ETo‐Brazil: A Daily Gridded Reference Evapotranspiration Data Set for Brazil - Repository

Published: 12-07-2020| Version 3 | DOI: 10.17632/sstjw74ryh.3
Daniel Althoff,
Santos Dias,
Roberto Filgueiras,
Lineu Neiva Rodrigues


Shortname: ETo-Brazil Longname: Gridded reference evapotranspiration for Brazil based on machine learning - 1-Day 0.1 degree x 0.1 degree DOI: 10.17632/sstjw74ryh.3 Version: 03 Format: TIFF Spatial Coverage: -74.0, -34.7, -33.8, 5.3 -- (xmin, xmax, ymin, ymax) Temporal Coverage: 2000-06-01 to 2019-12-31 File Size: ~0.3 MB per file Coordinate Reference System: WGS 84 - Geodetic coordinate system for World (EPSG: 4326) Data Resolution Spatial: 0.1° x 0.1° (rougly ~10 x 10 km) Temporal: 1-Day Highlights: - The grid cells of the ETo-Brazil matches the grid cells from the Global Precipitation Measurement data (position and resolution). - The overall root mean square error (RMSE) of this dataset (using weather stations as observed data) is 0.60 mm/day. This is an improvement on earlier attempts to map ETo using interpolation and geostatistics. ** ALWAYS CHECK FOR NEWEST VERSION ** History: - V3 (2020-07-12): - Fixed a small problem with the covariates projection. This fixed grid cells positioning and lead to an improvement in performance: RMSE reduced from 0.65 to 0.60 mm/day; - Extended the temporal coverage up to 2019-12-31; - V2 (2020-05-21): Masked files for Brazil boundaries only. Reduced file size in ~40%. - V1 (2020-03-20): First published