1Météo-France – CNRS, CNRM – GAME UMR3589, GMAP, Toulouse, France
2Météo-France – CNRS, CNRM – GAME UMR3589, CEN, Grenoble, France
3Université Joseph Fourier – CNRS – Grenoble INP, LTHE UMR5564, Grenoble, France
4School of Surveying, University of Otago, P.O. Box 56, Dunedin, New Zealand
5Institut National des Sciences Appliquées, Toulouse, France
6Météo-France – Dprévi, Toulouse, France
7Institut de Mathématiques, Université de Toulouse et CNRS (UMR5219), 31062 Toulouse, France
8Snow & Avalanche Study Establishment, Him Parisar, Sector-37A, Chandigarh 160 036, India
9Institute of Environmental Physics, University of Bremen, O. Hahn Allee 1, 28359 Bremen, Germany
Received: 11 Apr 2012 – Published in The Cryosphere Discuss.: 31 May 2012
Abstract. This study compares different methods to retrieve the specific surface area (SSA) of snow from satellite radiance measurements in mountainous terrain. It aims at addressing the effect on the retrieval of topographic corrections of reflectance, namely slope and aspect of terrain, multiple reflections on neighbouring slopes and accounting (or not) for the anisotropy of snow reflectance. Using MODerate resolution Imaging Spectrometer (MODIS) data for six different clear sky scenes spanning a wide range of snow conditions during the winter season 2008–2009 over a domain of 46 × 50 km in the French Alps, we compared SSA retrievals with and without topographic correction, with a spherical or non-spherical snow reflectance model and, in spherical case, with or without anisotropy corrections. The retrieved SSA values were compared to field measurements and to the results of the detailed snowpack model Crocus, fed by driving data from the SAFRAN meteorological analysis. It was found that the difference in terms of surface SSA between retrieved values and SAFRAN-Crocus output was minimal when the topographic correction was taken into account, when using a retrieval method assuming disconnected spherical snow grains. In this case, the root mean square deviation was 9.4 m2 kg−1 and the mean difference was 0.1 m2 kg−1, based on 3170 pairs of observation and simulated values. The added-value of the anisotropy correction was not significant in our case, which may be explained by the presence of mixed pixels and surface roughness. MODIS retrieved data show SSA variations with elevation and aspect which are physically consistent and in good agreement with SAFRAN-Crocus outputs. The variability of the MODIS retrieved SSA within the topographic classes of the model was found to be relatively small (3.9 m2 kg−1). This indicates that semi-distributed snowpack simulations in mountainous terrain with a sufficiently large number of classes provides a representation of the snowpack variability consistent with the scale of MODIS 500 m pixels.
Revised: 29 Mar 2013 – Accepted: 07 Apr 2013 – Published: 29 Apr 2013
Citation: Mary, A., Dumont, M., Dedieu, J.-P., Durand, Y., Sirguey, P., Milhem, H., Mestre, O., Negi, H. S., Kokhanovsky, A. A., Lafaysse, M., and Morin, S.: Intercomparison of retrieval algorithms for the specific surface area of snow from near-infrared satellite data in mountainous terrain, and comparison with the output of a semi-distributed snowpack model, The Cryosphere, 7, 741-761, doi:10.5194/tc-7-741-2013, 2013.