pith. machine review for the scientific record. sign in

arxiv: 2605.05255 · v1 · submitted 2026-05-05 · 📊 stat.AP · physics.ao-ph

Recognition: unknown

Prediction of Drought and Flash Drought in Africa at the Seasonal-to-Subseasonal Scale using the Community Research Earth Digital Intelligence Twin Framework

Amy McGovern, Jason Hickey, Stuart Edris

Pith reviewed 2026-05-08 16:33 UTC · model grok-4.3

classification 📊 stat.AP physics.ao-ph
keywords drought predictionmachine learningAfricasoil moistureflash droughtseasonal forecastingCrossFormer model
0
0 comments X

The pith

DroughtFormer, a machine learning model, predicts soil moisture and other drought indicators in Africa with skill out to 90-day lead times.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops and tests DroughtFormer to forecast droughts and flash droughts across Africa using reanalysis and satellite data. It incorporates physical constraints like dry air mass and moisture conservation into a CrossFormer architecture. The model underperforms on precipitation and flash drought indices but shows stable skill for soil moisture, vegetation health, and related variables that matches or exceeds climatology benchmarks. This matters because Africa relies heavily on rain-fed agriculture, so better long-range drought predictions could help with food security and planning. The work demonstrates that large ML weather models can retain usefulness even in data-sparse regions when focused on surface variables.

Core claim

DroughtFormer delivers skillful forecasts of soil moisture, vegetation health, and other surface variables out to 90 days, representing climate anomalies effectively though underestimating their magnitude, while struggling more with precipitation and flash drought indices.

What carries the argument

DroughtFormer, a CrossFormer model based on the CREDIT framework that ingests ERA5, GLDAS2, IMERG, and MODIS data with added conservation constraints to predict drought-related surface fields.

If this is right

  • Drought prediction in Africa can extend reliably to subseasonal and seasonal scales for agricultural variables.
  • ML models trained on global reanalyses can be adapted for data-sparse continents without losing all skill.
  • Early warning systems for drought could incorporate such models to support crop planning and resource allocation.
  • Focus on soil moisture and vegetation rather than precipitation alone yields more stable long-lead forecasts.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar models might improve flash drought prediction if trained with higher-resolution or additional convective-scale data.
  • Operational centers in Africa could test DroughtFormer for real-time seasonal outlooks.
  • Extending the framework to other vulnerable regions like South Asia or Australia would test its generalizability.
  • If physical constraints are key to stability, removing them might reveal how much they compensate for data sparsity.

Load-bearing premise

Historical patterns in reanalysis and satellite observations will continue to describe future drought conditions even as the climate changes.

What would settle it

An independent test on post-2023 data showing that DroughtFormer skill for soil moisture falls below climatology during a major African drought event.

read the original abstract

Droughts and flash droughts (rapidly developing droughts; FDs) remain impactful events that are known to desiccate landscape and destroy crops. In particular, droughts in Africa are often more impactful than in other locations, such as the United States or Europe, due to many regions in Africa heavily depending on local agriculture for sustenance. In recent years, large machine learning (ML) models, such as GraphCast and AIFS, have emerged as effective tools for global weather prediction. However, sparse data observations and few ML studies in Africa have left it unclear if these ML models retain their skill when focused on Africa. As such, this project seeks to examine the predictability of drought and FD in Africa using a CrossFormer model based on the Community Research Earth Digital Intelligence Twin (CREDIT) framework developed by NSF NCAR. Our CrossFormer model, termed DroughtFormer, incorporates variables from the ERA5 and GLDAS2 reanalyses and the IMERG and MODIS satellite observations, and employs dry air mass and moisture conservation, to predict soil moisture, vegetation health, and other drought-related surface variables. While DroughtFormer displayed lower accuracy in predicting precipitation and FD indices, it showed significant skill in predicting the remaining variables, delivering stable and skillful forecasts out to 90-day lead times (either beating out or having comparable skill to climatology). In particular, DroughtFormer skillfully represented climate anomalies for key variables, such as soil moisture (though it struggled with the magnitude of the anomalies). Thus, DroughtFormer showed significant promise in representing and predicting agricultural level drought in a region that is heavily impacted by drought events.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces DroughtFormer, a CrossFormer-based model within the CREDIT framework, that ingests ERA5, GLDAS2, IMERG, and MODIS data together with dry-air-mass and moisture-conservation constraints to forecast soil moisture, vegetation health, and other drought-related surface variables over Africa at lead times up to 90 days. The central claim is that the model delivers stable forecasts that match or exceed climatology for most variables (especially soil-moisture anomalies) while showing weaker performance on precipitation and flash-drought indices.

