Memory-based drought decomposition for regionalized drought assessment droughts exhibit complex patterns of variability that challenge the application of traditional drought indices, which often…

The National Oceanic and Atmospheric Administration awarded a Cooperative Agreement totaling $499,936 to Mississippi State University on July 1, 2026, under the Climate and Atmospheric Research program (CFDA 11.431) to develop memory-based drought decomposition methods for regionalized drought assessment across the Southeastern United States.

The project addresses the limitation that traditional drought indices—Standardized Precipitation Index (SPI), Standardized Precipitation Evaporation Index (SPEI), and Standardized Streamflow Index (SSI)—assume stationarity and may misrepresent drought risk in regions where climate characteristics are non-stationary. Mississippi State will apply time series decomposition using the STL method and a novel harmonic framework to quantify long- and short-term behaviors specific to regional conditions, assessing both meteorological and hydrological droughts across varying timesteps in the Southeastern US. The research will test non-stationary indices against local conditions to enhance drought index performance, with the intent to upscale findings across the broader United States. Cross-correlation analysis will determine the level and timing of drought development patterns.

Work is performed at Mississippi State University, Starkville, Mississippi, with a period of performance from July 1, 2026, through June 30, 2028.

Memory-based drought decomposition for regionalized drought assessment droughts exhibit complex patterns of variability that challenge the application of traditional drought indices, which often assume stationarity, defined as no change in mean, variance, or shape of the time series. Indices like the standardized precipitation index (SPI), standardized precipitation evaporation index (spei), and standardized streamflow index (SSI) quantify drought, but may underestimate risk in some regions due to stationarity assumptions. Non-stationary indices offer promise but are largely untested across diverse environments. This proposal aims to improve drought parametrization by addressing regional non-stationarity across the southeastern US (seus). We hypothesize that region-specific drought assessments better capture drought processes and quantify ongoing changes. Generalized indices may misrepresent diverse environments, leading to under- or overestimations of dry conditions. Literature and preliminary findings indicate that not all regions in the seus are equally impacted by non- stationarity. By adjusting indices to reflect local conditions and verifying non-stationarity assumptions, this research will enhance the performance of drought indices across initially in the seus and on further stages across the US using upscaling techniques. Meteorological and hydrological droughts will be studied in this project in varying timesteps. Hydrologic drought will be delineated by means of an objective threshold, and additionally, SSI will be computed. Meteorological droughts will be assessed by means of SPI and spei. To determine the potential scale of non-stationarity, time series will be decomposed using the STL method, and long- and short-term behaviors will be quantified by means of a novel harmonic framework with supporting tests for specific non-stationarity types. Cross-correlation analysis will be performed to determine the level and time shifts in the development of droughts. Non-stationary modeling for SSI, SPI, and spei will be developed using the gamlss framework, and the akaike information criterion will be applied to guide the selection process. Results will constitute the basis for regionalization using fuzzy clustering, with regional parameters and assessment methods provided for each region. The results will be disseminated through an online atlas and data hub. This project represents a novel approach to drought assessment through its integration of non-stationary modeling and drought regionalization across the seus based on drought characteristics and physical drivers rather than geographic region similarities. The interactive application will be used to deliver project outcomes and will be further adopted by the bureau of indian affairs and united south and eastern tribes to support drought resilience efforts on tribal lands, as well as the US golf association, to help guide their strategic decision-making for their program on reduction of water use on golf courses. Additionally, thanks to collaboration with local agricultural organizations and promotion boards the tool will be popularized in the sector to help agriculture managers improve their preparedness for future drought events. Thanks to our collaboration with the science-policy interface of the united nations convention to combat desertification, project outcomes will help improve policymaking on a national and global scale. The proposed project is highly relevant to the competition as IT is designed to address key problems identified by NOAA, as well as climatic research goals, overarching goals of nidis, and focus areas included in the NOAA drought assessment memorandum. This will be accomplished through improving the performance of drought indices on regional scales, providing detailed characteristics and tailored recommendations for each identified region, incorporating non-stationarity characteristics in the assessment of droughts, publicly accessible projec