Towards a better understanding of river network dynamics in a glacierized catchment

• We simulated 19 years of river network dynamics in a glacierized alpine catchment. • Quantified river network expansion and contraction using flow-based indicators. • Maximum contraction during winter and early spring, and expansion during summer. • Flow intermittency was most prevalent in low-order headwater streams. • Market interannual variability, especially during the contraction of the river network. River networks are increasingly recognized as highly dynamic components of catchments, expanding and contracting in response to hydrometeorological drivers such as precipitation, snow and glacier melt, and groundwater inputs. These spatial and temporal patterns shape not only hydrology but also key ecological and biogeochemical processes, including habitat connectivity, species interactions, nutrient cycling, and sediment transport. Accurately quantifying network expansion and contraction is therefore essential for understanding ecosystem functioning at the catchment scale. However, these dynamics remain poorly understood in high-mountain glacierized systems, largely due to limited observations and the complexity of cryosphere-hydrology interactions. In this study, we applied the recently developed SWAT-GL model to simulate river network dynamics in the Valsorey catchment (Swiss Alps) over the past 19 years. The model was calibrated and validated using observed discharge data (2016–2022) and then used to derive a set of flow-based indicators quantifying expansion and contraction patterns. Our results revealed strong seasonality, with up to 50% of the river network contracting during winter and early spring, and full connectivity during the summer melt season. Flow intermittency was most prevalent in low-order headwater streams, which exhibited on average 15% higher probability of intermittency than high-order streams, typically experiencing 2 to 3 non-flow events per year with an average duration of ∼ 30 days. Comparisons between wet and dry years highlighted significant interannual variability: wet years showed higher discharge (annual cumulative discharge 40 × 10 6 m 3 ), smaller dry surface areas (∼0.2 km 2 ), and shorter contraction phases, whereas dry years exhibited ∼ 40% lower cumulative discharge, nearly double dry surface area (∼0.40 km 2 ), and prolonged contraction periods. SWAT-GL effectively captured both hydrological and cryospheric processes, revealing strong seasonal and interannual contrasts that underscore the sensitivity of alpine river networks to climatic variability.