Hydrological Uncertainty Across Snow‐ and Rainfall‐Dominated Sub‐Basins of a Partially Glacierized Himalayan Catchment

Abstract

Uncertainty in hydrological modelling is often addressed through parameter estimation, while uncertainties from model inputs and structure remain underexplored. This limitation is critical in mountainous catchments, where sparse monitoring and high spatial heterogeneity amplify uncertainty. This study assesses precipitation input, parameter and model structure uncertainties in the Alaknanda River Basin, a partially glacierized Himalayan catchment with snow‐dominated upstream and rainfall‐driven downstream sub‐basins. Precipitation input and parameter uncertainties were evaluated using the SWAT+ model through Latin Hypercube Sampling and the Sequential Uncertainty Fitting algorithm, while model structure uncertainty was assessed using 47 conceptual hydrological models implemented in the MARRMoT framework. The precipitation input assessment shows that differences among precipitation datasets substantially affect simulated streamflow, particularly in areas with high precipitation variability. Parameter sensitivity analysis indicates that snow‐related parameters exert strong control on simulated streamflows in the upper basin, with parameter uncertainty producing the widest prediction uncertainty band at the glacier‐fed Joshimath station ( r ‐factor 0.85). The model structure assessment shows substantial variation among conceptual model structures, with wider prediction bands at the downstream rainfall‐ and baseflow‐dominated sub‐basins ( r ‐factor 0.63 at Rudraprayag‐A). Structural analysis further indicates that snow storage, soil moisture and routing stores are important for reliable simulations, whereas additional storage components do not consistently improve performance. Overall, the three uncertainty assessments reveal distinct spatial patterns associated with the dominant hydrological processes across the basin. These findings highlight the need for process‐specific calibration, data collection and model selection to improve hydrological predictions in complex mountainous catchments.