Micro Temperature Loggers: A Cost Effective Technology to Derive Input Data for Distributed Snow Melt Models

In snow dominated forested watersheds the quantification of snow melt is the most important process in hydrological modelling. The simplicity and in general satisfying accuracy of air temperature index or degree day models make them the most commonly used method to estimate snow melt rates when snow melt is based on radiation (and not on rain on snow events). Temperature is easy to measure and spatially explicit climate data is usually derived by laps rates computed from two or more climate stations that are commonly situated in the open field. However, the number of climate stations and as a result the accuracy of temperature input into a distributed model is limited and spatial variability of snowmelt cannot be modelled accurately since melt rates tend to vary with elevation, vegetation cover, aspect and shading for example. We can also incorporate these factors into our models, but then we increase the number of parameters significantly. In the Cotton Creek (17.4 km) watershed study, we use micro temperature loggers (iButtons) for measuring air temperature at 40 sites to generate temperature input data for a distributed snow melt model at a variety of elevations, aspects and forest covers. In this study, we compare temperature index models that incorporate the spatial variation derived from these temperature loggers with a conventional approach that only considers data from 3 climate stations in the watershed. In a next step we incorporate the temperature data from the iButtons to develop a scaling procedure for climate station temperature data that better describes the actual spatial variability of air temperature in the watershed. Snow depth and snow water equivalent (SWE) at 1st of April (e.i. end of snow accumulation) was measured at 20 stratified sites within the watershed (snow depth at 60 locations per site, SWE at 9 locations per site). Two additional snow courses at bi-weekly intervals during the snow melt are used to test the different models in space. In addition, we use iButtons at the soil surface that measure the time of snow disappearance (sudden jump in temperature after the snow has melted) at the stratified locations as a test for the different model performances.