J. Indian Water Resour. Soc.

,
Vol. 40, No. 3 & 4, Jul. & Oct., 2020
CLIMATE TRENDS OVER GOA: STATISTICAL ANALYSIS OF
OBSERVATIONS AND REANALYSIS DATASETS FOR BETTER WATER
MANAGEMENT
Ashwini Pai Panandiker1, Lewlynnde Mello2, B. Venkatesh3, Mahender Kotha4, A.G. Chachadi5
ABSTRACT
The study examines the significance and magnitude of trends in the monthly rainfall and daily temperatures in Goa for a period of 33 years (1980 – 2012). 11
IMD rain gauges, 3 Climate Forecast System Reanalysis (CFSR) monitoring points and 6 Water Resources Department (WRD) operated rain gauges were
selected for the study. 2 IMD weather monitoring stations, 3 CFSR points and 1 WRD monitoring station was used for the temperature data. Coefficient of
Variation (CV), Detrended Fluctuation analysis, seasonal Kendall Test, and Sen’s slope were used for statistical analysis. Percentage departure was
calculated for average monthly values separately for three decades. The CV shows that there is an increase in the annual rainfall during the post-monsoon and
pre-monsoon season indicating a shift in the rainfall pattern. From the trend analysis carried out it is evident that since 1980 till 2012, there was a minor
increase (~7-8mm/year) in the rainfall, however more recently (2009-2017) there is a decrease in the annual rainfall. An increase in temperature by
0.03°C/year has been observed. Long-range dependence analysis revealed a weak persistence for both rainfall and temperature. Results can be used for
projecting future trends, hydrological modeling and formulation ofadaptation strategies.

Keywords: Historical climate trends, statistical analysis, adaptation, water resource management

This study was done for the state of Goa which is located on
INTRODUCTION the west coast of India surrounded by the Arabian Sea on
Worldwide, several studies were engaged in assessing the the west and Western Ghats on the eastern side. The rivers
historical trends and variability in hydro-meteorological of Goa originate from the Western Ghats. These rivers carry
data, mainly rainfall and temperature, at regional levels as a large amount of water due to monsoonal rains, and all the
well as country levels (Raju et al., 2017, Jain and Kumar activities depend on the water availability in these rivers.
2012, Singh et al., 2005, Rupa et al., 1992, Rathore et al., Any change in the rainfall pattern would have an impact on
2013, Suprit and Shankar 2008). These studies provide an the agricultural and other activities (mainly tourism). Goa
insight into the climatic pattern; assist in hydrological has a resident population of 14.5 lakhs but has to also
modeling and thus aid in creation of the possible future provide water to additional 28-30 lakh tourists that visit the
scenarios. A common challenge in hydrological modeling is State every year. Despite of large amount of rainfall during
obtaining accurate weather input data (Kouwen et al., the monsoon months (June to Sept), many areas in Goa face
2005).When observed data is not available, satellite based a shortage of water. Therefore, it is necessary to understand
global datasets such as the Climate Forecast System the pattern of rainfall and their variability in the state for
Reanalysis (CFSR) are used. The validation or comparison better planning and utilization of water. It would also help
between these datasets is essential before using them for in future projections and hydrological modeling at a
future simulations(Kishore et al., 2015). Primarily the catchment level.
political and technical adaptation measures to climate Previously, there are only few studies have been carried out
change are taken at local level. Hence, local or regional to analyse spatiotemporal variability of precipitation.
studies are of paramount importance. Nandargi and Mulye (2014) studied the monthly rainfall
data from 1901-2012 for its spatiotemporal analysis, trend
1. The Energy and Resources Institute, St Cruz, Goa - 403 202& and mean intensity. The authors have reported that, in the
Earth Science Department, Goa University, Goa – 403 206
recent time period (2001-2012) the heavy precipitation
Corresponding Author Email: [email protected],
[email protected] events have increased compared to that of the earlier period.
2. The Energy and Resources Institute, St Cruz, Goa - 403 202; Further, it is reported that, the Monsoon Precipitation Index
Email: [email protected] (MPI) vary between 0.7 to 1.0 in both the district and are
3. National Institute of Hydrology, Belgaum, Karnataka - 590 019 influenced by the topographical variations. While studying
Email: [email protected] the cause of the extreme rainfall events of the recent times,
4. Head of the Department, Earth Science Department, Goa Metri and Singh(2010) observed that, active off-shore
University, Goa – 403 206 trough and low-pressure system in the southeast Arabian
Email: [email protected]
sea and orography play an important role for such extreme
5. Retired Professor, Earth Sciences Department, Goa University, rainfall events to happen. However, Preeti et al., (2017)
Goa - 403 206
Email: [email protected] reported an increasing trend of rainfall in the northern part
Manuscript No. 1530 of the Western Ghats (covering Goa and Konkan region). A
similar observation was reported by Ravadekar et al.,
10
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

