Assessment Of Snow Status Changes Using Temporal-coherence Components Decomposed From Multitemporal SAR Data At Alpine Mountains

Synthetic Aperture Radar(SAR) can observe the Earth surface without solar illumination, and penetrate frogs and clouds especially in the heavy cloud alpine regions. Therefore, SAR is one of important snow/glacier remote sensing techniques. The interferometry and coherence of InSAR technique reveal the variations of the geometry and physical properties of ground objects. The physical properties of different surface features vary greatly over time span. Thus, the coherence change map shows significant effects on surface features classification and change detection in alpine snow/glacier. Because the snow/glacier cover over alpine mountains is a vita water resource, and snow status changes are important parameters for snow runoff models, there is the need to assess spatial distribution of snow/glacier effectively. In this thesis, several important snow cover methods using SAR techniques, their characteristics and applicability, and possible limitation among the methods were discussed. Coupled with the characteristics of InSAR technique, a method to study the snow cover and snow status variation at Mtn. Dagu in western Sichuan Province was developed. Nine snow freezing/thawing status changes were analyzed. In particular, the observed multitemporal Phased Array type L-band Synthetic Aperture Radar(PALSAR) HH coherence data was decomposed into temporal-coherence, spatial-coherence, and thermal noise components. The data spanned between February of 2008 and May of 2008, and consisted of 2 pairs of interferometric SAR(In SAR) images formed by three consecutive repeat passes. With the analysis of ancillary data, a snow increase process and snow decrease process were determined during the time span. Then, the temporal-coherence component was used to study the variation of thawing and freezing statuses of snow because the component can mostly reflect the temporal change of snow happened between two data acquisitions. Combined with temperature records, digital elevation model(DEM) and optical data, we found that a threshold for the component could be used to detect snow status change through time. Furthermore, a method to map snow cover change using the component was developed. Compared with optical images, the outcomes from the snow increase process and snow decrease process reached an overall accuracy of 71.4% and 79.4%, respectively. Being capable of delineating not only the areas with or without snow cover but also status changes among free-snow, wet-snow, and dry-snow, a critical and an alternative means to assess the water resource in alpine areas has been developed.