15 36 57 52 36 108 108 180 180 159 129
92 87 108 129 124 108 180 180 252 252 231 201
164 159 180 201 196 180 252 252 324 324 303 273
236 231 252 273 268 252 308 303 324 345 340 324
yvals = 58 58 32 0 0 58 32 0 0 58 32 0 0 58 32
0 0 32 0 0 −58 −58 −32 −58 −32 −58 −32 −58 −32 −32
ybounds = 41 53 71 71 53 41 41 41 53 71 71 53
11 19 41 53 41 19 −19 −7 11 19 7 −11
−19 −11 7 19 11 −7 41 41 53 71 71 53
11 19 41 53 41 19 −19 −7 11 19 7 −11
−19 −11 7 19 11 −7 41 41 53 71 71 53
11 19 41 53 41 19 −19 −7 11 19 7 −11
−19 −11 7 19 11 −7 41 41 53 71 71 53
11 19 41 53 41 19 −19 −7 11 19 7 −11
−19 −11 7 19 11 −7 11 19 41 53 41 19
−19 −7 11 19 7 −11 −19 −11 7 19 11 −7
−41 −53 −71 −71 −53 −41 −53 −71 −71 −53 −41 −41
−19 −41 −53 −41 −19 −11 −53 −71 −71 −53 −41 −41
−19 −41 −53 −41 −19 −11 −53 −71 −71 −53 −41 −41
−19 −41 −53 −41 −19 −11 −53 −71 −71 −53 −41 −41
−19 −41 −53 −41 −19 −11 −19 −41 −53 −41 −19 −11

Figure D.2.: Orthographic and Robinson projection of the unstructured grid

245
Index

A daypctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
abs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 dayrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
acos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 daystd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
add . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 daystd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
addc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 daysub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
addtrend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 daysum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
adipot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 dayvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
adisit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 dayvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
aexpr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 delcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
aexprf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 delete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
after . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 delgridcell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
ap2pl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 delmulti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
apply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 delname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
asin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 delparam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
atan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 deltat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
atan2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 detrend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
dhouravg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
B dhourmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
bandpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 dhourmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
bitrounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 dhourmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
bottomvalue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 dhourrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
dhourstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
C dhourstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
cat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 dhoursum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
changemulti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 dhourvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
chcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 dhourvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
chlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 diff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
chlevelc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 diffn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
chlevelv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 distgrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
chname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 div . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
chparam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 divc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
chunit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 divcoslat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
clone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 divdpm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
cmorlite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 divdpy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
codetab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 duplicate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
collgrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 dv2ps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
consecsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 dv2uv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
consects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
const . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 E
copy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 enlarge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
cos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ensavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
ensbrs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
D enscrps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
dayadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 enskurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
dayavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 ensmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
daydiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 ensmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
daymax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 ensmedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
daymean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 ensmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
daymin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 enspctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
daymul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 ensrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

246
Index Index

ensrkhistspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 gradsdes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

ensrkhisttime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 gridarea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
ensroc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 gridboxavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
ensskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxkurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
ensstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
ensstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
enssum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxmedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
ensvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
ensvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 gridboxrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 gridboxskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eof3d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 gridboxstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eofcoeff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 gridboxstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eofspatial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 gridboxsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eoftime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 gridboxvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 gridboxvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
eqc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 gridcellindex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
exp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 griddes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
expr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 gridweights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
exprf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 gt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
gtc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
F
fdns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 H
fldavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 highpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
fldcor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 histcount . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
fldcount . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 histfreq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
fldcovar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 histmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
fldint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 histsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
fldkurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 houravg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldmedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
fldpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hoursum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hourvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
fldvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hpdegrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
fldvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 hpupgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
fourier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 hurr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

G I
ge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 ifnotthen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
gec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 ifnotthenc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
genbic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 ifthen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
genbil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 ifthenc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
gencon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 ifthenelse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
gencon2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 import_amsr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
gendis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 import_binary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
genlaf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 import_cmsaf . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
genlevelbounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
gennn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 infon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
gh2hl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
gheight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 inputext . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
gmtcells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 inputsrv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
gmtxyz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 int . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
gp2sp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 intlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

247
Index Index

intlevel3d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 monsub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

intlevelx3d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 monsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
intntime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 monvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
inttime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 monvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
intyear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 mrotuvb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
invertlat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 mul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
invertlev . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 mulc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
isosurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 mulcoslat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
muldpm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
L muldpy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
le . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
lec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 N
ln . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ndate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
log10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
lowpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 nec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
lt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 ngridpoints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
ltc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 ngrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
nint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
M nlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 nmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
maskindexbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
masklonlatbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 npar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
maskregion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 ntime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
mastrfu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 nyear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
max . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
maxc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 O
meravg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
merge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 outputext . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
mergegrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 outputf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
mergetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 outputint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
merkurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 outputsrv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
mermax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 outputtab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
mermean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
mermedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 P
mermin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 pack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
merpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 partab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
merrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 pow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
merskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 projuvLatLon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
merstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
merstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 R
mersum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 random . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
mervar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 reci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
mervar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 reducegrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
min . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 regres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
minc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 remap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
ml2hl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 remapavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
ml2pl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 remapbic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
monadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 remapbil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
monavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapcon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
mondiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 remapcon2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
monmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapdis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
monmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapeta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
monmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapkurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
monmul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 remaplaf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
monpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 remapmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
monrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
monstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapmedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
monstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 remapmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