(2018), while studying the variation of rainfall across the and satellite-derived products. At each analysis hour, the
entire Western Ghats of India and observed that, the rainfall CFSR includes both the forecast data, predicted from the
is showing increasing trend in the Goa and Konkan region. previous analysis hour, and the data from the analysis
All these studies have focused more on the statistical utilized to reinitialize the forecast models. The horizontal
analysis of the rainfall using few point measurements of resolution of the CFSR is 38 km (Saha et al., 2010). Three
Goa or the grid based data at 1o by 1o provided by the IMD. CFSR monitoring points were available for Goa at
As suggested by IPCC(2014) in its AR6 report, it is always Cavorem, Usagao and at the mouth of River Mandovi. The
beneficial to use the local (point measurement) rainfall data locations of the rainfall monitoring stations operated by
for analysis to gain more insight and information about various agencies are shown in Figure 1. The data was
their spatio-temporal variability, which is very important for obtained from these locations were used in this study. The
planning and development of water resources of any statistical methods used for the study are described in the
region. Keeping this in view, the present study is planned to next section.
use the point measured rainfall data of Goa state (by IMD,
WRD and from other sources) to understand their
spatiotemporal variability across the state.
Therefore, a study was initiated with specific objectives
was (i) to investigate the rainfall variability across the state
using the data obtained from three different organizations,
Water Resources Department (WRD), Indian
Meteorological Department (IMD) and the Climate
Forecast System Reanalysis (CFSR), and examine the trend
of the annual and seasonal rainfall, (ii) to identify the
influence of climate change in terms of behavioral change
of rainfall and temperature using the past 33 years data.
Since Goa is a very small state with an area of 3702 km2,
this analysis was done at State level instead of watershed
level. This study aimed to find out whether rainfall,
minimum and maximum air temperature reveal trends over
a long time-period. To accomplish this percentage departure
from the monthly mean values of the variables have been
characterized for different decades, and in order to examine
the existence of a trend and calculate the magnitude of trend
in rainfall and temperature data, the seasonal Kendall Test
and Sen’s slope estimators (Gocic and Trajkovic 2013)
were used. Further, to find out long-term persistence in the
time-series data, the Detrended Fluctuation Analysis (DFA)
method was used.

DATA
The long-term meteorological data from 1980 to 2012
period was obtained from National Data Centre of IMD for
11 weather monitoring stations spread across the entire
State of Goa. These stations are located at Mapusa, Panjim, Fig. 1: Rainfall monitoring stations operated by various
Pernem, Valpoi, Ponda, Margao, Mormugoa, Sanguem, agencies
Quepem, Canacona and Dabolim. The Water Resources
Department (WRD), Government of Goa has also started METHODOLOGY
monitoring climatic parameters more recently since 2009- Daily observations of the climaticvariables (rainfall and
10. Data from 6 WRD stations (Karapur, Colvale, temperature) collected from IMD, WRD and CFSR were
Cuncolim, Chapoli, Dhargal, Valpoi) were obtained for a 9 aggregated to monthly and annual means.The historical data
year period from 2009 to 2017. At Valpoi, the IMD station were scrutinized for consistency and missing records.
is located approximately 900 m away from the WRD Outliers were removed and missing data was filled-in using
monitoring station. linear interpolation for the temperature and rainfall. Annual
The CFSR dataset consists of hourly weather forecasts mean and standard deviation, mean of each decade (1980-
generated by the National Weather Service’s NCEP Global 1989, 1990-1999, 2000-2009 and 2010-2012) and the mean
Forecast System. Forecast models are reinitialized every 6h during the pre-monsoon (February, March, April, May),
using information from the global weather station network monsoon (June, July, August, September,) and the post-