Significance. If the quantitative verification holds, the work would be a useful demonstration of ML-based subseasonal drought prediction in a data-sparse, high-impact region, with the explicit inclusion of conservation constraints providing a modest advance over purely data-driven approaches. The multi-source data fusion and focus on agricultural drought variables address a genuine application gap.

major comments (2)
  1. [Abstract] Abstract: the statements that DroughtFormer 'showed significant skill' and 'delivering stable and skillful forecasts out to 90-day lead times (either beating out or having comparable skill to climatology)' are presented without any numerical skill scores, error bars, train-test split details, or explicit baseline definitions. Because these quantitative elements are load-bearing for the central claim of predictive skill, their absence prevents assessment of whether the reported performance is statistically meaningful or merely reflects in-sample fit.
  2. [Abstract] Abstract (and implied methods): the evaluation is described only against climatology and the same reanalysis/satellite distributions used for training; no independent out-of-sample verification set, cross-validation procedure, or comparison to operational dynamical forecasts is mentioned. This leaves open the possibility that reported skill largely reflects the model's ability to reproduce historical patterns rather than genuine predictive capability under distribution shift.
minor comments (2)
  1. [Abstract] The acronym 'FD' for flash drought is introduced without expansion on first use.
  2. [Abstract] The phrase 'the remaining variables' is used without listing which variables are included in that category versus the weaker precipitation/FD group.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which help clarify the presentation of our results. We address each major comment below and indicate where revisions will be made to the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the statements that DroughtFormer 'showed significant skill' and 'delivering stable and skillful forecasts out to 90-day lead times (either beating out or having comparable skill to climatology)' are presented without any numerical skill scores, error bars, train-test split details, or explicit baseline definitions. Because these quantitative elements are load-bearing for the central claim of predictive skill, their absence prevents assessment of whether the reported performance is statistically meaningful or merely reflects in-sample fit.

    Authors: We agree that the abstract would be strengthened by including quantitative context. In the revised manuscript we will add specific skill metrics (e.g., anomaly correlation coefficient ranges for soil-moisture and vegetation anomalies at 30-, 60-, and 90-day leads) and briefly state the temporal train-test split (training through 2015, independent test years 2016 onward) together with the climatology baseline definition. These additions will allow readers to assess the statistical support for the skill claims without lengthening the abstract excessively. revision: yes

  2. Referee: [Abstract] Abstract (and implied methods): the evaluation is described only against climatology and the same reanalysis/satellite distributions used for training; no independent out-of-sample verification set, cross-validation procedure, or comparison to operational dynamical forecasts is mentioned. This leaves open the possibility that reported skill largely reflects the model's ability to reproduce historical patterns rather than genuine predictive capability under distribution shift.

    Authors: The full methods section already implements a strict temporal train-test split that holds out later years for evaluation, providing an independent out-of-sample test. We will revise the abstract to explicitly note this temporal split and the resulting distribution shift. Cross-validation was omitted owing to the computational cost of the CrossFormer architecture; we will add a brief limitations statement. Direct comparison against operational dynamical forecasts (e.g., ECMWF) lies outside the present scope, which focuses on demonstrating the CREDIT framework for African drought variables; we will insert a sentence acknowledging this gap and identifying it as future work. These clarifications will address the concern about potential in-sample overfitting. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper trains DroughtFormer (a CrossFormer variant) on ERA5, GLDAS2, IMERG, and MODIS data with added conservation constraints, then reports forecast skill versus climatology on (presumably held-out) test periods. This is a standard supervised ML evaluation pipeline whose outputs are not definitionally identical to its training inputs. No equations or steps are shown that reduce a claimed prediction to a fitted parameter by construction, no load-bearing self-citations appear, and the physical constraints are external to the target variables rather than tautological. The derivation chain therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the representativeness of ERA5, GLDAS2, IMERG, and MODIS data for African conditions and on the assumption that adding conservation constraints allows the model to generalize beyond the training period. No new physical entities are postulated.

free parameters (1)
  • CrossFormer model hyperparameters and weights
    The neural network contains a large number of trainable parameters optimized on the input datasets; these are not enumerated or constrained in the abstract.
axioms (1)
  • domain assumption Conservation of dry air mass and moisture can be usefully imposed as soft constraints within the ML architecture
    Stated in the abstract as part of the model design.

pith-pipeline@v0.9.0 · 5605 in / 1421 out tokens · 97720 ms · 2026-05-08T16:33:54.028824+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

144 extracted references · 135 canonical work pages

  1. [2]

    Spatiotemporal analysis of meteorological drought across China based on the high-spatial-resolution multiscale SPI generated by machine learning , url =

    He, Qian and Wang, Ming and Liu, Kai and Li, Bohao and Jiang, Ziyu , date-added =. Spatiotemporal analysis of meteorological drought across China based on the high-spatial-resolution multiscale SPI generated by machine learning , url =. Weather and Climate Extremes , month = jun, pages =. 2023 , Bdsk-Url-1 =. doi:10.1016/j.wace.2023.100567 , issn =

    work page doi:10.1016/j.wace.2023.100567 2023

  2. [3]

    Anande and Jason A

    Doreen M. Anande and Jason A. Otkin , booktitle =. Examining climatological characteristics and drivers of rapid change events in hydrological conditions over Africa , year =

  3. [5]

    Lang, Simon and Alexe, Mihai and Clare, Mariana C. A. and Roberts, Christopher and Adewoyin, Rilwan and Ben Bouall. AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score , url =. npj Artificial Intelligence , month = feb, number =. 2026 , Bdsk-Url-1 =. doi:10.1038/s44387-026-00073-7 , issn =

    work page doi:10.1038/s44387-026-00073-7 2026

  4. [7]

    Horn of Africa floods and drought, 2020-2023 - Forensic analysis , url =

    UNDRR , date-added =. Horn of Africa floods and drought, 2020-2023 - Forensic analysis , url =

    2020

  5. [8]

    and Huang, Jinlong and Su, Buda and Jing, Cheng and Qiu, Fangdao and Jiang, Tong , date-added =

