Skip to main content
ARS Home » Pacific West Area » Riverside, California » Agricultural Water Efficiency and Salinity Research Unit » Research » Publications at this Location » Publication #414991

Research Project: Water Management for Crop Production in Arid and Semi-Arid Regions and the Safe Use of Alternative Water Resources

Location: Agricultural Water Efficiency and Salinity Research Unit

Title: Intercomparison of citrus evapotranspiration among eddy covariance, OpenET models, and water and energy balance model BAITSSS

Author
item Dhungel, Ramesh
item Anderson, Raymond
item French, Andrew
item Skaggs, Todd
item AJAMI, HOORI - University Of California, Riverside
item Wang, Dong

Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/13/2024
Publication Date: 11/1/2024
Citation: Dhungel, R., Anderson, R.G., French, A.N., Skaggs, T.H., Ajami, H., Wang, D. 2024. Intercomparison of citrus evapotranspiration among eddy covariance, OpenET models, and water and energy balance model BAITSSS. Agricultural Water Management. 304. https://doi.org/10.1016/j.agwat.2024.109066.
DOI: https://doi.org/10.1016/j.agwat.2024.109066

Interpretive Summary: Satellite remote sensing using thermal images have long been used for estimating crop evapotranspiration (ET). The OpenET project, recently formulated as a single platform of thermal-based remote sensing ET models, aims to provide field-scale estimates of ET across large landscapes and for many crop types. However, information is lacking for citrus. ET estimation for citrus is challenging due to its clumped canopy structure and evergreen leaves. In this study, we evaluated OpenET models and an independent water and energy balance model (BAITSSS) against in-situ eddy covariance ET measurements in two commercial citrus fields in San Joaquin Valley, California during 2021-2022. Results showed that ET estimates from the OpenET models were on average about 30% larger than measured ET. Simulated ET from BAITSSS was within 2 % of measured ET when actual irrigation and precipitation data were used in model inputs. The ensemble OpenET model diverged from measured ET mainly in the spring. This study highlighted the challenges in ET estimation using thermal-based remote sensing ET models for citrus.

Technical Abstract: Remote sensing-based surface energy balance algorithms have been used to estimate crop water use of various crops. However, citrus evapotranspiration (ET) estimation is challenging mainly due to evergreen leaves and a clumped canopy structure. In this study, we evaluated two methods for calculating ET: the OpenET models, which are mostly satellite thermal-based models, and the BAITSSS water and energy balance model. Calculated ET was compared with (i.) eddy covariance (EC) ET measurements for two San Joaquin Valley, California citrus orchards and (ii.) water received (irrigation plus precipitation) data for the two citrus sites. Polaris-based soil hydraulic properties and measured volumetric water content were used for BAITSSS initialization and water balance assessment. Sentinel-2 based NDVI was used for BAITSSS. Results showed that annual ET based on the OpenET ensemble model was on average 30% larger (r2 ~ 0.71, RMSE ~ 1.16 mm) than both EC ET and water received. The disparity was due mostly to disagreements that occurred in the spring. BAITSSS, on the other hand, showed mixed results when compared to EC ET and water received, with the annual total being on average 2% larger (r2 ~ 0.77, RMSE ~ 0.94 mm). Both measured and modeled ET values were below grass reference ET (ETo) for the majority of the simulation period. Soil moisture and water received data indicated the orchards may have been deficit irrigated. Overall, this study highlights the challenges of ET modeling in citrus orchards and the improvement needed for better estimation of ET for this specialty crop.