11
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

monsoon (October, November, December and January) contributes to less than 1% of the annual rainfall in Goa and
periods were calculated for the rainfall and temperature hence in some months no variation was seen.It was
data. observed that CV at Valpoi (IMD and WRD station) was
comparable to the closet CFSR station at Usgao.
The coefficient of variation (CV) and percentage departure
from the mean were calculated for all the data. Further we The CV for average monthly maximum and minimum
used (i) The Mann–Kendall statistical test has been temperatures as shown in Table 2 was analogous for all the
frequently used to quantify the significance of trends stations. The CV for both minimum and maximum
(increasing/decreasing/trendless) in hydro-meteorological temperature at the two stations that are in close proximity
time series (Kisi and Ay 2014; Duhan and Pandey 2013; i.e. Mormugoa (IMD) and Coastal (CFSR) was found to be
Silva et al. 2013), (ii) Sen’s slope method (Sen, 1968) and similar.
(iii) Detrended Fluctuation Analysis (DFA) (Yue et al.,
Table 3 shows meanannual rainfall and standard deviation
2010).
for all the monitoring stations. For IMD stations, on an
RESULTS & DISCUSSION average the maximum variation in rainfall was observed at
Mormugoa and minimum at Canacona. In each of the three
Statistical analysis successive decades from 1980-2009, majority of the stations
The CV for the monthly mean rainfall for the period of show a slight increase followed by a small decrease in the
1980-2012 been presented in Table 1. For all the rain rainfall. The pre-monsoon season shows a variation in
monitoring stations, variation observed in the rainfall during rainfall from 5 mm to 142.1 mm in the stations, while the
June, July and August was below 40%, however the monsoon season shows a distribution of 1388.5 mm to 3825
remaining months show a considerably higher variation, mm in the stations. The post-monsoon season has a
ranging from 60% to as high as 400%. It must be noted that variation of 104.7 mm to 525.1 mm.
the period from December to April (non-monsoon)
Table 1: Co-efficient of Variation (CV) for rainfall for all the monitoring stations
CV
STATIONS Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Rainfall (1980-2012)
IMD
Mapusa 2.66 0 3.40 2.01 1.37 0.26 0.32 0.34 0.80 0.79 1.51 2.40
Panjim 3.36 4.03 3.89 1.87 1.56 0.30 0.35 0.37 0.70 0.82 1.47 2.92
Pernem 0.93 4.80 4.80 4.31 1.47 0.17 0.30 0.19 0.52 0.48 0.70 0.73
Ponda 3.71 3.44 3.12 1.59 1.17 0.30 0.27 0.31 0.60 0.88 1.32 2.32
Valpoi 3.34 3.44 2.66 1.50 1.22 0.27 0.24 0.31 0.57 0.57 1.22 2.75
Canacona 0 0 0 1.22 0.79 0.22 0.25 0.11 0.55 0.42 0.84 0.93
Dabolim 2.77 3.42 2.96 1.24 2.15 0.29 0.35 0.39 0.70 0.75 1.12 2.00
Margao 3.57 0 3.40 2.08 1.31 0.28 0.30 0.38 0.70 0.79 1.46 2.35
Mormugoa 3.56 3.92 3.72 2.25 1.56 0.31 0.36 0.37 0.65 0.83 1.85 2.83
Quepem 4.77 0 3.34 1.58 1.11 0.27 0.30 0.35 0.61 0.76 1.29 2.13
Sanguem 3.77 0 2.81 1.42 1.39 0.30 0.24 0.33 0.57 0.58 1.10 2.00
CFSR
Cavorem 2.19 2.45 3.37 2.23 1.88 0.31 0.23 0.24 0.47 0.71 1.00 1.62
Coastal 2.73 3.93 3.82 1.80 1.98 0.31 0.23 0.25 0.48 0.71 1.07 1.81
Usgao 2.01 2.52 3.09 1.76 1.94 0.28 0.22 0.19 0.40 0.67 0.94 1.97
WRD
Karapur 2.83 2.62 2.83 1.58 0.93 0.30 0.28 0.25 0.27 0.34 1.49 2.29
Colvale 0 0 1.96 2.83 1.02 0.28 0.29 0.35 0.36 0.54 1.94 0
Cuncolim 0 0 0 1.58 1.00 0.23 0.32 0.39 0.46 0.98 1.23 2.83
Chapoli 0 0 0 1.88 1.54 0.40 0.38 0.36 0.32 0.87 2.26 2.83
Dhargal 0 0 2.83 0 1.42 0.25 0.25 0.34 0.29 0.72 1.92 0
Valpoi 0 0 2.83 2.83 1.44 0.27 0.29 0.25 0.29 0.35 2.83 0

12
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020
Table 2: Co-efficient of Variation (CV) for minimum and maximum temperature for all the monitoring stations
CV
Stations Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Min. Temp (1980-2012)
Mormugoa (IMD) 0.02 0.03 0.05 0.03 0.02 0.02 0.01 0.01 0.01 0.02 0.03 0.03
Panjim (IMD) 0.04 0.04 0.04 0.03 0.02 0.02 0.02 0.01 0.01 0.02 0.04 0.03

Cavorem (CFSR) 0.04 0.06 0.05 0.05 0.04 0.03 0.03 0.03 0.04 0.04 0.05 0.05
Coastal (CFSR) 0.02 0.03 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.02 0.03 0.02
Usgao (CFSR) 0.05 0.07 0.06 0.06 0.05 0.03 0.03 0.03 0.05 0.05 0.06 0.07

Karapur (WRD) 0.07 0.10 0.05 0.04 0.04 0.03 0.02 0.02 0.01 0.03 0.05 0.07
Max. Temp (1980-2012)
Mormugoa (IMD) 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03
Panjim (IMD) 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

Cavorem (CFSR) 0.04 0.04 0.03 0.03 0.04 0.04 0.02 0.03 0.03 0.04 0.04 0.04
Coastal (CFSR) 0.02 0.03 0.02 0.02 0.03 0.02 0.01 0.02 0.02 0.02 0.02 0.02
Usgao (CFSR) 0.03 0.04 0.03 0.02 0.04 0.05 0.02 0.03 0.05 0.04 0.04 0.03

Karapur (WRD) 0.03 0.03 0.02 0.02 0.03 0.08 0.03 0.02 0.01 0.03 0.03 0.03
Table 3. Mean rainfall and the standard deviation of the 33 years and the average rainfall in the past decades and
seasons for all the monitoring stations
Average rainfall in the past
Standa Seasonal average rainfall (mm)
Mean decades (mm)
rd
Station Rainfall
Deviati 1980- 1990- 2000- 2010- Pre- Post-
(mm) Monsoon
on (σ) 1989 1999 2009 2012 Monsoon Monsoon

