On Conceptualization and Routing Dynamics of a Glacierized Catchment

English Abstract:

Assuming that glaciers can be represented as linear reservoirs that transform surface melt into runoff, we investigate their dynamic behaviour by rewriting the differential equation of linear reservoirs to derive the storage coefficients (Ks) from input and outputs to the reservoir system. We use hourly ablation rates simulated by a distributed energy balance model (input) and hourly runoff at the glacier snout (outputs) and study the system characteristics derived from the linear proportionality between the time derivative of runoff and the difference between input and output. We derive the storage coefficients from the relationship between measured runoff and known inputs to the system and investigate their seasonal and interannual variability. We compare the resulted Ks with those obtained from time series analysis. We use data from three ablation seasons with different characteristics (extensive snow cover throughout the ablation season in 2001, limited/absent snow cover in 2006 and a ‘middle case’ in 2010) from Haut Glacier d’Arolla, Switzerland. Storage coefficients exhibit a similar general pattern of decline over time with local peaks corresponding to snowfall events with higher absolute values for the season with extensive snowcover and lower for the season with limited snowcover. Given the observed temporal evolution in storage characteristics, we modify the conventional reservoir approach in two ways. In the first type of models, we explore multiple reservoirs (from 1 to 5) with constant in time constant storage characteristics, while in the second approach we use one reservoir with varying in time storage coefficients, parameterised as a function of snow covered fraction or of the snowpack water equivalent (SWE). In the first category, models that separate the contribution of snowmelt from the glacierised area from that over the non-glacierised areas perform better, especially when applied to season with limited snowcover. In the second category, models based on SWE of the snowpack have a higher performance.