Exploring GEO–LEO fusion to improve the joint estimation of crop evapotranspiration and root-zone soil moisture in an ecohydrological model

Accurate estimation of crop evapotranspiration (ET) and root-zone soil moisture (RZSM) at agriculturally relevant spatial scales remains a challenge in ecohydrological modelling 1,2 . In a typical 30 m resolution surface energy balance (SEB) workflow, ET constraints usually rely on snapshots obtained from high-resolution Low-Earth-Orbit (LEO) satellites (e.g., Landsat), which are postprocessed using temporal interpolation. Consequently, day-to-day ET variability is driven primarily by reference ET (ETo), while the thermal constraint contributes only intermittently, which can dampen or mis-time actual ET dynamics and propagate into soil-profile water balance estimates. Geostationary (GEO) satellites (e.g., Himawari) provide a complementary pathway toward temporally dense monitoring of ET dynamics 3 , yet their integration into operational 30 m SEB workflows remains comparatively underexplored. Here, we present the spatiotemporal fusion 4,5 of GEO–LEO (i.e., Himawari–Landsat) data to improve the joint estimation of 30 m resolution ET and RZSM in an ecohydrological model, named the Sydney Soil Water–Energy Balance model (SWEB) model 6 . Himawari’s temporally dense observations provide an advantage of matching to Landsat overpass timing with minimal time offset (typically <10 minutes), offering a pathway to reduce temporal aliasing in ET estimation while retaining Landsat’s field-scale spatial details. By evaluating against ground-based networks, we showed that the spatiotemporally fused ET estimates exhibited improved agreement with flux tower observations, relative to ET estimates derived from temporally interpolated Landsat snapshots. Furthermore, when the fused ET data was embedded as the ET constraint in the subsequent soil water balance calibration, the optimised hydraulic parameter set became more physically plausible and better constrained, leading to systematic improvements in RZSM estimates. Overall, integrating GEO–LEO fusion into SWEB provides a robust pathway to better capture ET variability and its coupling with RZSM, enabling more reliable assessment of agricultural water availability.