IMD
Mapusa 3088 579 3139 3143 2971 3465 77 2859 151
Panjim 2898 489 2819 2962 2909 3229 82 2652 164
Pernem 3689 527 3389 3446 4158 4048 74 3311 370
Ponda 3392 558 3243 3519 3376 3725 82 3089 220
Valpoi 4156 575 4056 4245 4080 4764 78 3813 265
Canacona 3982 395 3614 4075 4333 3767 26 3431 525
Dabolim 2799 544 2709 2949 2787 2708 88 2531 179
Margao 2998 594 2995 3021 2943 3512 81 2749 167
Mormugoa 2731 458 2621 2832 2721 3079 88 2472 171
Quepem 3622 609 3732 3630 3461 3968 105 3303 214
Sanguem 3722 584 3705 3748 3702 4116 69 3396 256
CFSR
Cavorem 4279 828 4211 4656 3757 4836 142 3825 312
Coastal 4139 808 4228 4459 3548 4762 124 3751 264
Usgao 3919 654 3911 4328 3587 3578 128 3477 313
WRD
Karapur* 1825 1924 - - - - 37 1656 132
Colvale* 1519 1609 - - - - 26 1388 105

13
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

Cuncolim* 1550 1658 - - - - 26 1407 117

Chapoli* 1569 1733 - - - - 5.0 1442 122
Dhargal* 1616 1704 - - - - 13 1492 110
Valpoi* 2120 2228 - - - - 17 1974 129

The monthly maximum temperatures varied by 2-3% for the only 9 years, it was not considered for this exercise. The
33 year period, however Usgao and Cavorem sites (CFSR Figure 3 and Figure 4 shows the percentage departure of
data) showed a slightly higher variation of 4-5%. The average total rainfall in the three decades (1980–1989,
minimum temperatures on the other hand show a vast 1990–1999, 2000-2009) and recent three years from 2010–
difference, with variation of 0% to 7% in the monitoring 2012, from their corresponding normal value for IMD and
stations. In the three decades from 1980-2009, most stations CFSR stations respectively. All the stations exhibit
show a minor decrease followed by a minor increase in the departure in the range of +30 to -10% from the normal
temperature (maximum, minimum and mean) in each decadal value. For most of the stations, high departure was
successive decades. observed during non-monsoon months for certain decades.
However, departure in rainfall is moderate to low during the
The percentage departure in the monthly mean value was
monsoon season.
computed for the 33 years data that was collected from IMD
and CFSR. Since, the data from WRD was available for
Table 4. Mean rainfall and the standard deviation of the 33 years and the temperature in the past decades and
seasons for all the monitoring stations

Seasonal Average Temperature Average temperature in the

Avg. Standard (°C) past decades (°C)
Station Temp Deviation
Pre- Post- 1980- 1990- 2000- 2010-
(°C) (σ) Monsoon
Monsoon Monsoon 1989 1999 2009 2012
Maximum Temperature (Annual)
Panjim (IMD) 31.8 0.30 32.9 29.7 32.6 31.7 31.6 31.9 32.1
Mormugoa (IMD) 31.4 0.26 32.4 29.7 32.2 31.4 31.3 31.6 31.2
Cavorem (CFSR) 31.2 0.60 33.6 27.1 33.0 31.0 31.1 31.8 30.2
Coastal (CFSR) 29.3 0.36 30.5 27.6 29.8 29.1 29.2 29.7 28.9
Usgao (CFSR) 31.3 0.50 34.6 26.4 32.9 31.0 31.2 31.9 30.9
Karapur (WRD) 33.7 0.52 36.4 30.3 34.5 - - - -
Mean Annual Temperature
Panjim (IMD) 27.6 0.22 28.4 27.0 27.3 27.5 27.5 27.6 27.7
Mormugoa (IMD) 27.8 0.22 28.7 27.1 27.7 27.8 27.8 28.0 27.7
Cavorem (CFSR) 26.8 0.38 27.8 25.1 27.3 26.4 26.6 27.1 27.1
Coastal (CFSR) 27.5 0.30 28.3 26.9 27.2 27.2 27.3 27.7 27.6
Usgao (CFSR) 26.3 0.49 27.8 24.5 26.6 25.9 26.0 26.6 27.1
Karapur* (WRD) 27.8 0.51 29.3 27.1 27.1 - - - -
Minimum Temperature (Annual)
Panjim (IMD) 23.4 0.25 23.9 24.3 21.9 23.4 23.4 23.3 23.4
Mormugoa (IMD) 24.2 0.25 25.0 24.5 23.2 24.3 24.2 24.3 24.3
Cavorem (CFSR) 22.3 0.75 22.0 23.2 21.7 21.9 22.0 22.5 24.1
Coastal (CFSR) 25.6 0.35 26.1 26.1 24.6 25.3 25.5 25.8 26.2
Usgao (CFSR) 21.2 0.90 20.9 22.5 20.2 20.8 20.9 21.4 23.4
Karapur (WRD) 21.9 0.58 22.2 24.0 19.6 - - - -

