Incorporating field wind data to improve crop evapotranspiration parameterization in heterogeneous regions

Accurate parameterization of reference evapotranspiration (ET) is necessary for optimizing irrigation scheduling and avoiding costs associated with over-irrigation (water expense, loss of water productivity, energy costs, and pollution) or with under-irrigation (crop stress and suboptimal yields or quality). ET is often estimated using the FAO-56 method with meteorological data gathered over a reference surface, usually short grass. However, the density of suitable ET stations is often low relative to the microclimatic variability of many arid and semi-arid regions, leading to a potentially inaccurate ET for irrigation scheduling. In this study, we investigated multiple ET products from six meteorological stations, a satellite ET product, and integration (merger) of two stations’ data in Southern California, USA. We evaluated ET against lysimetric ET observations from two lysimeter systems (weighing and volumetric) and two crops (wine grapes and Jerusalem artichoke) by calculating crop ET (ET) using crop coefficients for the lysimetric crops with the different ET. ET calculated with ET products that incorporated field-specific wind speed had closer agreement with lysimetric ET, with RMSE reduced by 36 and 45% for grape and Jerusalem artichoke, respectively, with on-field anemometer data compared to wind data from the nearest station. The results indicate the potential importance of on-site meteorological sensors for ET parameterization; particularly where microclimates are highly variable and/or irrigation water is expensive or scarce.