The selection of paired watersheds affects the assessment of wildfire hydrological impacts

Wildfires strongly alter hydrological processes and surface and groundwater quality in forested environments. The paired-watershed method, consisting of comparing a burnt (altered) watershed with an unburnt (control) watershed, is commonly adopted in studies addressing the hydrological effects of wildfires. This approach requires a calibration period to assess the pre-perturbation differences and relationships between the control and the altered watershed. Unfortunately, in many studies, the calibration phase is lacking due to the unpredictability of wildfires and the large number of processes that should be investigated. So far, no information is available on the possible bias induced by the lack of the calibration period in the paired-watershed method when assessing the hydrological impacts of wildfires. Through a literature review, the consequences of the lack of calibration on the assessment of wildfire hydrological changes were evaluated, along with the most used watershed pairing strategies. The literature analysis showed that if calibration is lacking, misestimation of wildfire impacts is likely, particularly when addressing low-severity or long-term wildfire effects. The Euclidean distance based on physical descriptors (geology, morphology, vegetation) was proposed as a metric of watersheds similarity and tested in mountain watersheds in Central Italy. The Euclidean distance proved to be an effective metric for selecting the most similar watershed pairs. This work raises awareness of biases exerted by lacking calibration in paired-watershed studies and proposes a rigorous and objective methodology for future studies on the hydrological effects of wildfires.