Aligning Evapotranspiration from MOD16A2.061 Product to Ground Estimates in Piemonte (NW Italy): an analysis of temporal and spatial biases

Abstract. Evapotranspiration (ET), and in particular Reference Evapotranspiration (ET₀), is essential for agricultural planning, irrigation management, and water resource allocation—especially in regions facing water scarcity and limited observational data. While ground-based ET₀ is typically estimated using the FAO Penman-Monteith method, satellite-derived products such as MOD16 offer broader spatial coverage, although with conceptual and methodological differences. MOD16 provides Potential Evapotranspiration (PET) estimates, which, unlike ET₀, depend on local biome characteristics and are not standardized to a reference surface.
This study investigates whether the PET data from the MOD16A2GF product (version 6.1) can be adapted for ET₀ estimation in Piemonte (NW Italy), a region characterized by diverse climates and topographies. We compared PET data from 2010 to 2022 with ground-based ET₀ and applied a bias-correction method using linear regression models calibrated on local meteorological time series. The corrected dataset (ÊT₀) shows significantly improved agreement with ground-based ET₀, reducing the Mean Absolute Error from 10.06 mm/8 d to 2.48 mm/8 d, a 75% improvement. This correction proved robust across the region and particularly effective during the summer, when accurate ET₀ estimation is critical for crop irrigation.
Our results suggest that, with appropriate local calibration, MOD16A2GF PET data can serve as a practical surrogate for ET₀ in data-scarce environments. Future research should focus on exploring the impact of additional factors, such as altitude and land cover variability, to further refine the accuracy of satellite-derived ET₀ estimates and improve their applicability in diverse climatic and topographical conditions.