Effectiveness of drought indices in identifying impacts on major crops across the USA

In North America, the occurrence of extreme drought events has increased significantly in number and severity over the last decades. Past droughts have contributed to lower agricultural productivity in major farming and ranching areas across the US. We evaluated the relationship between drought indices and crop yields across the US for the period 1961-2014. In order to assess the correlations with yields from the major cash crops in the country, we calculated several drought indices commonly used to monitor drought conditions, including 4 Palmer-based and 3 multiscalar indices (Standardized Precipitation Index, Standardized Precipitation Evapotranspiration Index, Standardized Precipitation Drought Index). The 3 multiscalar drought indices were aggregated at 1 to 12 mo timescales. Besides the quantification of the similarities or differences between these drought indices using Pearson correlation coefficients, we identified spatial patterns illustrating this relationship. The results demonstrate that the flexible multiscalar indices can identify drought impacts on different types of crops for a wide range of time periods. The differences in spatial and temporal distribution of the correlations depend on the crop and timescale analyzed, but can also be found within the same type of crop. The moisture conditions during summer and shorter timescales (1 to 3 mo) turn out to be a determining factor for barley, corn, cotton and soybean yields. Therefore, the use of multiscalar drought indices based on both precipitation and the atmospheric evaporative demand (SPEI and SPDI) seems to be a prudent recommendation.