Skip to main content
ARS Home » Plains Area » El Reno, Oklahoma » Oklahoma and Central Plains Agricultural Research Center » Agroclimate and Hydraulics Research Unit » Research » Publications at this Location » Publication #407808
Research Project: Impacts of Variable Land Management and Climate on Water and Soil Resources

Location: Agroclimate and Hydraulics Research Unit

Title: Standardized precipitation and evapotranspiration index (spei) sensitivity analysis: influence of probability distributions and evapotranspiration estimation methods

Author
item LEE, SANGHYUN - Orise Fellow
item Moriasi, Daniel
item DANANDEH MEHR, ALI - Antalya Bilim University
item MIRCHI, ALI - Oklahoma State University

Submitted to: American Geophysical Union Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 8/23/2023
Publication Date: 12/11/2023
Citation: Lee, S., Moriasi, D.N., Danandeh Mehr, A., Mirchi, A. 2023. Standardized precipitation and evapotranspiration index (spei) sensitivity analysis: influence of probability distributions and evapotranspiration estimation methods. American Geophysical Union Meeting Abstract. American Geophysical Union (AGU) Fall Meeting 2023, Dec 11-15, 2023, San Francision, CA.

Interpretive Summary:

Technical Abstract: The standardized precipitation evapotranspiration index (SPEI) is a widely used meteorological drought index, which incorporates the effect of potential evapotranspiration (PET). However, the sensitivity of SPEI to the choice of probability distributions and PET equations is often overlooked. The sensitivity of SPEI to these alternative selections were assessed. For PET methods, Thornthwaite (TW), Hargreaves (HG), and Penman-Monteith (PM) equations were considered. The log-logistic and generalized extreme value distributions were examined to select the most suitable probability distribution to model the water deficit (P-PET) series. High-quality climate data measured at 107 stations in Oklahoma Mesonet were used. Results revealed that the log-logistic distribution was suitable for SPEI based on the Shapiro-Wilk test. In terms of the selection of PET method, the results of SPEI-TW and SPEI-HG were compared with those of SPEI-PM considering various time scales for SPEI. SPEI results computed from different PET methods at three aspects: 1) complete time series of SPEI, 2) extracted drought events, and 3) number of occurrences of droughts based on run theory were compared. For complete time series analysis, the SPEI series based on Pearson’s correlation coefficient (r) and mean absolute difference (MAD) were compared. For event-based, drought frequency, duration, severity, and intensity were evaluated against one another. Lastly, the occurrences of droughts for moderate, severe, and extreme droughts were characterized using drought classifications. The results showed that in general, SPEI-HG can be used as a suitable alternative to SPEI-PM for shorter timescales. However, for timescales longer than one year, both SPEI-TW and SPEI-HG methods showed no significant differences with SPEI-PM. These findings are expected to provide practical guidance to characterize drought using SPEI depending on the purpose of study and data availability for PET estimation. “USDA is an equal opportunity provider and employer.”