laureboudinaud/s2-cropland-abandonment

Google Earth Engine pipeline processing Sentinel-2 imagery to detect conflict-driven cropland abandonment in Sahelian landscapes. Underpins a peer-reviewed paper (ISPRS 2021) and WFP emergency food assistance targeting across central Mali. # s2-cropland-abandonment Assessing conflict-driven cropland abandonment at locality-level in Sahelian landscapes (here: central Mali) processing Sentinel-2 time series in Google Earth Engine. --- ## Overview Violent conflict disrupts agricultural systems in ways that are nearly impossible to document on the ground. This repository contains the remote sensing pipeline used to detect and categorize cropland abandonment across conflict-affected areas in central Mali, using Sentinel-2 multispectral imagery processed in Google Earth Engine. The methodology has since been replicated across similar environments (Burkina Faso, Niger, Nigeria, and Chad) from 2020 onwards. The core product is the **3-Period TimeScan (3PTS)** — a Red-Green-Blue composite that encodes the maximum NDVI value at three distinct phases of the agricultural season into a single image per year. Active cropland appears in vivid colour; bare or abandoned land appears dark or grey. By comparing composites before and after the onset of violence, reductions in cultivated area become immediately visible at the locality level. The analysis identified ~500 localities with significant cropland losses across the Mopti region in 2019, with a strong spatial correlation with ACLED-registered violent events. Results fed into the national food security framework (Cadre Harmonisé), contributing to the estimation of roughly 760,000 food-insecure persons for the 2020 lean season and targeting 65,000+ beneficiaries for early emergency assistance. --- ## Publication & Coverage **Peer-reviewed paper:** > **Boudinaud, L. and Orenstein, S. A. (2021)** > *Assessing Cropland Abandonment from Violent Conflict in Central Mali with Sentinel-2 and Google Earth Engine.* > ISPRS Archives, XLVI-4/W2-2021, pp. 9–16. > DOI: 10.5194/isprs-archives-XLVI-4-W2-2021-9-2021 > Presented at FOSS4G 2021 Academic Track. **Practitioner write-up:** > **Boudinaud, L., Kamara, N. and Ibrahim, A. (2021)** > *Using satellite imagery in c …