Preprints
https://doi.org/10.5194/egusphere-2025-1308
https://doi.org/10.5194/egusphere-2025-1308
14 Apr 2025
 | 14 Apr 2025
Status: this preprint is open for discussion and under review for Hydrology and Earth System Sciences (HESS).

Decoding the Architecture of Drought: SHAP-Enhanced Insights into the Climate Forces Reshaping the Sahel

Fabio Di Nunno, Mehmet Berkant Yildiz, and Francesco Granata

Abstract. The Sahel region faces increasing drought variability, driven by complex interactions between climatic indices and hydrological extremes. This study explores the correlation between the Standardized Precipitation Evapotranspiration Index (SPEI) and multiple climatic indices using trend analysis, cross-correlation, and an innovative SHAP-driven (SHapley Additive exPlanations) clustering approach. The Seasonal Kendall (SK) test identified statistically significant decreasing SPEI-12 trends in 57.5 % of the gridded data, particularly in the western (Senegal and Mauritania) and southeastern regions of the Sahel (South Sudan). In contrast, 19.3 % of the data, primarily in the central-western Sahel (Burkina Faso and Niger), exhibited statistically significant increasing trends. Correlation analysis between SPEI-12 and climatic indices revealed strong negative relationships between SPEI and Global Mean Temperature (GMT, correlation coefficient up to -0.76) and Indo-Pacific Warm Pool (IPWP, -0.71), underscoring their role in drought intensification. Conversely, the Atlantic Multidecadal Oscillation (AMO, 0.40) showed a positive correlation, emphasizing its influence on regional hydrology. Clustering delineated three distinct drought-prone regions, with Cluster C2, including the Sahel regions of Senegal, Mauritania, and Mali, (western Sahel) experiencing the most severe drought intensification (Z = -5.04). The SHAP-driven clustering approach, which incorporates a Machine Learning (ML) Random Forest (RF) model to classify data points into clusters, allowing the SHAP method to quantify the influence of each climatic variable on the clustering process, further highlights the dominant role of AMO and the North Tropical Atlantic Index (NTA) in shaping regional drought dynamics. This study provides a novel framework integrating explainable AI into drought assessment, offering valuable insights for climate adaptation and water resource management in the Sahel.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Share
Fabio Di Nunno, Mehmet Berkant Yildiz, and Francesco Granata

Status: open (until 30 May 2025)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Fabio Di Nunno, Mehmet Berkant Yildiz, and Francesco Granata
Fabio Di Nunno, Mehmet Berkant Yildiz, and Francesco Granata

Viewed

Total article views: 79 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
65 10 4 79 7 4 3
  • HTML: 65
  • PDF: 10
  • XML: 4
  • Total: 79
  • Supplement: 7
  • BibTeX: 4
  • EndNote: 3
Views and downloads (calculated since 14 Apr 2025)
Cumulative views and downloads (calculated since 14 Apr 2025)

Viewed (geographical distribution)

Total article views: 79 (including HTML, PDF, and XML) Thereof 79 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 23 Apr 2025
Download
Short summary
Droughts in the Sahel are becoming more severe and unpredictable due to climate change. This study explores how large-scale climate patterns influence drought trends in the region. Using advanced data analysis and machine learning, we identified key climate factors driving droughts and mapped areas most at risk. Our findings provide a clearer understanding of regional drought dynamics, helping policymakers and communities develop effective strategies for water management and climate adaptation.
Share