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 #405754
Research Project: Towards Resilient Agricultural Systems to Enhance Water Availability, Quality, and Other Ecosystem Services under Changing Climate and Land Use

Location: Agroclimate and Hydraulics Research Unit

Title: Sensitivity analysis of standardized precipitation and evapotranspiration index (SPEI) to probability distributions and potential evapotranspiration methods

Author
item LEE, SANGHYUN - US Department Of Agriculture (USDA)
item Moriasi, Daniel
item DANANDEHMEHR, ALI - Antalya Bilim University
item MIRCHI, ALI - Oklahoma State University

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/27/2024
Publication Date: 3/29/2024
Citation: Lee, S., Moriasi, D.N., Danandehmehr, A., Mirchi, A. 2024. Sensitivity analysis of standardized precipitation and evapotranspiration index (SPEI) to probability distributions and potential evapotranspiration methods. Journal of Hydrology: Regional Studies.53(2024).101761. https://doi.org/10.1016/j.ejrh.2024.101761.
DOI: https://doi.org/10.1016/j.ejrh.2024.101761

Interpretive Summary: Drought has harmful effects on water resources and affects all aspects of life. Climate change is expected to worsen these effects, as longer and more severe droughts are projected to occur more frequent. The standardized precipitation evapotranspiration index (SPEI) is widely used to monitor drought conditions to help decision-makers better prepare and address drought impacts. This index can be used to monitor droughts on a short (1 - 6 months)- or long (12 - 24 months)- term basis. The drought prediction accuracy of this index depends on the evapotranspiration (ET) method it uses. In this study, three commonly used ET methods on drought prediction accuracy were evaluated. Two methods require only temperature input, and the third requires more climate data inputs that are not readily available in many regions. Climate data from 107 weather stations across Oklahoma were used. The method using more data inputs produced the most reliable predictions. However, the methods that require only temperature input can also be used depending on the purpose of the study. One of the methods using only temperature input showed good results when the droughts were monitored on a short-term basis, and both worked well for long-term drought monitoring in regions with limited available data inputs. These findings will help researchers choose an appropriate ET method to use depending on data availability and study purpose. USDA is an equal opportunity provider and employer.

Technical Abstract: The standardized precipitation evapotranspiration index (SPEI) is a widely used meteorological drought index that incorporates potential evapotranspiration (PET) into precipitation-based indices. However, the choice of PET equation for SPEI is often not given much consideration despite its importance. In this study, we evaluated the impacts of different PET equations, namely Thornthwaite (TW), Hargreaves (HG), and Penman-Monteith (PM), on drought assessments, based on SPEI. Both TW and HG require mainly temperature as input parameter, whereas PM requires extensive meteorological datasets. We considered three temporal scales: 1) complete time series, 2) event-based, and 3) monthly to evaluate SPEI at various accumulation periods. In addition, we considered the log-logistic and generalized extreme value distributions to test the normality of SPEI computed from the three PET methods. To do this, we utilized high-quality climate datasets measured at 107 stations across Oklahoma State, which has a diverse climate ranging from semi-arid to humid subtropical. Based on the Shapiro-Wilk test results, we selected the log-logistic distribution for SPEI. For complete time series drought analysis, we compared the SPEI series based on Pearson’s r and mean absolute difference (MAD). For event-based, drought characteristics such as frequency, duration, severity, and intensity were calculated to compare. Lastly, on a monthly scale, the occurrences of drought months for moderate, severe, and extreme droughts were extracted using drought classifications. For the complete time series comparisons, SPEI-HG showed better agreement with SPEI-PM than SPEI-TW. For event-based and monthly drought analyses, HG and TW equations can be used as a suitable alternative to PM, respectively, for accumulation periods less than one year. However, for accumulations longer than one year, both TW and HG equations showed no significant differences with SPEI-PM. The findings provide practical guidance for selecting an appropriate PET equation depending on the purpose of study without resorting to data-intensive methods for PET estimation. USDA is an equal opportunity provider and employer.