Postprocessing East African Rainfall Forecasts Using a Generative Machine Learning Model

Abstract Existing weather models are known to have poor skill at forecasting rainfall over East Africa. Improved forecasts could reduce the effects of extreme weather events and provide significant socioeconomic benefits to the region. We present a novel machine learning (ML)‐based method to improve...

Full description

Saved in:
Bibliographic Details
Main Authors: Bobby Antonio, Andrew T. T. McRae, David MacLeod, Fenwick C. Cooper, John Marsham, Laurence Aitchison, Tim N. Palmer, Peter A. G. Watson
Format: Article
Language:English
Published: American Geophysical Union (AGU) 2025-03-01
Series:Journal of Advances in Modeling Earth Systems
Subjects:
Online Access:https://doi.org/10.1029/2024MS004796
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Existing weather models are known to have poor skill at forecasting rainfall over East Africa. Improved forecasts could reduce the effects of extreme weather events and provide significant socioeconomic benefits to the region. We present a novel machine learning (ML)‐based method to improve precipitation forecasts in East Africa, using postprocessing based on a conditional generative adversarial network (cGAN). This addresses the challenge of realistically representing tropical rainfall, where convection dominates and is poorly simulated in conventional global forecast models. We postprocess hourly forecasts made by the European Centre for Medium‐Range Weather Forecasts Integrated Forecast System at 6–18 hr lead times, at 0.1° resolution. We combine the cGAN predictions with a novel neighborhood version of quantile mapping, to integrate the strengths of ML and conventional postprocessing. Our results indicate that the cGAN substantially improves the diurnal cycle of rainfall, and improves predictions up to the 99.9th percentile (∼10mm/hr). This improvement extends to the March–May 2018 season, which had extremely high rainfall, indicating that the approach has some ability to generalize to more extreme conditions. We explore the potential for the cGAN to produce probabilistic forecasts and find that the spread of this ensemble broadly reflects the predictability of the observations, but is also characterized by a mixture of under‐ and over‐dispersion. Overall our results demonstrate how the strengths of ML and conventional postprocessing methods can be combined, and illuminate what benefits ML approaches can bring to this region.
ISSN:1942-2466