Precipitation forecasting utility for proactive agroecosystem management: A case study from central Texas

Authors: Schantz, Merilynn; Adhikari, Kabindra; Harmel, Robert; Hardegree, Stuart; ABATZOGLOU, JOHN - University Of California; FULHART, ANDREW - University Of Arizona; Thorp, Kelly; Smith, Douglas

Submitted to: Journal of Environmental Quality
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/21/2025
Publication Date: N/A
Citation: N/A

Interpretive Summary: Precipitation inputs directly affect plant production, sediment and nutrient loading, and soil health across agroecosystems depending on precipitation abundance, timing and intensity. Gridded weather and climate data have been used for nearly 20 years to assist in understanding precipitation timing and intensity across small spatial scales (1-4 km2) and are critical to the development and use of many agroecosystem decision support tools and models. There are, however, numerous gridded data products available and their quality is rarely evaluated since data from most ground-based monitoring stations necessary for testing are often used in the production of gridded data. At the Texas Gulf Long-Term Agroecosystem Research (LTAR) site in Riesel, TX we found that all gridded data were significantly related to on-site weather station precipitation data. We also found that gridded data acquired from earlier decades (1982-1989) had greater error compared to those from recent decades (2000-2024), likely due to upgraded instrumentation at ground-based monitoring sites used to produce gridded data. Moreover, we identified that all historical data sources accurately accounted for ENSO events. Collectively this study provides a critical quality assurance test and a template for quality assurance tests of gridded precipitation data across LTAR sites.

Technical Abstract: Precipitation inputs directly affect plant production, sediment and nutrient loading, and soil health across agroecosystems depending on precipitation abundance, timing and intensity. Gridded weather and climate data have been used for nearly 20 years to assist in understanding precipitation timing and intensity across small spatial scales (1-4 km2) and are critical to the development and use of many agroecosystem decision support tools and models. There are, however, numerous gridded data products available and their quality is rarely evaluated since data from most ground-based monitoring stations necessary for testing are often used in the production of gridded data. At the Texas Gulf Long-Term Agroecosystem Research (LTAR) site in Riesel, TX, however, a nearly 90-year precipitation record exists that provides a key region for quality assurance tests of gridded precipitation data as this region has extremely dynamic precipitation associated with frequent convective storm activity. Our objective in this study was to quantify the differences between historical precipitation data acquired from an on-site precipitation monitoring station to the common gridded databases of PRISM, DayMET and GridMET at the Texas Gulf LTAR site in Riesel, TX. We hypothesized that differences among gridded data sources would 1. Vary by grid scale where the smaller the grid scale the greater the accuracy and 2. Vary by time where gridded precipitation data calculated for times prior to 1990 would be less accurate than gridded data calculated for recently acquired data and 3. That all historical data sources would accurately account for ENSO events. Our findings suggest that all gridded data were significantly related to on-site weather station precipitation data. Contrary to our first hypothesis, the grid scale did not affect data accuracy. In support of our second hypothesis, however, gridded data acquired from earlier decades (1982-1989) had greater error compared to those from recent decades (2000-2024), likely due to upgraded instrumentation at ground-based monitoring sites used to produce gridded data. Lastly, we found that all historical data sources accurately accounted for ENSO events. Collectively this study provides a critical quality assurance test and a template for quality assurance tests of gridded precipitation data across LTAR sites.