    Yahaya, Ibrahim and Zhou, Jian and Jiang, Shan and Xu, Runhong H. and Huang, Jinlong and Su, Buda and Jing, Cheng and Qiu, Fangdao and Jiang, Tong , date-added =. Depiction of Drought Over Africa in the Light of Changing Climate from CMIP6 Models , url =. Earth Systems and Environment , month = nov, number =. 2024 , Bdsk-Url-1 =. doi:10.1007/s41748-024-00...

    work page doi:10.1007/s41748-024-00502-0 2024

  6. [13]

    and Young, Matthew P

    de Andrade, Felipe M. and Young, Matthew P. and MacLeod, David and Hirons, Linda C. and Woolnough, Steven J. and Black, Emily , date-added =. Subseasonal Precipitation Prediction for Africa: Forecast Evaluation and Sources of Predictability , url =. Weather and Forecasting , month = feb, number =. 2021 , Bdsk-Url-1 =. doi:10.1175/waf-d-20-0054.1 , issn =

    work page doi:10.1175/waf-d-20-0054.1 2021

  7. [18]

    The State of AI Ethics Report (Volume 7) - AI at the Crossroads: A Practitioner's Guide To Community-Centered Solutions , url =

    Butalid, Renjie and Wright, Connor and Kherroubi Garcia, Ismael , copyright =. The State of AI Ethics Report (Volume 7) - AI at the Crossroads: A Practitioner's Guide To Community-Centered Solutions , url =. 2025 , Bdsk-Url-1 =. doi:10.5281/ZENODO.17328882 , publisher =

    work page doi:10.5281/zenodo.17328882 2025

  8. [22]

    and Ault, Toby R

    Newman, Matthew and Alexander, Michael A. and Ault, Toby R. and Cobb, Kim M. and Deser, Clara and Di Lorenzo, Emanuele and Mantua, Nathan J. and Miller, Arthur J. and Minobe, Shoshiro and Nakamura, Hisashi and Schneider, Niklas and Vimont, Daniel J. and Phillips, Adam S. and Scott, James D. and Smith, Catherine A. , date-added =. The Pacific Decadal Oscil...

    work page doi:10.1175/jcli-d-15-0508.1 2016

  9. [23]

    and Mestas‐Nu

    Enfield, David B. and Mestas‐Nu. The Atlantic Multidecadal Oscillation and its relation to rainfall and river flows in the continental U.S. , url =. Geophysical Research Letters , month = may, number =. 2001 , Bdsk-Url-1 =. doi:10.1029/2000gl012745 , issn =

    work page doi:10.1029/2000gl012745 2001

  10. [24]

    Jones, P. D. and Jonsson, T. and Wheeler, D. , date-added =. Extension to the North Atlantic oscillation using early instrumental pressure observations from Gibraltar and south-west Iceland , url =. International Journal of Climatology , month = nov, number =. 1997 , Bdsk-Url-1 =. doi:10.1002/(sici)1097-0088(19971115)17:13<1433::aid-joc203>3.0.co;2-p , issn =

    work page doi:10.1002/(sici)1097-0088(19971115)17:13 1997

  11. [41]

    and Zolin, Cornelio A

    Anderson, Martha C. and Zolin, Cornelio A. and Sentelhas, Paulo C. and Hain, Christopher R. and Semmens, Kathryn and Tugrul Yilmaz, M. and Gao, Feng and Otkin, Jason A. and Tetrault, Robert , date-added =. The Evaporative Stress Index as an indicator of agricultural drought in Brazil: An assessment based on crop yield impacts , url =. Remote Sensing of En...

    work page doi:10.1016/j.rse.2015.11.034 2016

  12. [42]

    and Hain, Christopher and Otkin, Jason and Zhan, Xiwu and Mo, Kingtse and Svoboda, Mark and Wardlow, Brian and Pimstein, Agustin , date-added =

    Anderson, Martha C. and Hain, Christopher and Otkin, Jason and Zhan, Xiwu and Mo, Kingtse and Svoboda, Mark and Wardlow, Brian and Pimstein, Agustin , date-added =. An Intercomparison of Drought Indicators Based on Thermal Remote Sensing and NLDAS-2 Simulations with U.S. Drought Monitor Classifications , url =. Journal of Hydrometeorology , month =. 2013 ...

    work page doi:10.1175/jhm-d-12-0140.1 2013

  13. [52]

    and Fotiadis, Stathi , copyright =

    Lino, Mario and Cantwell, Chris and Bharath, Anil A. and Fotiadis, Stathi , copyright =. Simulating Continuum Mechanics with Multi-Scale Graph Neural Networks , url =. 2021 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2106.04900 , keywords =

    work page doi:10.48550/arxiv.2106.04900 2021

  14. [60]

    Scalable and Consistent Graph Neural Networks for Distributed Mesh-based Data-driven Modeling , url =

    Barwey, Shivam and Balin, Riccardo and Lusch, Bethany and Patel, Saumil and Balakrishnan, Ramesh and Pal, Pinaki and Maulik, Romit and Vishwanath, Venkatram , copyright =. Scalable and Consistent Graph Neural Networks for Distributed Mesh-based Data-driven Modeling , url =. 2024 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2410.01657 , keywords =

    work page doi:10.48550/arxiv.2410.01657 2024

  15. [63]

    arXiv:2202.07575 [physics]

    Keisler, Ryan , copyright =. Forecasting Global Weather with Graph Neural Networks , url =. 2022 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2202.07575 , keywords =

    work page doi:10.48550/arxiv.2202.07575 2022

  16. [64]