14
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

Fig. 3: Percentage departure of monthly rainfall from 33 years normal value for IMD stations

Similarly the percentage departure for minimum and except for Colvale and Dhargal, which show a negative
maximum temperature for IMD and CFSR stations is trend at 5% significance for short period of data (only 9
depicted in Figure 5 and 6 respectively. It was observed that years of data is used) and Usgao shows a negative trend at
the minimum temperatures show ±3% departure, except 10% significance for longer period of data (1979-2010).
Usgao and Cavorem sites (CFSR data), where large These stations show a decrease of 138 mm/year and 157.72
variations for 2010-2012 years was seen. The maximum mm/year in annual rainfall respectively. It must be noted
temperatures have a high departure from the normal values that the extremely high values maybe on account of the
for the years 2010-2012 for all the monitoring stations. The limited period (9 years) of available rainfall for these
decade 2000-2009 also shows a higher departure compared stations.Pernem and Canacona(IMD) also show a positive
to the other decades under study. trend at significance of 0.1%. However it must be noted
these stations had large gaps in the data and hence linear
The rainfall trend analysis was done using 33 years of data
interpolation was used to rectify the same. Panjim show an
that was available from IMD and CFSR and 9 years data
increase of 7.9 mm/year while Mormugoa shows an
(2009-17) from WRD and is presented in Table 5. All the
increase of 9.5 mm/year in annual rainfall at 10%
stations show no significant trend using the Kendall’s tau,
significance.

15
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

Fig. 4: Percentage departure of monthly rainfall from 33 years normal value for CFSR station

Fig. 5: Percentage departure of minimum temperature from33 years normal value

Fig. 6: Percentage departure of maximum temperature from33 years normal value

16
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

Trend analysis temperature at the various stations monitored has been

presented in Table 6. Most of the studied stations show a
The rainfall trend analysis was done using 33 years of data
trend in the average annual temperatures, except for the
that was available from IMD and CFSR and 9 years data
minimum temperatures at Panjim and Mormugao. The
(2009-17) from WRD and is presented in Table 5. All the
remaining stations all show a positive trend in their
stations show no significant trend using the Kendall’s tau,
max/min temperatures at 1% and lower significance, except
except for Colvale and Dhargal, which show a negative
Karapur at 5% significance. The Sen’s slope analysis of the
trend at 5% significance for short period of data ( only 9
temperature shows a general increase of 0.01° to 0.03°C in
years of data is used) and Usgao shows a negative trend at
the max/min temperatures of the monitored stations, except
10% significance for longer period of data (1979-2010).
Karapur showing an average increase in the annual
These stations show a decrease of 138 mm/year and 157.72
temperatures of 0.2°C and 0.17°C for the maximum and
mm/year in annual rainfall respectively. It must be noted
minimum temperatures respectively. The analysis revealed
that the extremely high values maybe on account of the
that, none of the stations indicate in-significant trends both
limited period (9 years) of available rainfall for these
positive and negative for the data considered for analysis.
stations. Pernem and Canacona(IMD) also show a positive
However it must be noted that for this station only 9 years
trend at significance of 0.1%. However it must be noted
of temperature data was used which could be a limitation.
these stations had large gaps in the data and hence linear
interpolation was used to rectify the same. Panjim show an Table 6: Trend analysis for minimum and maximum
increase of 7.9 mm/year while Mormugoa shows an temperature using statistical methods
increase of 9.5 mm/year in annual rainfall at 10%
Sen's p- Kendall's DFA
significance.
Slope Value tau(τ) exponent
Station
Table 5: Trend analysis for rainfall using statistical (mm/yr)
methods Raw DS
Maximum Temperature (Annual)
Sen's p- Kendall's DFA
Panjim (IMD) 0.02 0.999 0.379 0.321 0.342
Slope Value tau(τ) exponent
Station Mormugoa (IMD) 0.01 0.877 0.144 0.285 0.332
(mm/yr)
Raw DS Cavorem (CFSR) 0.03 0.998 0.352 0.544 0.522
IMD Coastal (CFSR) 0.02 0.999 0.375 0.438 0.426
Mapusa 3.16 0.568 0.023 0.348 0.291 Usgao (CFSR) 0.03 0.998 0.364 0.412 0.380
Panjim 7.90 0.850 0.129 0.267 0.171 Karapur* (WRD) 0.20 0.012 0.611 3.194 3.112
Pernem 67.87 1.000 0.527 0.497 0.093 Mean Annual Temperatures
Ponda 9.10 0.767 0.091 0.311 0.228 Panjim (IMD) 0.01 0.906 0.163 0.180 0.211
Valpoi 8.46 0.767 0.091 0.322 0.289 Mormugoa (IMD) 0.00 0.864 0.136 0.228 0.243
Canacona 24.64 1.000 0.754 0.464 0.493 Cavorem (CFSR) 0.03 1.000 0.561 0.358 0.305
Dabolim 3.14 0.592 0.030 0.330 0.180 Coastal (CFSR) 0.02 1.000 0.492 0.320 0.318
Margao 9.34 0.717 0.072 0.338 0.246 Usgao (CFSR) 0.04 1.000 0.583 0.291 0.276
Karapur* (WRD) 0.13 0.022 0.556 2.547 2.691
Mormugoa 9.50 0.871 0.140 0.226 0.138
Minimum Temperatures (Annual)
Quepem -6.89 0.315 -0.061 0.397 0.212
Panjim (IMD) 0.00 0.327 -0.057 0.168 0.186
Sanguem 2.13 0.555 0.019 0.302 0.389
Mormugoa (IMD) 0.00 0.685 0.061 0.250 0.252
CFSR
Cavorem (CFSR) 0.03 1.000 0.515 0.378 0.367
Cavorem -17.76 0.164 -0.121 0.282 0.117
Coastal (CFSR) 0.03 1.000 0.598 0.289 0.282
Coastal -21.88 0.099 -0.159 0.370 0.110 Usgao (CFSR) 0.04 1.000 0.496 0.406 0.395
Usgao -20.29 0.028 -0.235 0.314 0.173 Karapur* (WRD) 0.17 0.012 0.611 1.649 1.860
WRD
Karapur* -129.51 0.060 -0.444 2.162 1.608 Long-term persistence in time series
Colvale* -138.00 0.022 -0.556 1.835 1.606 The long range dependence of the meteorological
Cuncolim* -62.28 0.460 -0.056 3.143 1.955 parameters was checked, using DFA, for raw and de-
Chapoli* -240.62 0.179 -0.278 2.164 1.851 seasonalized time series. If the value of DFA is 0.5, it
Dhargal* -157.72 0.006 -0.667 2.208 2.009 indicates that there is a long range persistence in the time
Valpoi* -149.78 0.060 -0.444 2.756 1.507 series. The DFA exponent (α) for rainfall was found to be in
Table 6 gives the trend analysis results for temperature the range 0.23–0.497 for the raw time series and between
using statistical methods. The trend analysis (Sen’s slope, p- 0.09 and 0.49 for the de-seasonalized (DS) time series.
value and Kendall’s tau) for the minimum and maximum Although the raw data shows a weak persistence in the time

