This is a project providing equations, code and input data to compute the area of forest loss due to fire and estimate accuracy of the resulting map in the following study: Tyukavina et al. (in review) "Global trends of forest loss due to fire, 2001-2019". For a detailed description of the study please refer to the publicaiton.

Project files:

• Equations.docx - Equations from the Supplementary Information of Tyukavina et al. article

• Compute_area_and_accuracy.ipynb - Python 3 code implementing the equations

• Strata_info.txt - sampling design information

  Columns:

  • "Region" - Unique region name, for region boundaries please refer to the article
  • "Stratum" - Stratum ID (1-N)
  • "Area_km2" - Stratum area
  • "Sample size" - Number of pixels sampled in each stratum

• Sample_data.txt - input data for each sample pixel

  Columns:

  • "ID" - Unique sample pixel ID (1-n)
  • "Region" - Unique region name, for region boundaries please refer to the article
  • "Stratum" - Stratum ID (1-N)
  • "Reference" - Reference sample value, 1 if sample pixel was identified as forest loss due to fire during the validation process, and 0 otherwise
  • "Pixarea" - Area of pixel in km2
  • "Map" - Map value from the map adjusted to match sample-based estimate, with 1 corresponding to forest loss due to fire, and 0 - to forest loss due to other drivers or no forest loss. This is the version of the map used for area and trend reporting in the paper, where any forest loss within year 2000 zero % tree canopy cover (Hansen et al. 2013) is mapped as "no forest loss".