Rainfall structure and modelled surface soil moisture in relation to agricultural response in rainfed lands of eastern Sudan

Dryland farming can be exposed to markedly different rainfall conditions within seasons receiving similar total rainfall. This study examined rainfall structure, modelled surface soil moisture, and administrative agricultural response in Gedaref State, eastern Sudan, during 1982–2020 using daily CHIRPS rainfall, monthly FLDAS Noah soil moisture, and annual sorghum and sesame records. CHIRPS showed moderate seasonal correlation with the available monthly station record ( r = 0.625), but agreement in absolute seasonal rainfall was limited (bias = − 77.5 mm; RMSE = 130.0 mm). A CHIRPS-derived Daily Rainfall-Structure Risk Index (DRSRI) combined seasonal rainfall deficit, maximum June–September consecutive dry days, onset delay, and rainfall concentration. DRSRI was associated with lower modelled FLDAS surface soil moisture ( r = − 0.776, p < 0.001). In rolling-origin evaluation, the four rainfall variables produced only a small and non-significant improvement over seasonal rainfall alone (RMSE 0.544 versus 0.554; p = 0.739), while DRSRI alone had a higher RMSE of 0.681. The clearest agricultural association was with the detrended sorghum harvested-area ratio ( r = − 0.448; FDR-adjusted p = 0.015); yield associations were not significant after correction and the harvested-area relationship weakened after 1996. Eight years met the primary compound-risk definition and showed lower modelled surface soil moisture and sorghum harvested-area ratio anomalies, although the classification was sensitive to threshold choice. DRSRI is therefore used here as a compact regional screening indicator of within-season rainfall structure rather than as a replacement for seasonal rainfall or its individual components.