Hydrological Dynamics of Debris Flow-Prone Pyroclastic Catchments: A Budyko-Based Modeling Approach

Rainfall-induced geohazards in pyroclastic slopes are known to be influenced by antecedent soil wetness, but the assessment of this condition based on the catchment water balance is still unexplored. A lumped Dynamic Water Balance Model (DWBM) based on the Budyko framework is developed to characterize the hydrological behavior of eight landslide-prone catchments underlain by pyroclastic deposits over limestone bedrock in Campania, Italy. Using monthly rain-gauge, satellite, and reanalysis data between 2003 and 2023, model parameters representing catchment retention, evapotranspiration efficiency, maximum soil storage, baseflow response, and groundwater losses were calibrated against the discharge, estimated from reanalysis runoff, for seven ungauged catchments and from a stage–discharge relationship for one gauged catchment. This allowed an assessment of the consistency of reanalysis runoff for setting up hydrological models of ungauged catchments. During calibration, the models reproduced discharge satisfactorily (RMSE = 11.66–25.95 mm; NNSE = 0.80–0.90; KGE = 0.76–0.91), although the performance decreased in some catchments in model validation (RMSE = 13.60–28.97 mm; NNSE = 0.61–0.92; KGE = 0.55–0.91). Monte Carlo and Sobol analyses revealed distinct parameter sensitivities, with evapotranspiration efficiency, catchment retention, and maximum soil storage contributing directly to model output variability, while groundwater-related parameters were primarily influential through interactions with other parameters. Physically interpretable relationships of model parameters with catchment characteristics were found, allowing a hydrological characterization of poorly gauged pyroclastic catchments. Modeled soil storage provides a physically meaningful proxy for antecedent slope wetness, potentially supporting the assessment of rainfall-induced geohazards.