Monthly snowmelt modelling for large-scale climate change studies using the degree day approach

The degree day or temperature index snowmelt modelling approach has been adapted to provide a simple conceptual, semi-distributed snowmelt-runoff model suitable for large-scale climate impact investigations. The model is forced by pseudo-daily data derived from standard monthly meteorological data. Generic rather than site-specific parameters are used where possible to allow spatial and temporal transferability. Spatial distribution is achieved by elevation and broad vegetation cover. Snow accumulation and melt are determined from air temperature. Catchment runoff is simulated with a one-dimensional water balance sub-model consisting of two soil layers. Available water capacity, percolation and drainage rates are defined by soil texture. The model was tested in three catchments: Valuoja in Estonia and Gimdalsbyn and Kultsjön in Sweden. The catchment areas ranged from approximately 4–2200 km2, simulation lengths were 10, 21 and 26 years. The snow season timing matched that observed in all three catchments. Comparisons of measured and estimated snow water equivalence in Kultsjön demonstrated that the model can adequately predict snow cover below the tree line. Simulated monthly runoff at Valuoja gave good agreement with observations (r = 0.87) and a paired t-test showed no significant differences. The seasonal discharge pattern at Kultsjön also fitted well (r = 0.85) although the hydrograph generated was too low, probably due to precipitation undercatch. The discharge pattern at Gimdalsbyn was less well modelled (r = 0.55), but the annual runoff totals were similar suggesting a lagged runoff response. Sensitivity analysis showed that the model is insensitive to the generic parameters but is sensitive to local soil texture. The model is able to capture, from standard meteorological data, patterns of snow accumulation, melt and runoff with a reasonable degree of accuracy without the need to optimise parameters. These features make the model a potentially useful tool for evaluating the impacts of climate change on regional and continental water balances.