The need for improved processes understanding to develop enhanced knowledge and modelling strategies is a central issue within the Prediction in Ungauged Basins (PUB) initiative. Prediction of snow-cover depletion and spring-melt runoff in arctic basins is particularly challenging due to the spatial heterogeneity of the snow-cover as a result of topographic and vegetation effects on snow accumulation, wind redistribution, and ablation processes. Additionally the remote location and winter inaccessibility of Arctic basins contribute to the lack of proper data to model the hydrology of these sites. This study was conducted in two topographically distinct cold regions research basins in the north-west of Canada characterised as tundra environments, one an alpine mountain tundra at 60°N and the other a rolling planar tundra at approximately 70°N. This paper examines the transference of parameters of a distributed physically based land surface hydrological model from one site to another across 1350 km using a step-wise calibration procedure. The transference methodology uses a physiographic similarity criterion to identify similar landscapes units and makes use of landcover-based parameters rather than the usual relationships between parameters and basin characteristics. Results showed that simulations using regionalised vegetation parameters performed better than those using default parameters and accurately described both snow-cover ablation and snowmelt runoff in both basins.