248
Index Index

remapnn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 selstdname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

remaprange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 seltabnum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
remapskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 seltime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
remapstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 seltimestep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
remapstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 selyear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
remapsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 selyearidx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
remapvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 selzaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
remapvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 selzaxisname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
replace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 seq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
rhopot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 setattribute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
rotuvb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 setcalendar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
rotuvNorth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 setcindexbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
runavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setclonlatbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
runmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
runmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setcodetab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
runmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setctomiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
runpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 setdate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
runrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setday . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
runstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setgrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
runstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setgridarea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
runsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setgridcell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
runvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setgridmask . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
runvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 setgridtype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
sethalo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
S setlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
samplegrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 setltype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
sealevelpressure . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 setmaxsteps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
seasavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setmisstoc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
seasmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setmisstodis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
seasmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setmisstonn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
seasmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setmissval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
seaspctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 setmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
seasrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
seasstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setparam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
seasstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setpartabn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
seassum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setpartabp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
seasvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setreftime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
seasvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 setrtoc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
selcircle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 setrtoc2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
selcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 setrtomiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
seldate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 settaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
selday . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 settbounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
select . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 settime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
selgrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 settunits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
selgridcell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 setunit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
selhour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 setvals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
selindexbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 setvrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
sellevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 setyear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
sellevidx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 setzaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
sellonlatbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 shifttime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
selltype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 shiftx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
selmonth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 shifty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
selmulti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 showattribute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
selname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 showcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
selparam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 showdate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
selregion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 showformat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
selseason . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 showlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
selsmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 showltype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

249
Index Index

showmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timselstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

showname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timselsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
showstdname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timselvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
showtime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timselvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
showtimestamp . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timsort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
showyear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 timstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
sin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 timstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
sinfo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 timsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
sinfon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 timvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
smooth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 timvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
smooth9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 topo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
sp2gp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 topvalue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
sp2sp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 trend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
splitcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
splitdate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 U
splitday . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 unpack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
splitgrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 uv2dv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
splithour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 uvDestag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
splitlevel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
splitmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
V
varsavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splitname . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
varsmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splitparam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
varsmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splitseas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
varsmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splitsel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
varsrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splittabnum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
varsstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splityear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
varsstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splityearmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
varssum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
splitzaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
varsvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
sqr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
varsvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
sqrt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
vct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
stdatm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
verifygrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
strbre . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
vertavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
strgal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
vertfillmiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
strwin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
vertmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
sub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
vertmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
subc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
vertmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
subtrend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
vertrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
T vertstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
tan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 vertstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
tee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 vertsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
timavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 vertvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
timcor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 vertvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
timcovar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
timcumsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
W
wct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
timfillmiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
timmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 X
timmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 xsinfo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
timmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 xsinfop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
timpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
timrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Y
timselavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydayadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
timselmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydayavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
timselmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydaydiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
timselmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydaymax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
timselpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 ydaymean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
timselrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydaymin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
timselstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 ydaymul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

250
Index Index

ydaypctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 ymonmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

ydayrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydaystd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydaystd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonmul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
ydaysub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 ymonne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
ydaysum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
ydayvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydayvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 ymonstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydrunavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 ymonstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydrunmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 ymonsub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
ydrunmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 ymonsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydrunmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 ymonvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydrunpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 ymonvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ydrunstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 yseasadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
ydrunstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 yseasavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
ydrunsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 yseasdiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
ydrunvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 yseasmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
ydrunvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 yseasmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 yseasmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseasmul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
yeardiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 yseaspctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
yearmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseasrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearmaxidx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseasstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseasstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseassub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
yearminidx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 yseassum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearmonmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 yseasvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearmul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 yseasvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
yearpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
yearrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Z
yearstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 zaxisdes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
yearstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 zonavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yearsub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 zonkurt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yearsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 zonmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yearvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 zonmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yearvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 zonmedian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhouradd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 zonmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhouravg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonpctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourdiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 zonrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonskew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourmean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourmul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 zonsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourrange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourstd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 zonvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
yhourstd1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
yhoursub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
yhoursum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
yhourvar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
yhourvar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
ymonadd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
ymonavg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
ymondiv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
ymoneq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
ymonge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
ymongt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
ymonle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
ymonlt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

251