Demonstration activities for snow cover monitoring and hydrology using satellite data

Activities and examples for operational applications of satellite images for snow monitoring and snowmelt runoff forecasting in mountainous basins are presented. Because of the topographically induced variability of snow accumulation, ablation, and meteorological parameters spatially distributed information is an essential requirement for runoff modelling and forecasting. Therefore we developed versatile and robust tools for retrieval of snow cover information from optical and SAR satellite sensors and for using these products for modelling and forecasting snow and glacier melt. Automated procedures are applied for producing snow cover maps over large areas or customized products for individual drainage basins. The processing steps include radiometric calibration, geocoding, classification of snow, and post-processing to match the output to the needs of a particular application. A statistical approach is used to correct for sensor specific differences in the obtained snow information. This is very relevant for matching optical and SAR snow maps, because of systematic differences particularly in areas of patchy snow cover. The main sensors presently applied are medium resolution optical sensors (MODIS) and wide swath data of ENVISAT ASAR. During the last winter digital snow cover maps, superimposed by cartographic information, were produced on weekly basis for the whole European Alps, and also for the Himalayan mountains of Pakistan. These data area used for hydropower management and planning activities, and are as well very relevant for climate monitoring.

We demonstrated also the use of the satellite snow products for forecasting the daily runoff of 1 to 6 days in advance during several melting periods of the last years for the Zillertal and Ötztal basins, Austria. Individual forecasts were made for sub-basins where runoff gauges area available. Data links were established to receive ENVISAT ASAR data and MODIS data for snow information retrieval within less than 24 hours after image acquisition. Numerical weather forecasts were received from the Austrian weather service including deterministic forecasts as well as ensemble forecast products of the ECMWF model. Ensemble predictions (consisting of 51 forecasts fore each parameter and time step) enable to take into account the uncertainty of meteorological forecasts. Examples of the runoff forecasts and evaluations of the quality are presented. Time series of snow maps for the last winter periods were also produced for several basins in the Himalayan mountains of Pakistan the in order to support the planning of a network of hydropower reservoirs. The snow maps provide basic information on the contribution of snow melt to the runoff in areas where only very limited information is available. Examples of the snow cover changes for several basins in the investigation area will be presented.