nanaamuah/ghana_vegetation_drought

Reproducible analysis of vegetation condition and rainfall deficit across Ghana, 2001–2025 # Vegetation Condition and Rainfall Deficit Across Ghana, 2001–2025 A reproducible analysis of standardised MODIS EVI anomalies against multi-scale CHIRPS rainfall deficit across three analysis belts in Ghana, covering 25 years and 1,725 satellite composites. **Full report:** View the rendered HTML report **Headline findings** - The delay between rainfall deficit and vegetation response follows a clean north-to-south gradient: about 32 days in the North, 16 days in the Middle, and none detectable in the South. - Longer-accumulation rainfall (SPI-6) retains predictive value only in the North, consistent with soil-moisture carryover through a single long dry season. - Below-normal vegetation composites declined significantly in the North (Sen's slope −0.268 per year, p = 0.0003), but this pattern is inseparable from a concentration of poor conditions early in the record and should not be read as falling drought risk. - Associations are moderate throughout: peak correlations of 0.202 to 0.419, explaining roughly 4% to 18% of variance in the vegetation anomaly. **Scope.** These are statistical patterns and associations. They are not a formal drought declaration, evidence of crop loss, or proof that rainfall alone drove the vegetation changes observed. --- ## Contents - Research questions - Study design - Quick start - Project structure - Pipeline - Methods - Results - Limitations - Data sources and licensing --- ## Research questions Vegetation greenness in Ghana changes naturally between wet and dry seasons, so a low value during the dry season is not by itself evidence of poor condition. This project removes the expected seasonal pattern and examines departures from it. 1. In which years and parts of Ghana did vegetation fall below the level normally expected for that time of year? 2. How closely were these anomalies associated with rainfall deficit? 3. How long after a rainfall deficit did the strongest vegetation response occur? 4. Did the strength and t …