Résumé
In the Sudano-Sahelian zone of Cameroon, agriculture — predominantly rainfed — constitutes a major economic sector, underpinning food security but remaining vulnerable to rainfall variability. This study explores statistical relationships between 25 rainfall indices and the yields of maize (Zea mays L.) from 1999 to 2021 and cotton (Gossypium hirsutum L.) from 1991 to 2010, employing Pearson correlation tests across temporal and spatial scales. At the temporal scale, results indicate that maize and cotton yields respond similarly to two indices: the end of the rainy season (EOS) and cumulative dry spells (CDS15). A longer rainy season is associated with higher yields for both crops, whereas dry spells exert a negative influence across the entire study area. However, spatial analyses reveal significant local variations in crop responses. Specifically, maize yields exhibit positive correlations with indices such as rainfall amount (PRCPSEAS), rainy days (R1mm), wet days (R20mm), season length (SL) in the northern and southwestern parts of the study area, reflecting the importance of consistent moisture availability for optimal growth. Conversely, cotton yields are strongly negatively correlated with these same indices in the northern, northwestern, central, and southeastern parts, likely reflecting the crop's lower tolerance for excessively humid conditions. The findings highlight the need for crop-specific adaptation strategies to rainfall variability, including the selection of appropriate crop varieties, adjustments to planting calendars, improved water management practices, particularly in the context of increasing rainfall trends. Policymakers could invest in localized agro-climatic forecasting systems and improve the integration of climate data into agricultural advisory services.