    Graph-based Neural Weather Prediction for Limited Area Modeling , url =

    Oskarsson, Joel and Landelius, Tomas and Lindsten, Fredrik , copyright =. Graph-based Neural Weather Prediction for Limited Area Modeling , url =. 2023 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2309.17370 , keywords =

    work page doi:10.48550/arxiv.2309.17370 2023

  17. [66]

    Connolly, Daniel M

    Connolly, Charlotte J. and Hueholt, Daniel M. and Burt, Melissa A. , date-added =. Datasheets for Earth Science Datasets , url =. Bulletin of the American Meteorological Society , month = apr, number =. 2025 , Bdsk-Url-1 =. doi:10.1175/bams-d-24-0203.1 , issn =

    work page doi:10.1175/bams-d-24-0203.1 2025

  18. [67]

    Geoscientific Model Development17(21), 7915–7962 (2024) https: //doi.org/10.5194/gmd-17-7915-2024

    Olivetti, Leonardo and Messori, Gabriele , date-added =. Do data-driven models beat numerical models in forecasting weather extremes? A comparison of IFS HRES, Pangu-Weather, and GraphCast , url =. Geoscientific Model Development , month = nov, number =. 2024 , Bdsk-Url-1 =. doi:10.5194/gmd-17-7915-2024 , issn =

    work page doi:10.5194/gmd-17-7915-2024 2024

  19. [68]

    arXiv preprint arXiv:2411.07814 , year=

    Schreck, John and Sha, Yingkai and Chapman, William and Kimpara, Dhamma and Berner, Judith and McGinnis, Seth and Kazadi, Arnold and Sobhani, Negin and Kirk, Ben and Gagne, David John , copyright =. Community Research Earth Digital Intelligence Twin (CREDIT) , url =. 2024 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2411.07814 , keywords =

    work page doi:10.48550/arxiv.2411.07814 2024

  20. [69]

    Appa: Bending weather dynamics with latent diffusion models for global data assimilation.arXiv preprint arXiv:2504.18720, 2025

    Andry, G. Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation , url =. 2025 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2504.18720 , keywords =

    work page doi:10.48550/arxiv.2504.18720 2025

  21. [70]

    and Chapman, William and Gagne, David John , copyright =

    Sha, Yingkai and Schreck, John S. and Chapman, William and Gagne, David John , copyright =. Improving AI weather prediction models using global mass and energy conservation schemes , url =. 2025 , Bdsk-Url-1 =. doi:10.48550/ARXIV.2501.05648 , keywords =

    work page doi:10.48550/arxiv.2501.05648 2025

  22. [72]

    Wilks, D. S. , date-added =. ``The Stippling Shows Statistically Significant Grid Points'': How Research Results are Routinely Overstated and Overinterpreted, and What to Do about It , url =. Bulletin of the American Meteorological Society , month = dec, number =. 2016 , Bdsk-Url-1 =. doi:10.1175/bams-d-15-00267.1 , issn =

    work page doi:10.1175/bams-d-15-00267.1 2016

  23. [73]

    A Gentle Introduction to Graph Neural Networks , url =

    Sanchez-Lengeling, Benjamin and Reif, Emily and Pearce, Adam and Wiltschko, Alex , date-added =. A Gentle Introduction to Graph Neural Networks , url =. Distill , month = aug, number =. 2021 , Bdsk-Url-1 =. doi:10.23915/distill.00033 , issn =

    work page doi:10.23915/distill.00033 2021

  24. [74]

    MODIS/Terra+Aqua Downward Shortwave Radiation Daily/3-Hour L3 Global 0.05Deg CMG V062 , url =

    Wang, Dongdong , date-added =. MODIS/Terra+Aqua Downward Shortwave Radiation Daily/3-Hour L3 Global 0.05Deg CMG V062 , url =. 2024 , Bdsk-Url-1 =. doi:10.5067/MODIS/MCD18C1.062 , publisher =

    work page doi:10.5067/modis/mcd18c1.062 2024

  25. [83]

    D., Gentine, P., Barnes, E

    Eyring, Veronika and Collins, William D. and Gentine, Pierre and Barnes, Elizabeth A. and Barreiro, Marcelo and Beucler, Tom and Bocquet, Marc and Bretherton, Christopher S. and Christensen, Hannah M. and Dagon, Katherine and Gagne, David John and Hall, David and Hammerling, Dorit and Hoyer, Stephan and Iglesias-Suarez, Fernando and Lopez-Gomez, Ignacio a...

    work page doi:10.1038/s41558-024-02095-y 2024

  26. [85]

    A global flash drought inventory based on soil moisture volatility , url =

    Osman, Mahmoud and Zaitchik, Benjamin and Otkin, Jason and Anderson, Martha , date-added =. A global flash drought inventory based on soil moisture volatility , url =. Scientific Data , month = sep, number =. 2024 , Bdsk-Url-1 =. doi:10.1038/s41597-024-03809-9 , issn =

    work page doi:10.1038/s41597-024-03809-9 2024

  27. [86]

    Reducing drought emergencies in the Horn of Africa , url =

    Thomas, Evan and Jordan, Elizabeth and Linden, Karl and Mogesse, Beshah and Hailu, Tamene and Jirma, Hussein and Thomson, Patrick and Koehler, Johanna and Collins, Greg , date-added =. Reducing drought emergencies in the Horn of Africa , url =. Science of The Total Environment , month = jul, pages =. 2020 , Bdsk-Url-1 =. doi:10.1016/j.scitotenv.2020.13877...

