Abstract
The seasonal snow cover is an important resource in mountain regions such as the Alps, the monitoring of which is crucial for water management, snow hydrology, tourism, and natural hazard mitigation. Snow has a high societal economic impact on people living in the Alps and is an important source of headwater for drainage basins downstream. AlpSnow (2020-2022) is a science activity within ESA's Alpine Regional Initiative, addressing the development of novel Earth Observation techniques and algorithms for the generation of innovative and consistent snow products optimized for specific scientific and operational applications. Input data for AlpSnow products are primarily provided by the Copernicus Sentinel satellite family, but also data of third-party missions such as the SAR missions TerraSAR-X and TanDEM-X (X-band SAR), SAOCOM (L-Band SAR), and PRISMA (hyperspectral sensor) are used. The products cover several physical parameters of the seasonal snowpack, namely snow area extent, snow surface albedo and grainsize, snow water equivalent, snow depth, snowmelt area extent and liquid water content. The usability of the products will be demonstrated within six scientific and operational use cases in the fields of meteorology, snow science, hydrology, and water management.
For development and validation of the algorithms we selected five test areas in different Alpine regions, covering a wide range of elevation zones and meteorological conditions. Each of the test areas, two of which are part of the INARCH activity, is equipped with operational field stations providing time series of meteorological and snow measurements. These are supplemented by specific activities in snow research and hydrology, including the collection of additional snow reference data for validation, as snow pits and snow transects measured during field campaigns.
In the first phase of the project several candidate algorithms for each of the snow parameters were selected, implemented, tested, and evaluated in the test areas. Based on this intercomparison we selected a set of preferred algorithms which are further developed, optimized for environments with complex topography and diverse surface cover. Regarding high resolution snow extent, we intercompare two algorithms, an advanced linear multispectral unmixing approach and a new method applying machine learning techniques. Both methods, developed within the project, are exploiting the spectral capabilities of Sentinel-2 and Landsat satellite sensors. Surface albedo and grain size products are generated adopting the algorithms from Dozier and Painter (2004) and Kokhanovsky (2015). A critical issue for both algorithms is the correction of topographic effects on the spectral irradiance and on the bidirectional reflectance distribution. The parameters wet snow extent and snow water equivalent or snow depth are derived from SAR data. For snow water equivalent two approaches are explored. The first approach develops methods for assimilation of EO snow extent products into the physical SNOWGRID model of the Austrian meteorological service (ZAMG) that is driven by numerical meteorological data and provides operationally spatially distributed snow information over the Alps. Additionally, we study the suitability and limitations of repeat-pass SAR interferometry for generating maps of snow accumulation in mountain regions with a spatial resolution in the order of about hundred meters. The interferometric phase provides a direct physical measurement of the change of the snow mass during the time span between two image acquisitions but works only for dry snow conditions. These developments are tested using L-Band SAR data from ALOS-2 PALSAR and SAOCOM and prepare for the Copernicus Expansion Mission ROSE-L. To derive maps of total snow depth we investigate the suitability of differencing DEMs from bistatic TanDEM-X data acquired during snow free conditions and conditions with maximum snow accumulation at the start of the snowmelt season.
We will present the results of the experiments dedicated to the selection of retrieval algorithms for the different snow parameters and report on the accuracy of the prototype algorithms and products developed within the project.