Spatiotemporal Comparison of Drought Metrics over the Western United States

Abstract

Drought definitions and drought metrics come in many different forms and it is not always clear which definition or metric is most useful for describing or forecasting drought impacts. Drought metrics range from those which only incorporate precipitation anomalies to more comprehensive measures such as the United States Drought Monitor (USDM), which incorporates on-the-ground observations of drought impacts to arrive at a consensus-based drought status. Here, we develop a method to compare drought metrics, as reported in the Climate Toolbox (https://climatetoolbox.org/), and investigate their level of agreement in assessing drought conditions in the western continental United States. We compare two drought metrics and evaluate how they spatially evolve over the two decade period from 2000 to 2021. Our drought metrics are the USDM and Standardized Precipitation Index (SPI) at 30 day (SPI30d) and 180 day (SPI180d) intervals. We propose utilizing contiguous drought area (CDA) analysis to extract drought tracks from a larger network of drought characterizations. CDA blobs are organized through a directed graph that we can pull out individual paths called threads. Threads are aligned between different metric graphs through spatio-temporal intersection. USDM drought characterizations were found to move slowly and be more persistent than SPI characterizations. Meanwhile SPI30d drought characterizations were fragmented and moved quickly with dynamic changes in size. SPI180d drought characterizations moved faster than USDM drought characterizations, but not as dynamically as SPI30d drought characterizations. Contiguous areas between metrics were found to not necessarily result in the same drought evolution, yet we now have the means to compare evolutions. We found that the definition of drought used highly impacts the evolution observed.