A national network for snow monitoring in Norway: Snow pillow verification using observations and models

Abstract

Snowmelt makes a substantial contribution to spring floods in Norway. The most severe floods, such as the flood in southeast Norway in 1995, are fed from extensive snowcovered high-mountain areas. However, monitoring of the temporal and spatial variability of snow on a real-time basis is particularly difficult due to the vast extent, remote location and high-frequency variability of snow. To monitor the temporal evolution of the snow mass and its water content during winter and spring, a network of 23 snow pressure pillows has been established in Norway, covering 58°N–71°N, 6°E–28°E, and 30–1400 m above sea level. Hourly data are supplied twice a day to government agencies.

During the 1998/1999 winter and spring, extensive manual sampling was conducted on a monthly basis to verify the measurements obtained from the snow pillows. Furthermore, nearby meteorological data were used to simulate snow accumulation and ablation using a snow model. To investigate the performance of the snow pillow network, manual snow surveys (depth and density, liquid water, stratigraphy and grain size), snow models (SWE, snow runoff, LWC) and nearby air temperature and precipitation data were analysed. The results are important for snow pillow deployment and maintenance, as well as snowmodelling in terms of historical simulations and spatial-temporal variation in model performance and parameter settings. The results show that snow accumulation was well simulated using the model. Snowmelt was not so easy to simulate. The snow pillow performance was not as good as expected, and it was obvious that the snow pillows did not respond well during periods of repeated melting and refreezing. Discrepancies were also observed between snow pillow and manual observations during the melting period in spring, which may be attributable to difficulties during the snow survey sampling.