MULTI-SOURCE SATELLITE DATA ASSESSMENT OF LAND–ATMOSPHERE COUPLING IN THE SUDAN SAVANNA: EVIDENCE OF DROUGHT-AMPLIFIED FEEDBACKS

This repository contains the reproducible Python/Jupyter Notebook workflow used to assess multi-scale land–atmosphere coupling in the Sudan Savanna of Kano State, Nigeria, for the period 2016–2025. The analysis integrates satellite-derived land surface temperature (LST), soil moisture (SM), normalized difference vegetation index (NDVI), and rainfall data to investigate spatial and temporal coupling, differences between wet and dry years, rainfall-mediated relationships, and spatial heterogeneity in land–atmosphere interactions. The workflow implements: Annual spatial correlations between soil moisture, land surface temperature, NDVI, and rainfall with correction for spatial dependence and bootstrap confidence intervals. Pixel-wise temporal correlations for 2016–2025 with false discovery rate (FDR) correction. Wet- and dry-year classification using rainfall anomalies and composite analysis to assess drought-related changes in soil moisture–land surface temperature coupling. Partial correlation analysis controlling for rainfall to examine rainfall-mediated relationships between soil moisture, land surface temperature, and vegetation. Spatial mapping of coupling metrics using interpolation and ordinary kriging, together with local spatial analysis of dry–wet coupling differences. Sensitivity analyses evaluating the influence of wet/dry classification thresholds and local-neighbourhood specifications. Additional statistical visualizations, including correlation matrices, annual correlation trends, regression analyses, and robust temporal trend assessment using Mann–Kendall and Theil–Sen methods. The repository is intended to support transparency, reproducibility, and reuse of the analytical workflow associated with the manuscript: “Multi-Source Satellite Data Assessment of Land–Atmosphere Coupling in the Sudan Savanna: Evidence of Drought-Amplified Feedbacks.” The original satellite datasets are not redistributed in this repository. Users should obtain the required datasets from their respective data providers and organize the annual raster files according to the file-naming conventions documented in the repository README.