Browsing by Author "Jutila, Arttu"

Studying and modelling the snow distribution processes is important because snow influences the ground, flora, and fauna by affecting among other things the energy balance both in large and small scales and the near-surface atmospheric conditions due to its highly reflective and insulating properties. The aim of the study was to use the spatially distributed high-resolution snow-evolution modelling system SnowModel to simulate the snow conditions in winter 2015-2016 in the Saariselkä region in Northern Finland and assess the model's performance. SnowModel has not been used to study a domain in Finland before, and the model gives information about variables that are hardly measured in Finland, such as snow sublimation. The simulations were first run without snow water equivalent assimilation and then assimilating the available snow water equivalent (SWE) observations. The simulation results show that in the default mode the model needs assimilation and SWE observations, preferably more frequent observations towards the spring, to produce physically sensible results. The domain averaged simulated end-of-winter maximum SWE value of 220 mm was reached on 21 April 2016. The simulated SWE patterns match with known elevation and vegetation dependencies. Timing of the first snow, the beginning of the snow season and the end-of-winter SWE are simulated well, whereas the melt and the snowfree date depend on the amount of snow. The assimilation run suggests that the needed summed precipitation is as much as 18 % larger than the observed increasing towards the northeast. Similarly, the simulated summed melt reaches locally up to 70 % larger values compared to the non-assimilation run. Blowing-snow sublimation takes place in open areas and its simulated summed value is up to 27 mm. Simulated static-surface sublimation varies between 4-22 mm. The simulated sublimation from the canopy-intercepted snow peaks at 110 mm. Up to 16 % of the precipitation is returned to the atmosphere by sublimation. The simulation results could be improved by utilizing more detailed data of the study domain and modifying the hard-coded variables to suit the surroundings, which could in turn decrease the need for assimilating SWE observations.