Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming

Modifications in precipitation regimes significantly affect various socio-economic sectors, including agriculture and water resource management. Although the rainfall regimes characterizing Central Africa (CA) have just recently been described, it is equally urgent to investigate potential changes i...

Full description

Saved in:
Bibliographic Details
Main Authors: Alain T Tamoffo, Fernand L Mouassom, Torsten Weber
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:Environmental Research Communications
Subjects:
Online Access:https://doi.org/10.1088/2515-7620/adf2f9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849246264711970816
author Alain T Tamoffo
Fernand L Mouassom
Torsten Weber
author_facet Alain T Tamoffo
Fernand L Mouassom
Torsten Weber
author_sort Alain T Tamoffo
collection DOAJ
description Modifications in precipitation regimes significantly affect various socio-economic sectors, including agriculture and water resource management. Although the rainfall regimes characterizing Central Africa (CA) have just recently been described, it is equally urgent to investigate potential changes in their spatial extent under different global warming pathways, which motivates the present study. For this purpose, we utilized results from the dynamical downscaling performed by regional climate models (RCMs) under the CORDEX-CORE (Coordinated Regional Climate Downscaling Experiment–Coordinated Output for Regional Evaluations) initiative. The warming pathways are based on low (RCP2.6) and high (RCP8.5) emission scenarios. The K-means clustering technique is employed to classify areas with homogeneous rainfall regimes. Our findings indicate that the ability of experiments to mimic the spatial patterns of these subregions is model-dependent. REMO and CCLM5 RCMs outperform RegCM4, achieving the highest Adjusted Rand (AR) index values compared to the observational datasets CHIRPS2 and TAMSAT3.1. Projections based on individual experiments and the multimodel ensemble-mean suggest that the warming level will influence clusters’ spatial extent. Broadly, the ensemble mean shows that an expansion of Equatorial CA is projected (4.8% and 9.7%, respectively), while a contraction of Southern CA is anticipated (4.2% and 4.5%, respectively), consistently under both scenarios. In contrast, the signal of change in Northern CA differs between the two warming pathways. Under the highly mitigated RCP2.6 scenario, an expansion of the cluster is projected (1%), whereas the low-mitigation RCP8.5 scenario projects a shrinking of its spatial extent (0.8%).
format Article
id doaj-art-6a4bf431de634fab87d7b988ebce731f
institution Kabale University
issn 2515-7620
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series Environmental Research Communications
spelling doaj-art-6a4bf431de634fab87d7b988ebce731f2025-08-20T03:58:32ZengIOP PublishingEnvironmental Research Communications2515-76202025-01-017707100710.1088/2515-7620/adf2f9Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warmingAlain T Tamoffo0https://orcid.org/0000-0001-8482-8881Fernand L Mouassom1https://orcid.org/0000-0001-7995-3880Torsten Weber2https://orcid.org/0000-0002-8133-8622Climate Service Center Germany (GERICS), Helmholtz-Zentrum Hereon, Hamburg, GermanyDepartment of Mathematics and Statistics, Memorial University of Newfoundland , St. John’s, Newfoundland, CanadaClimate Service Center Germany (GERICS), Helmholtz-Zentrum Hereon, Hamburg, GermanyModifications in precipitation regimes significantly affect various socio-economic sectors, including agriculture and water resource management. Although the rainfall regimes characterizing Central Africa (CA) have just recently been described, it is equally urgent to investigate potential changes in their spatial extent under different global warming pathways, which motivates the present study. For this purpose, we utilized results from the dynamical downscaling performed by regional climate models (RCMs) under the CORDEX-CORE (Coordinated Regional Climate Downscaling Experiment–Coordinated Output for Regional Evaluations) initiative. The warming pathways are based on low (RCP2.6) and high (RCP8.5) emission scenarios. The K-means clustering technique is employed to classify areas with homogeneous rainfall regimes. Our findings indicate that the ability of experiments to mimic the spatial patterns of these subregions is model-dependent. REMO and CCLM5 RCMs outperform RegCM4, achieving the highest Adjusted Rand (AR) index values compared to the observational datasets CHIRPS2 and TAMSAT3.1. Projections based on individual experiments and the multimodel ensemble-mean suggest that the warming level will influence clusters’ spatial extent. Broadly, the ensemble mean shows that an expansion of Equatorial CA is projected (4.8% and 9.7%, respectively), while a contraction of Southern CA is anticipated (4.2% and 4.5%, respectively), consistently under both scenarios. In contrast, the signal of change in Northern CA differs between the two warming pathways. Under the highly mitigated RCP2.6 scenario, an expansion of the cluster is projected (1%), whereas the low-mitigation RCP8.5 scenario projects a shrinking of its spatial extent (0.8%).https://doi.org/10.1088/2515-7620/adf2f9Central AfricaK-means clusteringhomogeneous subregionsspatial changesglobal warming
spellingShingle Alain T Tamoffo
Fernand L Mouassom
Torsten Weber
Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
Environmental Research Communications
Central Africa
K-means clustering
homogeneous subregions
spatial changes
global warming
title Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
title_full Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
title_fullStr Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
title_full_unstemmed Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
title_short Machine-learning unravels spatial shifting in homogeneous rainfall subregions in Central Africa under global warming
title_sort machine learning unravels spatial shifting in homogeneous rainfall subregions in central africa under global warming
topic Central Africa
K-means clustering
homogeneous subregions
spatial changes
global warming
url https://doi.org/10.1088/2515-7620/adf2f9
work_keys_str_mv AT alainttamoffo machinelearningunravelsspatialshiftinginhomogeneousrainfallsubregionsincentralafricaunderglobalwarming
AT fernandlmouassom machinelearningunravelsspatialshiftinginhomogeneousrainfallsubregionsincentralafricaunderglobalwarming
AT torstenweber machinelearningunravelsspatialshiftinginhomogeneousrainfallsubregionsincentralafricaunderglobalwarming