Drought Monitoring in the Southwestern United States: Analysis of Seasonal Precipitation, Multiscalar Indices, and Soil Water

Drought is a complex, natural hazard that can cause widespread socioeconomic and environmental impacts. Drought can be defined as a water deficit that arises compared to normal conditions and lasts long enough to cause a lasting hydrological imbalance; however, what constitutes normal conditions varies based on the local climate regime and water source being studied. It is therefore that drought can be subjective to the observer and accounting for the correct environmental drivers is important to accurately assess drought impacts. In water-limited ecosystems like the Southwestern United States, hereby referred to as the Southwest, monitoring drought conditions presents unique challenges as annual potential evapotranspiration is significantly greater than precipitation. Primary production of Southwestern vegetation is adapted to the timing and magnitude of soil water recharge from seasonal precipitation. Variations in precipitation timing and magnitude can thus lead to lasting drought impacts and therefore tracking soil water at different depths is key for understanding overall ecosystem health. Studies have monitored soil water availability by placing Time Domain Reflectometry (TDR) probes at different depths within a soil profile. However, varying measurements between TDR probes due to differences in soil properties, calibration complications, and maintenance issues have led to the lack of long-term, reliable soil water datasets. This has restricted drought analysis using soil water data to a more local scale. As an alternative approach, land managers and decisions makers use meteorological drought indices as proxies for soil water availability. Meteorological drought indices are numerical timeseries that convey the frequency and magnitude of meteorological indicators, such as precipitation and temperature. Examples of meteorological drought indices are the Standardized Precipitation Index (SPI) and Standardized Precipitation-Evapotranspiration Index (SPEI), which use monthly total precipitation and water balance values, respectively, to communicate drought conditions. A key feature of the SPI and SPEI is the ability to be calculated at any monthly timescale length (‘multiscalar’), allowing for drought conditions of different water sources, such as shallow soil water, to be evaluated. Furthermore, the SPI and SPEI require minimal data inputs and are simple to calculate. It is therefore that the SPI and SPEI are commonly used by land managers for drought monitoring and decision making. Using SPI and SPEI timescales to estimate soil water in the semi-arid Southwest can be complicated as not all falling precipitation can be assumed to infiltrate into a soil profile. This is due to low intensity precipitation events being captured by tree canopy interception or evaporation and increased runoff during high intensity events. These issues complicate objectively identifying the index and timescale length that best represents soil water availability at different depths, leaving a significant gap between applying meteorological drought index information to land management action for the Southwest. Furthermore, the lack on in-situ soil water datasets prevents characterizing the full relationship between index timescale and soil water. Moreover, the lack of soil water datasets limit the study of links between Southwestern climate, intra-annual soil water variability, and soil drought dynamics on a regional scale. This dissertation address these gaps by creating a regional soil water dataset for the Southwestern United States for the purpose of improving drought monitoring and our understanding of drought development in soils. By coupling sophisticated computer modeling, site-specific soils information, and spatially continuous, high resolution meteorological datasets, this dissertation simulated daily matric potential values from 0-200cm at 240 locations across the Southwest. Through a series of investigative studies central to the Southwest, this dissertation quantified historical drought events in soils (appendix A), defined the relationship between meteorological drought index timescale and soil water depth (Appendix B), and developed a time-varying approach designed to improve multiscalar index approximation of soil water in climates with seasonal precipitation distributions (Appendix C). The first study (Appendix A) quantified how changes in the timing or magnitude of seasonal precipitation translated to soil drought onset and cessation patterns. Results showed that annual matric potential values followed a location’s seasonal precipitation distribution. Short-term droughts (60 – 270 days) were frequent, and typically resulted from delayed or slowed starts to a locations major rainy season. Long-term droughts (>270 days) were infrequent and occurred only during specific years, requiring anomalous below average precipitation in one or more consecutive rainy seasons to develop. Long-term droughts were more likely to occur in locations with unimodal precipitation distributions (a majority of rain occurring once annually), due to soil water anomalies likely remaining unresolved until the following rainy season. Locations with bimodal precipitation distributions (rainy seasons occurring twice annually) made long-term drought development difficult as consecutive below average rainy seasons were needed. The second study (Appendix B) defined the relationship between multiscalar index timescale and soil water availability for the Southwest. For all 240 locations, a new matric potential index (MPI) was created at 5cm intervals between 0-200cm and correlated with timescales from 1-24 months for the SPI and SPEI. Results showed the relationship between the highest correlating index timescale at each MPI depth operates roughly on a 1:1 step progression at shallow depths. Further analysis showed that soil type impacts the timescale-depth relationship, with clay loam soils correlating at longer timescales than sandy soils when correlating with the same depth MPI. Additionally, the SPI produced higher correlations and with the MPI compared to the SPEI. Therefore, this studied recommended SPI usage for shallow (