    work page doi:10.1016/j.scitotenv.2020.138772 2020

  28. [88]

    A contribution to drought resilience in East Africa through groundwater pump monitoring informed by in-situ instrumentation, remote sensing and ensemble machine learning , url =

    Thomas, Evan and Wilson, Daniel and Kathuni, Styvers and Libey, Anna and Chintalapati, Pranav and Coyle, Jeremy , date-added =. A contribution to drought resilience in East Africa through groundwater pump monitoring informed by in-situ instrumentation, remote sensing and ensemble machine learning , url =. Science of The Total Environment , month = aug, pa...

    work page doi:10.1016/j.scitotenv.2021.146486 2021

  29. [91]

    Applying climate reparative finance toward water security , url =

    Thomas, Evan and Ntazinda, Jean and Kathuni, Styvers , date-added =. Applying climate reparative finance toward water security , url =. Science of The Total Environment , month = jun, pages =. 2023 , Bdsk-Url-1 =. doi:10.1016/j.scitotenv.2023.162506 , issn =

    work page doi:10.1016/j.scitotenv.2023.162506 2023

  30. [93]

    , date-added =

    Stallwood, P. , date-added =. Desertification in Africa: Causes, Effects and Solutions , url =. 2022 , Bdsk-Url-1 =

    2022

  31. [94]

    , date-added =

    Chu, J. , date-added =. Climate change to worsen drought, diminish corn yields in Africa , url =. 2017 , Bdsk-Url-1 =

    2017

  32. [95]

    Improving Drought Reslience in East Africa with Sensors, Satellites, and Machine Learning , url =

    Alumni, and Friends , date-added =. Improving Drought Reslience in East Africa with Sensors, Satellites, and Machine Learning , url =. 2019 , Bdsk-Url-1 =

    2019

  33. [100]

    and Pan, Ming and Lin, Peirong and Anghileri, Daniela and Forsythe, Nathan and Pritchard, David M.W

    Gebrechorkos, Solomon H. and Pan, Ming and Lin, Peirong and Anghileri, Daniela and Forsythe, Nathan and Pritchard, David M.W. and Fowler, Hayley J. and Obuobie, Emmanuel and Darko, Deborah and Sheffield, Justin , date-added =. Variability and changes in hydrological drought in the Volta Basin, West Africa , url =. Journal of Hydrology: Regional Studies , ...

    work page doi:10.1016/j.ejrh.2022.101143 2022

  34. [101]

    Expansion of the Sahara Desert and shrinking of frozen land of the Arctic , url =

    Liu, Ye and Xue, Yongkang , date-added =. Expansion of the Sahara Desert and shrinking of frozen land of the Arctic , url =. Scientific Reports , month = mar, number =. 2020 , Bdsk-Url-1 =. doi:10.1038/s41598-020-61085-0 , issn =

    work page doi:10.1038/s41598-020-61085-0 2020

  35. [106]

    and Klar, Manuel and Schlueter, Andreas , date-added =

    Vogel, Peter and Knippertz, Peter and Gneiting, Tilmann and Fink, Andreas H. and Klar, Manuel and Schlueter, Andreas , date-added =. Statistical Forecasts for the Occurrence of Precipitation Outperform Global Models over Northern Tropical Africa , url =. Geophysical Research Letters , month = feb, number =. 2021 , Bdsk-Url-1 =. doi:10.1029/2020gl091022 , issn =

    work page doi:10.1029/2020gl091022 2021

  36. [109]

    and Osborn, Timothy J

    Conway, Declan and Dalin, Carole and Landman, Willem A. and Osborn, Timothy J. , date-added =. Hydropower plans in eastern and southern Africa increase risk of concurrent climate-related electricity supply disruption , url =. Nature Energy , month = dec, number =. 2017 , Bdsk-Url-1 =. doi:10.1038/s41560-017-0037-4 , issn =

    work page doi:10.1038/s41560-017-0037-4 2017

  37. [110]

    Water Conflicts in Sub-Saharan Africa , url =

    Pe. Water Conflicts in Sub-Saharan Africa , url =. Frontiers in Environmental Science , month = mar, publisher =. 2022 , Bdsk-Url-1 =. doi:10.3389/fenvs.2022.863903 , issn =

    work page doi:10.3389/fenvs.2022.863903 2022

  38. [111]

    Does precipitation and runoff variability affect treaty cooperation between states sharing international bilateral rivers? , url =

    Dinar, Ariel and Blankespoor, Brian and Dinar, Shlomi and Kurukulasuriya, Pradeep , date-added =. Does precipitation and runoff variability affect treaty cooperation between states sharing international bilateral rivers? , url =. Ecological Economics , month = oct, number =. 2010 , Bdsk-Url-1 =. doi:10.1016/j.ecolecon.2010.07.036 , issn =

    work page doi:10.1016/j.ecolecon.2010.07.036 2010

  39. [114]

    N. H. Saji and T. Yamagata , date-added =. Possible impacts of Indian Ocean Dipole mode events on global climate , ty =. Climate Research , journal1 =. 2003 , Bdsk-Url-1 =

    2003

  40. [127]

    Hoell, Andrew and Funk, Chris and Magadzire, Tamuka and Zinke, Jens and Husak, Greg , date-added =. El Ni. Climate Dynamics , month = nov, number =. 2014 , Bdsk-Url-1 =. doi:10.1007/s00382-014-2414-z , issn =

    work page doi:10.1007/s00382-014-2414-z 2014

  41. [128]