17
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

series, the de-seasonalized data shows a lower persistence monsoons have a significant increasing trend for most of the
compared to the raw data. The DFA exponent for the WRD stations. The pre-monsoon shows an increase of 0.6163
stations shows a very high value, but it must be noted that mm/year to 2.8563 mm/year, while the post monsoon shows
these stations have only 9 years of rainfall data. In order to an increase of 2.3463 mm/year to 14.63 mm/year. Thus, it is
be reliable the time range must be greater than 20 years. observed that percentage of rainfall falling in the monsoons
(June to September) is decreasing and there is an increase in
The maximum temperature for the Cavorem station shows
the percentage in the remaining months.
long term persistence with a DFA exponent (α) of 0.54 in
the raw series and 0.52 in the de-seasonalized series. The On comparing the three stations (IMD, WRD station at
DFA exponent (α), of the remaining monitored stations, for Valpoi and Usgao of CFSR) rainfall data for the period
the maximum temperature is in the range 0.29–0.44 for the from 2009 to 2015 using the Kendall’s seasonal test, it was
raw time series and between 0.33 and 0.43 for the de- observed that the stations show a significant trend. The
seasonalized time series. While for the minimum WRD station indicates the decrease of 150.42mm/year of
temperatures, the DFA exponent (α) is in the range between annual rainfall while the IMD station exhibits a decrease of
0.17-0.41 for the raw series and 0.19-0.40 for the de- 111.48 mm/year of annual rainfall and the CFSR station
seasonalized time series data. indicates an increase of 72.27 mm/year. The lower annual
trend and rainfall of the CFSR stations may be attributed to
Spatio-Temporal Rainfall Variability
its inland location compared to the Ghats region of the other
The average annual rainfall falling on the coastal belt of two stations. Satellite data generally underestimate both the
Goa (Panjim, 2898 mmand Mormugoa, 2731 mm) is less mean and variability of rain, but the phase of inter-seasonal
than the interior of Goa (Ponda, 3392 mm) and far less than variations is accurate in Goa which is west coast of India
the regions near the Western Ghats (Valpoi, 4157mm).The (Rahmanet.al., 2009). Overall, all the stations show a
IMD stations in the North Goa district (Panjim, Mapusa, significant increasing trend. However, it is observed that, in
Pernem, Valpoi and Ponda)receive an average annual the recent times, there is an increasing rainfall amount in
rainfall of 3444.62 mm/year, while the South Goa district some of the station which areboardering the Western Ghats.
(Mormugoa, Dabolim, Margao, Canacona, Quepem and But in the inland a decreasing rainfall amount is observed.
Sanguem) receive an average annual rainfall of 3309.08
mm/year for the 33 year study period. The North Goa SUMMARY AND CONCLUSION
district stations show a Sen’s slope of 11.79 mm/year while
From the long term trend analysis carried out it is evident
the South Goa district stations show a Sen’s slope of 6.33
that since 1980 till 2012, there was a minor increase (~7-
mm/year.
8mm/year) in the rainfall, however, for more recent years
The trend analysis was carried out for the pre-monsoon, (2009-2017) there is a decrease in the annual rainfall that is
monsoon and post-monsoon seasons of the 11 IMD stations. received by the state.
The results (Table 7) show that the pre-monsoons and post-
The North Goa district stations have received a higher
Table 7: Trend analysis for seasonal periods using statistical methods
Sen's p- Kendall Sen's p- Kendall Sen's p- Kendall
Slope Value tau (τ) Slope Value tau(τ) Slope Value tau (τ)
Station (mm/year) (mm/ (mm/year)
year)
Pre-Monsoon Monsoon Post Monsoon
Mapusa 0.71 0.828 0.117 -0.84 0.469 -0.008 0.75 0.685 0.061
Panjim 1.39 0.911 0.167 6.48 0.757 0.087 1.14 0.767 0.091
Pernem -0.18 0.000 -0.621 53.41 1.000 0.500 14.63 1.000 0.644
Ponda 1.35 0.906 0.162 0.24 0.494 0.000 1.80 0.864 0.136
Valpoi 0.76 0.850 0.129 4.50 0.639 0.045 2.23 0.828 0.117
Canacona 1.88 1.000 0.598 18.34 1.000 0.705 4.43 1.000 0.754
Dabolim 2.85 1.000 0.405 -3.81 0.396 -0.030 2.34 0.921 0.174
Margao 2.34 0.934 0.186 9.47 0.685 0.061 -0.38 0.420 -0.023
Mormugoa 1.67 0.963 0.220 5.46 0.776 0.095 1.53 0.776 0.095
Quepem 0.61 1.000 0.598 -4.98 1.000 0.705 -0.80 1.000 0.754
Sanguem 1.55 0.980 0.254 -3.17 0.38 -0.034 3.72 0.890 0.152

