Quantifying Wildfire Impacts on Water Yield via Multi‐Scale <scp>BACI</scp> and Remotely Sensed Evapotranspiration

ABSTRACT Wildfires disrupt vegetation structure, soil properties and hydrologic processes that regulate watershed water yield. Quantifying wildfire impact on water‐yield remains difficult because conventional Before–After–Control–Impact (BACI) analyses rely on in situ discharge records and a single external control watershed, which are rarely available for sufficiently long and comparable periods. To address these limitations, we develop a multi‐scale, multi‐control BACI framework that integrates remote sensing‐based actual evapotranspiration (AET) from the OpenET ensemble to infer post‐wildfire changes in annual water yield without requiring dense gauge networks or external control watershed. Unlike traditional paired designs, our method allows for performing the BACI analysis within single watershed. The watershed is first divided into gridded 4 km × 4 km tiles, ensuring the tile size is smaller than the burned area perimeters. Unburned or vegetation‐stable tiles are used as potential controls; for each burned tile, the model uses a weighted ensemble of all control tiles, to construct a composite BACI model for the watershed. This structure reduces spatial bias and quantifies the strength and consistency of post‐fire hydrological responses. The framework was applied to the Middle Fork Salmon River watershed (Idaho, USA), which experienced three major wildfires (2003, 2006, 2007) and the analysis revealed consistent hydrological outcomes: streamflow increased by 36 mm year −1 (traditional BACI), while AET decreased by 31 mm year −1 , consistent with watershed water balance (Δ Q WY ≈ −ΔAET). The composite model reproduced these patterns with lower total uncertainty ( ≈ 0.4 mm/year), while resolving spatial variability linked to burn severity and extent. This research introduces a scalable BACI methodology that couples physically hydrologic reasoning with satellite‐derived flux estimates. The framework enhances the reproducibility and transferability of post‐wildfire water yield assessments, providing a generalizable tool for ungauged basins and a new pathway to integrate remote sensing and hydrologic modelling in disturbance hydrology.