Abstract
In New Zealand it has long been recognised that vegetation plays an important role in the distribution of snow cover. However, most studies have concentrated on snow tussock grassland and the role of other types of grassland have been ignored. A field study on the Pisa Range during the 1992 snow season measured snow cover patterns in induced short tussock grassland. This provided information on snow cover processes that led to construction of a computer simulation model of vegetation and snow cover. Snow cover measurements were taken at four-weekly intervals within an area of induced short tussock grassland where topographic influences were minimal. An automatic climate station was set up nearby. A plot study measured the effects of overgrazing on snow cover patterns. Two measurement plots, one of which was clipped to simulate overgrazing, were established at 1400m. This experiment was repeated at 1200m. Regular snow surveys were completed at plots throughout the snow season. Results of the field study show that snow cover in induced short tussock grassland is highly variable both temporally and spatially. Much of the spatial variability is caused by vegetation. During periods of accumulation, individual tussocks trap snow producing snow drifts with the form of a half cone. Once these drifts begin to overlap one another, the snow pack builds up until the vegetation is buried. During warm sunny periods, tussocks induce snow melt around them because of their radiative properties. This results in a distinctive pattern of snow melt hollows around individual tussocks. Effects of overgrazing on snow cover patterns vary depending upon whether winter ablation is an important process or not. If winter ablation is negligible, overgrazing reduces the accumulation of snow prior to the vegetation becoming buried. However, it also reduces the overall rate of snow melt. Overgrazing has little effect on snow accumulation in areas where winter ablation is an important process but it reduces the rate of snow melt during the early stages of snow melt. An empirically-based numerical model is developed to simulate snow cover distribution in short tussock grassland. The model requires further testing but results from initial simulation runs are presented. With further development, the model could be applied to examine effects of landuse change on snow cover data of Central Otago block mountains.