    , date-added =

    Zaitchik, Benjamin F. , date-added =. Madden-Julian Oscillation impacts on tropical African precipitation , url =. Atmospheric Research , month = feb, pages =. 2017 , Bdsk-Url-1 =. doi:10.1016/j.atmosres.2016.10.002 , issn =

    work page doi:10.1016/j.atmosres.2016.10.002 2017

  42. [129]

    Reconciling Theories for Human and Natural Attribution of Recent East Africa Drying , url =

    Hoell, Andrew and Hoerling, Martin and Eischeid, Jon and Quan, Xiao-Wei and Liebmann, Brant , date-added =. Reconciling Theories for Human and Natural Attribution of Recent East Africa Drying , url =. Journal of Climate , month = feb, number =. 2017 , Bdsk-Url-1 =. doi:10.1175/jcli-d-16-0558.1 , issn =

    work page doi:10.1175/jcli-d-16-0558.1 2017

  43. [130]

    The Perfect Ocean for Drought , url =

    Hoerling, Martin and Kumar, Arun , date-added =. The Perfect Ocean for Drought , url =. Science , month = jan, number =. 2003 , Bdsk-Url-1 =. doi:10.1126/science.1079053 , issn =

    work page doi:10.1126/science.1079053 2003

  44. [131]

    Rainfall variability over the East African coast , url =

    Gamoyo, Majambo and Reason, Chris and Obura, David , date-added =. Rainfall variability over the East African coast , url =. Theoretical and Applied Climatology , month = may, number =. 2014 , Bdsk-Url-1 =. doi:10.1007/s00704-014-1171-6 , issn =

    work page doi:10.1007/s00704-014-1171-6 2014

  45. [132]

    and Barlow, Mathew and Byrne, Michael P

    Allan, Richard P. and Barlow, Mathew and Byrne, Michael P. and Cherchi, Annalisa and Douville, Herv. Advances in understanding large‐scale responses of the water cycle to climate change , url =. Annals of the New York Academy of Sciences , month = apr, number =. 2020 , Bdsk-Url-1 =. doi:10.1111/nyas.14337 , issn =

    work page doi:10.1111/nyas.14337 2020

  46. [133]

    and van de Steeg, J

    Thornton, P.K. and van de Steeg, J. and Notenbaert, A. and Herrero, M. , date-added =. The impacts of climate change on livestock and livestock systems in developing countries: A review of what we know and what we need to know , url =. Agricultural Systems , month = jul, number =. 2009 , Bdsk-Url-1 =. doi:10.1016/j.agsy.2009.05.002 , issn =

    work page doi:10.1016/j.agsy.2009.05.002 2009

  47. [139]

    and Allan, Richard P

    Maidment, Ross I. and Allan, Richard P. and Black, Emily , date-added =. Recent observed and simulated changes in precipitation over Africa , url =. Geophysical Research Letters , month = oct, number =. 2015 , Bdsk-Url-1 =. doi:10.1002/2015gl065765 , issn =

    work page doi:10.1002/2015gl065765 2015

  48. [140]

    Hydroclimatic trends during 1950--2018 over global land , url =

    Dai, Aiguo , date-added =. Hydroclimatic trends during 1950--2018 over global land , url =. Climate Dynamics , month = feb, number =. 2021 , Bdsk-Url-1 =. doi:10.1007/s00382-021-05684-1 , issn =

    work page doi:10.1007/s00382-021-05684-1 1950

  49. [142]

    Adloff, M., M. B. Singer, D. A. MacLeod, K. Michaelides, N. Mehrnegar, E. Hansford, C. Funk, and D. Mitchell, 2022: Sustained water storage in horn of africa drylands dominated by seasonal rainfall extremes. Geophysical Research Letters, 49 (21), doi:10.1029/2022gl099299

    work page doi:10.1029/2022gl099299 2022

  50. [143]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 380 (2238), doi:10.1098/rsta.2021.0288

    AghaKouchak, A., and Coauthors, 2022: Status and prospects for drought forecasting: opportunities in artificial intelligence and hybrid physical statistical forecasting. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 380 (2238), doi:10.1098/rsta.2021.0288

    work page doi:10.1098/rsta.2021.0288 2022

  51. [144]

    M., and J

    Anande, D. M., and J. A. Otkin, 2026: Examining climatological characteristics and drivers of rapid change events in hydrological conditions over africa. 106th Annual Meteorological Society Meeting, 40th Conference on Hydrology

    2026

  52. [145]

    Pure and Applied Geophysics, 179 (4), 1365--1386, doi:10.1007/s00024-022-02988-z

    Ayugi, B., and Coauthors, 2022: Review of meteorological drought in africa: Historical trends, impacts, mitigation measures, and prospects. Pure and Applied Geophysics, 179 (4), 1365--1386, doi:10.1007/s00024-022-02988-z

    work page doi:10.1007/s00024-022-02988-z 2022

  53. [146]

    Basara, J. B., J. I. Christian, R. A. Wakefield, J. A. Otkin, E. H. Hunt, and D. P. Brown, 2019: The evolution, propagation, and spread of flash drought in the central united states during 2012. Environmental Research Letters, 14 (8), 084\,025, doi:10.1088/1748-9326/ab2cc0

    work page doi:10.1088/1748-9326/ab2cc0 2019

  54. [147]

    Rodell, and NASA/GSFC/HSL , 2020: Gldas noah land surface model l4 3 hourly 0.25 x 0.25 degree, version 2.1

