Tracing snow-hydrological processes in a pre-alpine catchment with stable water isotopes

Abstract

In winter, the snowpacks in mountain headwater catchments store significant amounts of water, which is released during the melting period. As snowfall and snow accumulation vary significantly with elevation and forest cover, the release of meltwater is highly heterogeneous across the catchment landscape, particularly during rain-on-snow events. As a consequence, streamflow generating mechanisms during rain-on-snow events and the snowmelt period are complex in mountainous headwater catchments. In order to improve predictions of water availability in the future, these processes need to be better understood. In this PhD thesis, the hydrological processes during snow accumulation, snow melt and rain-on-snow events were investigated in the 47 km2 Alp catchment in Central Switzerland. For this, a snowmelt lysimeter system was designed to obtain high-resolution measurements of snowpack outflow (defined as snowmelt and rain) and its chemical composition at the plot scale (Part I). These data were complemented by regular measurements of stream water, precipitation and bulk snowpack properties, so that the resulting database allows for studying the variable hydrological responses of different water sources (Part II). In particular, the spatiotemporal variability of snowpack outflow generation across the catchment landscape was investigated and the snowmelt contribution to streamflow during rain-on-snow events was estimated (Part III). Part I of this thesis compares two different systems used to capture the rate of snowmelt and its isotopic composition at daily intervals during two snowfall-snowmelt seasons in mountainous terrain. The systems were a low-cost unheated precipitation collector and a newly developed snowmelt lysimeter system. The analysis showed that the isotopic compositions of the precipitation samples derived from the unheated precipitation collector were similar to those derived from the snowpack lysimeter system. Thus the sampling method based on the unheated precipitation collector could be a useful alternative for estimating the isotopic composition of snowmelt, especially for remote areas, as it requires little effort to install. The snowmelt lysimeter system requires a power supply and significantly more effort in installation; however, it allows for monitoring snowmelt rates at 10-minute resolution at multiple nearby locations. Part II aims at providing and evaluating a database containing environmental tracers such as geochemical solutes and stable water isotopes in a spatially distributed sampling network in the research Alptal catchment. During a 3.3-year period, different water sources such as precipitation, stream water and snowpack outflow were sampled at daily intervals. To reduce evaporative fractionation effects in the automatic water samplers during each three-week sampling period, the sample bottles were retrofitted. This new design significantly reduced isotopic fractionation in the collected water samples, especially during warm conditions in summer. The environmental tracer data in stream water showed good agreement with bi-weekly measurements carried out by the Swiss National River Monitoring and Survey Programme (NADUF). The database is separated into two parts and contains water fluxes and environmental tracer data at daily intervals in (1) precipitation and stream water and (2) snowpack outflow. Database (1) also includes measurements of snowpack outflow volumes at ten-minute resolution. Both datasets are publically available on www.envidat.ch (10.16904/envidat.70 and 10.16904/envidat.71). In Part III, the generation of snowmelt and streamflow are investigated for several rain-on-snow events in a 0.7 km2 tributary of the Alptal valley. Snowpack outflow volumes were measured with snowmelt lysimeter systems at three different elevations characterized by different vegetation (e.g., open grassland and evergreen forest). The differences in snowpack outflow volumes at the three sites could mainly be explained by initial snow depth and rainfall volumes, which in turn are affected by vegetation cover and elevation. An end-member mixing analysis based on stable water isotopes showed that snowpack outflow contributions to streamflow during rain-on-snow events were highly variable between the three lysimeter sites. Our results suggest that hydrological processes during rain-on-snow events are very heterogeneous across the catchment landscape mainly also due to effects of vegetation and elevation.