Persistent daily OpenET bias drives seasonal water-use divergence in California avocado orchards

Satellite-based evapotranspiration (ET) products such as OpenET are increasingly used for irrigation management and agricultural water accounting, yet their reliability in heterogeneous avocado orchard systems remains insufficiently characterized across operational timescales. This study evaluated OpenET against eddy covariance (EC) measurements across four commercial avocado orchards in California during 2022–2025 to determine whether recurring daily ET offsets accumulated into considerable cumulative seasonal water-use divergence. Across sites and years, the OpenET ensemble reproduced broad temporal ET variability while consistently generating higher ET estimates than ET₍EC₎. Regression slopes generally exceeded unity (1.08–1.38), with ET divergence increasing consistently under high atmospheric-demand conditions. Despite relatively modest daily RMSE values (0.59–1.19 mm d⁻¹), cumulative seasonal divergence between ET₍OpenET₎ and ET₍EC₎ reached 23–282 mm, corresponding to relative seasonal divergence of 2.8–33.9%. Approximately 60–75% of cumulative seasonal divergence developed during the summer high-demand period. Mixed-effects analyses further showed that daily ET bias increased progressively with ET magnitude and atmospheric demand, with the strongest positive divergence occurring during the mid-season high-demand period. Sensitivity analyses indicated that cumulative seasonal divergence remained evident under conservative EC uncertainty assumptions across most site-year combinations. Although model responses varied among OpenET components, the overall tendency toward ET amplification remained broadly consistent across orchard systems. These findings indicate that apparently acceptable daily ET agreement can mask substantial cumulative seasonal water-use divergence in avocado production systems, highlighting the importance of evaluating satellite-based ET products across operational seasonal timescales rather than relying solely on daily performance metrics.