    Beaudoing, H., M. Rodell, and NASA/GSFC/HSL , 2020: Gldas noah land surface model l4 3 hourly 0.25 x 0.25 degree, version 2.1. NASA Goddard Earth Sciences Data and Information Services Center, ://disc.gsfc.nasa.gov/datacollection/GLDAS_NOAH025_3H_2.1.html, doi:10.5067/E7TYRXPJKWOQ

    work page doi:10.5067/e7tyrxpjkwoq 2020

  55. [148]

    Bi, K., L. Xie, H. Zhang, X. Chen, X. Gu, and Q. Tian, 2023: Accurate medium-range global weather forecasting with 3d neural networks. Nature, 619 (7970), 533--538, doi:10.1038/s41586-023-06185-3

    work page doi:10.1038/s41586-023-06185-3 2023

  56. [149]

    Br \"o cker, J., A. J. Charlton-Perez, and A. Weisheimer, 2023: A statistical perspective on the signal‐to‐noise paradox. Quarterly Journal of the Royal Meteorological Society, 149 (752), 911--923, doi:10.1002/qj.4440

    work page doi:10.1002/qj.4440 2023

  57. [150]

    Chapman, W. E., J. S. Schreck, Y. Sha, D. J. Gagne, D. Kimpara, L. Zanna, K. J. Mayer, and J. Berner, 2025: Camulator: Fast emulation of the community atmosphere model. arXiv, ://arxiv.org/abs/2504.06007, doi:10.48550/ARXIV.2504.06007

    work page doi:10.48550/arxiv.2504.06007 2025

  58. [151]

    FuXi : a cascade machine learning forecasting system for 15-day global weather forecast

    Chen, L., X. Zhong, F. Zhang, Y. Cheng, Y. Xu, Y. Qi, and H. Li, 2023: Fuxi: a cascade machine learning forecasting system for 15-day global weather forecast. npj Climate and Atmospheric Science, 6 (1), doi:10.1038/s41612-023-00512-1

    work page doi:10.1038/s41612-023-00512-1 2023

  59. [152]

    Christian, J. I., J. B. Basara, E. D. Hunt, J. A. Otkin, J. C. Furtado, V. Mishra, X. Xiao, and R. M. Randall, 2021: Global distribution, trends, and drivers of flash drought occurrence. Nature Communications, 12 (1), doi:10.1038/s41467-021-26692-z

    work page doi:10.1038/s41467-021-26692-z 2021

  60. [153]

    Christian, J. I., J. B. Basara, E. D. Hunt, J. A. Otkin, and X. Xiao, 2020: Flash drought development and cascading impacts associated with the 2010 russian heatwave. Environmental Research Letters, 15 (9), 094\,078, doi:10.1088/1748-9326/ab9faf

    work page doi:10.1088/1748-9326/ab9faf 2020

  61. [154]

    Christian, J. I., J. B. Basara, L. E. Lowman, X. Xiao, D. Mesheske, and Y. Zhou, 2022: Flash drought identification from satellite-based land surface water index. Remote Sensing Applications: Society and Environment, 26, 100\,770, doi:10.1016/j.rsase.2022.100770

    work page doi:10.1016/j.rsase.2022.100770 2022

  62. [155]

    Christian, J. I., J. B. Basara, J. A. Otkin, E. D. Hunt, R. A. Wakefield, P. X. Flanagan, and X. Xiao, 2019: A methodology for flash drought identification: Application of flash drought frequency across the united states. Journal of Hydrometeorology, 20 (5), 833--846, doi:10.1175/jhm-d-18-0198.1

    work page doi:10.1175/jhm-d-18-0198.1 2019

  63. [156]

    Dale, A., C. Fant, K. Strzepek, M. Lickley, and S. Solomon, 2017: Climate model uncertainty in impact assessments for agriculture: A multi‐ensemble case study on maize in sub‐saharan africa. Earth's Future, 5 (3), 337--353, doi:10.1002/2017ef000539

    work page doi:10.1002/2017ef000539 2017

  64. [157]

    Deman, V. M. H., A. Koppa, W. Waegeman, D. A. MacLeod, M. Bliss Singer, and D. G. Miralles, 2022: Seasonal prediction of horn of africa long rains using machine learning: The pitfalls of preselecting correlated predictors. Frontiers in Water, 4, doi:10.3389/frwa.2022.1053020

    work page doi:10.3389/frwa.2022.1053020 2022

  65. [158]

    Pradhan, A

    Dikshit, A., B. Pradhan, A. Huete, and H.-J. Park, 2022 a : Spatial based drought assessment: Where are we heading? a review on the current status and future. Science of The Total Environment, 844, 157\,239, doi:10.1016/j.scitotenv.2022.157239

    work page doi:10.1016/j.scitotenv.2022.157239 2022

  66. [159]

    Pradhan, and M

    Dikshit, A., B. Pradhan, and M. Santosh, 2022 b : Artificial neural networks in drought prediction in the 21st century a scientometric analysis. Applied Soft Computing, 114, 108\,080, doi:10.1016/j.asoc.2021.108080

    work page doi:10.1016/j.asoc.2021.108080 2022

  67. [160]

