Comparison of meteorological and satellite-based drought indices as yield predictors of Spanish cereals.

Highlights • Models based on VCI/TCI explained 70% of wheat and barley yield level annual variability. • Validation tests at agricultural districts confirmed better fit of VCI/TCI-based crop models compared to SPI-based models. • Both families of indices (meteorological and satellite) fail to explain yield variability of maize, mostly rainfed in Spain. • VCI/TCI-based models performed better than SPI-based as in-season, real-time yield forecasting tools. Abstract In the context of global warming, as drought episodes become increasingly frequent, it is crucial to accurately measure the impacts of droughts on the overall performance of agrosystems. This study aims to compare the effectiveness of meteorological drought indices against satellite-based agronomical drought indices as crop yield explanatory factors in statistical models calibrated at a local scale. The analysis is conducted in Spain using a spatially detailed, 12-year (2003–2015) dataset on crop yields, including different types of cereals. Yields and drought indices were spatially aggregated at the agricultural district level. The Standardised Precipitation Index (SPI), computed at different temporal aggregation levels, and two satellite-based drought indices, the Vegetation Condition Index (VCI) and the Temperature Condition Index (TCI), were used to characterise the dynamics of drought severity conditions in the study area. Models resting on satellite-based indices showed higher performance in explaining yield levels as well as yield anomalies for all the crops evaluated. In particular, VCI/TCI models of winter wheat and barley were able to explain 70% and 40% of annual crop yield level and crop yield anomaly variability, respectively. We also observed gains in explanatory power when models for climate zones (instead of models at the national scale) were considered. All the results were cross-validated on subsamples of the whole dataset and on models fitted to individual agricultural districts and their predictive accuracy was assessed with a real-time forecasting exercise. Results from this study highlight the potential for including satellite-based drought indices in agricultural decision support systems (e.g. agricultural drought early warning systems, crop yield forecasting models or water resource management tools) complementing meteorological drought indices derived from precipitation grids. [ABSTRACT FROM AUTHOR]