A comparison of drought indices for crop yield loss detection: The role of green-up onset alignment and spatial resolution

Central Plateau of Mexico This study investigated the relationships between yield anomalies of rainfed crops and six drought indices: evaporative stress indices (ESIs) derived from the ALEXI, MOD16, and GLEAM evapotranspiration products; the MODIS NDVI anomaly; the standardized precipitation index (SPI) based on spatially interpolated rain gauge data; and the Mexico Drought Monitor (MDM). Drought indices were calculated as a function of calendar time (all indices) and time from MODIS-derived green-up onset (all indices except MDM). To evaluate the effects of spatial resolution, indices were analyzed both at their native spatial resolutions and at a common 0.1° grid. Significant delays in satellite-derived green-up onset were observed during drought years, likely due to postponed sowing. Aligning drought variables to green-up onset improved temporal synchronization of crop growth stages across years. For corn, this alignment revealed that the strongest relationships between yield anomalies and cumulative NDVI anomaly/evaporative stress indices occurred during the mid-season growth stage. Degrading the spatial resolution generally weakened the drought index-yield anomaly relationships, particularly for the NDVI anomaly. The ALEXI-based ESI and MODIS NDVI anomaly showed the strongest relationships with crop yield anomalies. Mixed results were obtained for ESI MODIS , while the performance of ESI GLEAM was likely affected by its lower spatial resolution. Both the SPI and the MDM showed the weakest relationships. • Green-up onset alignment improved crop growth stage synchronization across years. • Coarser spatial resolution weakened drought index-yield anomaly relationships. • ALEXI ESI and MODIS NDVI showed strongest relationships with crop yields. • SPI and Mexico Drought Monitor showed weakest relationships with crop yields. • Cumulative ESIs/NDVI anom -corn yield anomaly relationships peaked mid-season.