18
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

amount of average annual rainfall for the 33 year study REFERENCES

period compared to the South Goa stations. The North Goa
district stations also show a higher increase in rainfall in the 1. Duhan D, Pandey, A., 2013. “Statistical analysis of
3 decades rainfall data studied. While the rainfall increases long term spatial and temporal trends of precipitation
from West (Coast) to East (Ghats), as the Western Ghat during 1901–2002 at Madhya Pradesh, India”,
ranges acts as a barrier to the monsoon winds originating in Atmosphere Research 122:136–149.
the Indian ocean and moving north east (Putty & 2. Gocic M, Trajkovic S., 2013.“Analysis of changes in
Madhusoodhanan 2013). Furthermore, on closer analysis of meteorological variables using Mann–Kendall and
the monthly rainfall patterns it reveals that there is a shift in Sen’s slope estimator statistical tests in Serbia”, Global
the amount of rainfall falling in the monsoon period with an Planet Change, 100:172–182.
increase in the post monsoon and pre monsoon months.
3. Jain, Sharad& Kumar, Vijay. 2012.“Trend analysis of
The long term study also conclusively shows that there is an rainfall and temperature data for India”, Current
increase in the annual maximum temperatures experience Science (Bangalore)102 (1), 37-49.
by Goa. This has also been reported by (TERI 2016). The
report on climate change trends in India (Rathore, L.S. et al. 4. Kendall, M,G., 1975. “Rank correlation methods”,
2013), by the India Meteorological Department, mentions Griffin, London
that the average annual mean maximum temperature for 5. Kishore, P, Jyothi, S, Basha, G, Rao, SVB, Rajeevan,
Goa increasedby +0.04 °C/year for the period of 1951-2010. M, Velicogna, I, Sutterley, TC. 2015. “Precipitation
This study has observed an annual increase of +0.03°C/year climatology over India: Validation with observations
from 1980 to 2012. and reanalysis datasets and spatial trends”, Climate
Although the overlapping period of 7 years for the IMD and Dynamics, 2015. 46 (1-2)
WRD Valpoi stations is not significantly large to establish 6. Kisi O, Ay M. 2014 Comparison of Mann–Kendall and
any long term co-relationship, it is evident from the data at innovative trend method for water quality pa-rameters
hand that the rainfall pattern report by both institutions is of the Kizilirmak River, Turkey. J Hydrol513(26):362–
fairly consistent with one another. The IMD station on an 375.
average reports a slightly higher amount of rainfall
7. Kouwen N, Danard M, Bingeman A, Luo W,
annually. The difference in the values at these monitoring
Seglenieks FR, Soulis ED. 2005.“Case study:
stations could be due to the surrounding topography,
watershed modeling with distributed weather model
vegetation at the specific location of the individual station.
data”, Journal of Hydrologic Engineering, 10(1): 23–
Furthermore thetype of rain gauge employed and the
38.
accuracy with which the data is collected and human errors
can affect the reported rainfall for a particular station. 8. Mann, H, B., 1945. “Nonparametric tests against
trend”, Econometrica13:245–259
The findings of this study can provide crucial inputs to
climate change related studies that involve hydrological 9. Metri,S.M., and Singh, K.,2010 Study of rainfall
modeling. More importantly, to the water resources features over goa state during Southwest monsoon
manages and dam operators, as they need to develop season. MAUSAM 61(2), 155-162
appropriate dam operating mechanism to efficiently use the
10. Nandargi, S., andMulye, S.S., 2014. Spatio-Temporal
available water in the dam. It would also assist the planners
Rainfall Variability and Rainstorm Analysis over the
and decision-makers in developing and implementing
Goa State, India, Studies in Atmospheric Science 1, 8-
adaptation strategies at local level.
23.

