Evaluation of machine learning approaches for large-scale agricultural drought forecasts to improve monitoring and preparedness in Brazil

Gallear, Joseph W.; Valadares Galdos, Marcelo; Zeri, Marcelo; Hartley, Andrew

doi:https://doi.org/10.5194/nhess-25-1521-2025

Articles | Volume 25, issue 4

https://doi.org/10.5194/nhess-25-1521-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/nhess-25-1521-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 25, issue 4

Research article

|

25 Apr 2025

Research article |

| 25 Apr 2025

Evaluation of machine learning approaches for large-scale agricultural drought forecasts to improve monitoring and preparedness in Brazil

Joseph W. Gallear, Marcelo Valadares Galdos, Marcelo Zeri, and Andrew Hartley

Data sets

ERA5 monthly averaged data on single levels from 1940 to present Hans Hersbach et al. https://doi.org/10.24381/cds.f17050d7

A quasi-global precipitation time series for drought monitoring C. C. Funk et al. https://data.chc.ucsb.edu/products/CHIRPS-2.0/

GPCC First Guess Product at 1.0: Near real-time first guess monthly land-surface precipitation from rain-gauges based on SYNOP data Markus Ziese et al. https://doi.org/10.5676/DWD_GPCC/FG_M_100

Groundwater and Soil Moisture Conditions from GRACE and GRACE-FO Data Assimilation L4 7-days 0.125 x 0.125 degree U.S. V4.0 Hiroko Beaudoing et al. https://doi.org/10.5067/UH653SEZR9VQ

NESDIS STAR - Global Vegetation Health Products NOAA https://www.star.nesdis.noaa.gov/smcd/emb/vci/VH/vh_ftp.php

SPEIbase v.2.10: A Comprehensive Tool for Global Drought Analysis Santiago Beguería et al. https://digital.csic.es/handle/10261/364137

Model code and software

Jgallear/CSSP_brazil_23_24: Initial release, code for Evaluation of machine learning approaches for large-scale agricultural drought forecasts to improve monitoring and preparedness in Brazil (v1.0.0) Joe Gallear https://doi.org/10.5281/zenodo.15210667

Short summary

In Brazil, drought is of national concern and can have major consequences for agriculture. Here, we determine how to develop forecasts for drought stress on vegetation health using machine learning. Results aim to inform future developments in operational drought monitoring at the National Centre for Monitoring and Early Warning of Natural Disasters (CEMADEN) in Brazil. This information is essential for disaster preparedness and planning of future actions to support areas affected by drought.