Land Use Dependent Snow Cover Retrieval

The DLR-funded project “Integration of remote sensing data in operational water balance and flood forecasting systems” (InFerno) has been originated to establish an operational synergy of remote sensing and flood forecasting. Upon availability, InFerno intends to make excessive use of ENVISAT ASAR data, which will provide microwave imagery at a specifically appropriate spatial and temporal resolution. New methodologies, based on existing and ongoing research initiatives for the derivation of soil moisture and snow properties, are being developed in the catchments of the Ammer, Neckar, and Mosel rivers. These are brought to an operational level, by means of algorithm retrieval and software development, to automatically process remote sensing data and to assimilate this information in the hydrological model LARSIM. In the presented study, ENVISAT performance is simulated in a multitemporal and multisensoral attempt to delineate snow cover from RADARSAT and ERS imagery. For the retrieval of snow covered area (SCA) by means of SAR, image rationing has been established as the state-of-art method. It derives information on the areal extent of wet SCA, which, due to its critical energetic stage, is a tremendously important indicator for flood risk. To obtain areal maps of SCA, a landuse-dependent thresholding for the difference between the reference and the data take of interest is applied. For crop fields a threshold of –4 dB, for pasture a threshold of –2 dB is determined. The snow line obtained from SAR imagery is validated by point observations of the German Weather Service (DWD). Analyses show, that SAR imagery, presuming high spatial and temporal resolution, provides most promising results in terms of SCA detection. INTRODUCTION The potential of remote sensing data has gained increasing consideration in numerous hydrological applications (3-5). The possibility to image spatial characteristics and to retrieve manifold quantitative information over large areas has been recognised as a useful tool in complex water-related issues, such as water balance modelling or yield estimation. However, due to the coarse temporal resolution of available spaceborne imagery, the utilisation of remote sensing is yet uncommon in the field of operational flood forecasting and flood management, where information on soil and snow properties, as fundamental flood-steering parameters, needs to be provided at high temporal frequency. Methods to accurately assess and forecast flood discharge are a fundamental requirement in practical hydrology. Existing rainfall-runoff models seldom consider the spatial characterisation of the land surface. In order to obtain reliable flood predictions, an accurate physically based description of the processes contributing to runoff formation is essential. The new generation of remote sensing systems, such as ENVISAT, will deliver relevant spatial model input parameters at a sufficient temporal frequency to gather the information required for trustworthy flood forecasting. Proceedings of EARSeL-LISSIG-Workshop Observing our Cryosphere from Space, Bern, March 11 – 13, 2002 129