    Doi, T., S. K. Behera, and T. Yamagata, 2022: On the predictability of the extreme drought in east africa during the short rains season. Geophysical Research Letters, 49 (22), doi:10.1029/2022gl100905

    work page doi:10.1029/2022gl100905 2022

  68. [161]

    arXiv, ://arxiv.org/abs/2509.17601, doi:10.48550/ARXIV.2509.17601

    Dunstan, T., and Coauthors, 2025: Fastnet: Improving the physical consistency of machine-learning weather prediction models through loss function design. arXiv, ://arxiv.org/abs/2509.17601, doi:10.48550/ARXIV.2509.17601

    work page doi:10.48550/arxiv.2509.17601 2025

  69. [162]

    Endris, H. S., L. Hirons, Z. T. Segele, M. Gudoshava, S. Woolnough, and G. A. Artan, 2021: Evaluation of the skill of monthly precipitation forecasts from global prediction systems over the greater horn of africa. Weather and Forecasting, 36 (4), 1275--1298, doi:10.1175/waf-d-20-0177.1

    work page doi:10.1175/waf-d-20-0177.1 2021

  70. [163]

    Macharia, J

    Fankhauser, K., D. Macharia, J. Coyle, S. Kathuni, A. McNally, K. Slinski, and E. Thomas, 2022: Estimating groundwater use and demand in arid kenya through assimilation of satellite data and in-situ sensors with machine learning toward drought early action. Science of The Total Environment, 831, 154\,453, doi:10.1016/j.scitotenv.2022.154453

    work page doi:10.1016/j.scitotenv.2022.154453 2022

  71. [164]

    Chihaoui, and A

    Ferchichi, A., M. Chihaoui, and A. Ferchichi, 2024: Spatio-temporal modeling of climate change impacts on drought forecast using generative adversarial network: A case study in africa. Expert Systems with Applications, 238, 122\,211, doi:10.1016/j.eswa.2023.122211

    work page doi:10.1016/j.eswa.2023.122211 2024

  72. [165]

    Ford, T. W., J. A. Otkin, S. M. Quiring, J. Lisonbee, M. Woloszyn, J. Wang, and Y. Zhong, 2023: Flash drought indicator intercomparison in the united states. Journal of Applied Meteorology and Climatology, 62 (12), 1713--1730, doi:10.1175/jamc-d-23-0081.1

    work page doi:10.1175/jamc-d-23-0081.1 2023

  73. [166]

    Gavahi, and H

    Foroumandi, E., K. Gavahi, and H. Moradkhani, 2024: Generative adversarial network for real‐time flash drought monitoring: A deep learning study. Water Resources Research, 60 (5), doi:10.1029/2023wr035600

    work page doi:10.1029/2023wr035600 2024

  74. [167]

    Sulla-Menashe, 2022: Modis/terra+aqua land cover type yearly l3 global 0.05deg cmg v061

    Friedl, M., and D. Sulla-Menashe, 2022: Modis/terra+aqua land cover type yearly l3 global 0.05deg cmg v061. NASA EOSDIS Land Processes Distributed Active Archive Center, ://lpdaac.usgs.gov/products/mcd12c1v061/, doi:10.5067/MODIS/MCD12C1.061

    work page doi:10.5067/modis/mcd12c1.061 2022

  75. [168]

    Quarterly Journal of the Royal Meteorological Society, 144 (S1), 360--383, doi:10.1002/qj.3266

    Funk, C., and Coauthors, 2018: Examining the role of unusually warm indo‐pacific sea‐surface temperatures in recent african droughts. Quarterly Journal of the Royal Meteorological Society, 144 (S1), 360--383, doi:10.1002/qj.3266

    work page doi:10.1002/qj.3266 2018

  76. [169]

    Journal of Hydrology, 604, 127\,224, doi:10.1016/j.jhydrol.2021.127224

    Gou, Q., and Coauthors, 2022: Application of an improved spatio-temporal identification method of flash droughts. Journal of Hydrology, 604, 127\,224, doi:10.1016/j.jhydrol.2021.127224

    work page doi:10.1016/j.jhydrol.2021.127224 2022

  77. [170]

    Climate Risk Management, 30, 100\,247, doi:10.1016/j.crm.2020.100247

    Guido, Z., and Coauthors, 2020: Farmer forecasts: Impacts of seasonal rainfall expectations on agricultural decision-making in sub-saharan africa. Climate Risk Management, 30, 100\,247, doi:10.1016/j.crm.2020.100247

    work page doi:10.1016/j.crm.2020.100247 2020

  78. [171]

    Guo, Y., T. Zhou, W. Jiang, B. Wu, L. Sun, and R. Jin, 2024: Maximizing the impact of deep learning on subseasonal-to-seasonal climate forecasting: The essential role of optimization. arXiv, ://arxiv.org/abs/2411.16728, doi:10.48550/ARXIV.2411.16728

    work page doi:10.48550/arxiv.2411.16728 2024

  79. [172]

    doi:10.24381/cds.adbb2d47 , note =

    Hersbach, H., and Coauthors, 2023: Era5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS), ://cds.climate.copernicus.eu/doi/10.24381/cds.adbb2d47, doi:10.24381/CDS.ADBB2D47

    work page doi:10.24381/cds.adbb2d47 2023

  80. [173]

    Funk, and M

    Hoell, A., C. Funk, and M. Barlow, 2013: The regional forcing of northern hemisphere drought during recent warm tropical west pacific ocean la ni \ n a events. Climate Dynamics, 42 (11--12), 3289--3311, doi:10.1007/s00382-013-1799-4

    work page doi:10.1007/s00382-013-1799-4 2013

Showing first 80 references.