INVESTIGATING POLARIMETRIC SAR DATA FOR CRYOSPHERIC MONITORING IN A CANADIAN ENVIRONMENT Monique Bernier, Yves Gauthier, Stéphane Mermoz, Imen Gherboudj, Ali El Battay and Jalal Khaldoune INRS-EAU, Terre & Environnement, 490 de la Couronne, Québec (Qc), G1K 9A9, Canada. Phone: 1-418-654-2585, e-mail: Monique.Bernier@ete.inrs.ca Abstract - Since the early 90’s, INRS has been developed tools for snow monitoring, river ice characterization and seasonal frost mapping in Canada. The focus to date has been on the use of monopolarized or multipolarized SAR data. With the forthcoming of RADARSAT-2, we have undergone a series of studies to asses the potential information gain from polarimetric SAR data. Airborn polarimetric SAR data from the Canadian CV-580 have been acquired over three different Canadian sites in winter: 1) a Boreal forest 2) an agricultural watershed and 3) the Saint- François River. This paper presents the preliminary results obtained from the polarimetric data set over the Saint-François River in February 2003. KEYWORDS: Polarimetric SAR data, Convair-580, River ice monitoring, frozen soil, snow, RADARSAT, ASAR. I. INTRODUCTION The terrestrial cryosphere consists of snow cover, glaciers, lake and river ice, and seasonally frozen ground and permafrost. Since the early 90’s, INRS has been using remote sensing technology to develop tools for snow monitoring, river ice characterization and seasonal frost mapping in Canada [1][2][3][7]. The focus to date has been on the use of monopolarized or multipolarized SAR data (ERS, RADARSAT, ASAR). However, with the forthcoming of RADARSAT-2 and its polarimetric features, we have undergone a series of studies to asses the potential information gain from such data. Recently, Airborne polarimetric SAR data from the Canadian CV-580 have been acquired for the Canadian Space Agency over three different Canadian sites in winter: 1) a Boreal forest 2) an agricultural watershed, 3) the Saint-François River. The image acquired over the boreal site (mainly lakes, conifers, clear cuts and regrowth areas) presented some calibration problems and could not be fully exploited. We nonetheless compared the Cloude polarimetric parameters from a winter image (March 2001) and a snow-free image (May 2001), in the context of monitoring the snow parameters. Although few field data were collected, it is observed that in the presence of a snow cover, the backscattering coefficients are lower (frozen soil), as are the scattering entropy (H) and the alpha angle (α). Furthermore, the anisotropy is higher in regrowth areas. Over the agricultural site, polarimetric SAR images were acquired under frozen (February 2003) and non- frozen (November 2002) conditions. The seasonal soil freezing is a frequent phenomenon which touches practically all the agricultural areas in northern climates. It represents a determining parameter which has an impact on climate, hydrology, agriculture production, soil erosion, and water quality [12][13]. The potential of polarimetric parameters and decomposition algorithms for monitoring the frozen soil is presently under study through a GIS approach (backscattering coefficients, scattering entropy and alpha angle, land use, soil type, soil temperature, field contours). Over the Saint-François river, our objective is to derive a river ice map from the polarimetric data (February 2003). This map will also be compare with an unsupervised classification of a RADARSAT-1 image (HH, Fine mode) acquired simultaneously. This paper focuses on the first interpretation of the polarimetric decomposition algorithms for river ice. II. MONITORING RIVER ICE TYPES A. Context Rivers do certainly play an important role in the water resource distribution, particularly in winter, when the land surface runoff is minimal. However, the presence of an ice cover will affect streamflow, modify the ecosystem, cause flooding, restrict navigation, influence hydropower generation and even impact on the climate. Canada is renowned for its magnificent rivers, its harsh but beautiful winters and its strong hydroelectric industry. This combination is reason enough for the need to monitor the river ice cover. However, aerial surveys and field trips in remote areas are costly and only provide surface or localised information. In that sense, SAR remote sensing could represent a reliable alternative, combining coverage, 0-7803-9050-4/05/$20.00 ©2005 IEEE. 36 Authorized licensed use limited to: BIBLIOTHEQUE DE L'UNIVERSITE LAVAL. Downloaded on February 5, 2010 at 10:29 from IEEE Xplore. Restrictions apply.