ACKNOWLEDGEMENT 11. Putty, M.R.Y. &Madhusoodhanan, C. G. 2013. Water

Resources and Hydrology of the Western Ghats: Their
Authors gratefully acknowledge the Water Resources Role and Signifcance in South India. Bulletin of the
Department (WRD), Government of Goa for providing the National Natural Resources Management Systems.57-
financial support to carry out this study. The authors are 66.
grateful to IMD, WRD and CFSR for providing data.We
deeply acknowledge the timely suggestions extended by Dr. 12. Rahman, S, H, Debasis, Sengupta, and M,
Nandagiri (Professor, National Institute of Technology Ravichandran., 2009. “Variability of Indian summer
Karnataka, Surathkal). monsoon rainfall in daily data from gauge and

19
 J. Indian Water Resour. Soc., Vol. 40, No. 3 & 4, Jul. & Oct., 2020

satellite”, Journal of Geophysical Research, Vol. 114, 18. Silva RM, Santos CAG, Macedo MLA, Silva L, Freire
D17113, doi:10.1029/2008JD011694 PKMM 2013 Space–time variability of rainfall and
hydrological trends in the Alto São Francisco River
13. Raju, Kumar, B.C. and LakshmanNandagiri. 2017.
basin. IAHS-AISH Publ359:48–54.
“Analysis of historical trends in hydrometeorological
variables in the upper Cauvery basin, Karnataka, 19. Singh, N., Sontakke, N. A., Singh, H. N. and Pandey,
India”, Current Science, 112(3) pp. 577-587 A. K. 2005.“Recent trend in spatiotemporal variation of
rainfall over Indian investigation into basin-scale
14. Rathore, L &Attri, S D &Jaswal, Ashok. 2013. “State
rainfall fluctuations”, IAHS Publication296, pp. 273–
Level Climate Change Trends in India”, Ministry of
282.
Earth Science, Government of India.
20. Suprit, K. and D. Shankar. 2008. “Resolving
15. Rupa Kumar, K., Pant, G. B., Parthasarathy, B. &
orographic rainfall on the Indian west coast”,
Sontakke, N. A., 1992. “Spatial and sub seasonal
International Journal of Climatology,28: 643-657.
patterns of the long term trends of Indian summer
monsoon rainfall”, International Journal of 21. Revadekar J.V, Varikoden H, Murumka, P.K, et.al.,
Climatology, 12, 257–268 (2018) Latitudinal variation in summer monsoon
rainfall over Western Ghat of India and its association
16. Saha, S., S. Moorthi, H. Pan, X. Wu, J. Wang, S.
with global sea surface temperatures. Science of the
Nadiga, P. Tripp, R. Kistler, J. Woollen, D. Behringer,
Total Environment (613–614), 88–97.
H. Liu, D. Stokes, R. Grumbine, G. Gayno, J. Wang, Y.
Hou, H. Chuang, H.H. Juang, J. Sela, M. Iredell, R. 22. Preethi B, Mujumdar M, Kripalani R.H, et.al., ( 2017)
Treadon, D. Kleist, P. Van Delst, D. Keyser, J. Derber, Recent trends and tele-connections among South and
M. Ek, J. Meng, H. Wei, R. Yang, S. Lord, H. van den East Asian summer monsoons in a warming
Dool, A. Kumar, W. Wang, C. Long, M. Chelliah, Y. environment. Clim. Dyn. 48, 2489–2505.
Xue, B. Huang, J. Schemm, W. Ebisuzaki, R. Lin, P.
23. TERI. 2014. “Climate Resilient Infrastructure Services
Xie, M. Chen, S. Zhou, W. Higgins, C. Zou, Q. Liu, Y.
– Panaji and Vishakhapatnam, New Delhi: The Energy
Chen, Y. Han, L. Cucurull, R.W. Reynolds, G.
and Resources Institute. 258 pp (Project Report No.
Rutledge, and M. Goldberg, 2010. “The NCEP Climate
2013UD04).
Forecast System Reanalysis”, Bulletin of the American
Meteorological Society, 91, 1015–1058. 24. Yue, J., Zhao, X., and Shang, P., 2010. Effect of
Trends on Detrended Fluctuation Analysis of
17. Sen, P.K., 1968. “Estimates of the regression
Precipitation Series, Mathematical Problems in
coefficient based on Kendall’s tau”, Journal of
Engineering doi:10.1155/2010/749894
American Statistical Association, 63:1379–1389.

20