A multivariate flash drought climatology based on the framework of the flash drought intensity index (FDII)

The high impact of severe flash drought events during the past decade has greatly increased public awareness and the demand for improved understanding and monitoring of this type of drought. All existing flash drought indices singularly focus on evaporative demand, soil moisture, or vegetation stress, which limits the capability of depicting compound climate phenomena such as flash drought. Here we develop a multivariate climatology of flash drought occurrence and severity across the U. S. within the framework of the flash drought intensity index (FDII) that accounts for both the rapid intensification and the actual drought severity. The FDII is applied to a set of drought monitoring datasets including the Standardized Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), Evaporative Demand Drought Index (EDDI), Evaporative Stress Index (ESI) and soil moisture (SM) in the upper 40cm. A combined FDII is then generated by averaging the FDIIs from all of the monitoring datasets when at least two of them indicate flash drought conditions. The combined FDII and the individual FDIIs are also compared to the FDII computed using the U. S. Drought Monitor (USDM), which is a multivariate drought monitoring tool itself.

The maps of FDII climatology for the growing season (from May to October) between 2000 and 2021 show that the SPI and SPEI FDIIs are largest in the central and western U. S., the ESI, SM, and combined FDIIs are largest in the central U. S., whereas the USDM FDII is largest in the Great Plains, the Rocky Mountains, and the southeast U. S. The spatial patterns of the FDII climatology are primarily determined by the drought severity component rather than the intensification component of the FDII. Interestingly, the ESI FDII indicates the lowest occurrence frequency of flash drought, pointing toward it being a reliable indicator of flash drought. Lagged covariances of the combined FDII with the individual FDIIs suggest that the combined FDII largely covaries simultaneously with the SPI and SPEI FDIIs, lagging the EDDI, ESI and SM FDIIs and leading the USDM FDII. In addition, the climatology of flash drought event frequency, onset dates, duration, and maximum FDII during the growing season is analyzed for four U. S. regions. The multivariate climatology based on the FDII framework provides valuable information regarding the timing and severity of flash drought depicted by each dataset and can inform development of a comprehensive flash drought monitor.