Multi-depth actively heated fiber optics sensing quantifies spatial-support uncertainty in soil-water-balance evapotranspiration estimates under greenhouse subsurface drip irrigation

Estimating actual crop evapotranspiration (ET c act ) in greenhouse vegetable systems under subsurface drip irrigation is complicated by strong small-scale heterogeneity in root-zone soil water dynamics and by mismatch between local measurements and plot-scale water balance calculations. We evaluated actively heated fiber optics (AHFO) sensing as a distributed approach for monitoring soil water content at multiple depths (0.10–0.50 m) along 8.0 m greenhouse plots planted with Omanda No. 3 tomato, Strawberry 9051 tomato, and cucumber. AHFO-derived soil water content agreed well with co-located time-domain reflectometry measurements (R² = 0.91–0.97), with the smallest errors in the main root zone (RMSE = 0.018–0.028 m³ m⁻³ at 0.10–0.30 m). The distributed records showed crop-dependent depth patterns and strong horizontal variability near the buried drip line. Plot-integrated ET c act increased with irrigation amount during the AHFO monitoring period, but local-support estimates remained widely dispersed even when their treatment means were close to plot-integrated values. Layer-wise storage analysis showed that depletion and replenishment were uneven through the 0–0.60 m profile. Increasing AHFO support length from 0.25 to 4 m reduced the worst-case 95th-percentile ET c act error from 24.9% to 7.5% and the mean probability of exceeding 10% error from 21.9% to 0.6%. These results show that AHFO can quantify spatial-support effects and reduce representativeness uncertainty in soil-water-balance ET c act estimation under heterogeneous greenhouse subsurface drip irrigation.