TR(BR)-6/96-97 : Automated mapping of snow cover using IRS 1C data

Abstract

Snow is an important phase of the hydrological cycle and is of much importance in India because of the presence of mighty snow clad Himalayas. Some of the great rivers of India, like Indus, Ganga and Brahmaputra originate from Himalayas. Estimation of snowmelt runoff in these rivers, either seasonal or daily, is of immense use in proper management and use of water resources in these basins. Real-time and long-term forecasting of runoff in these basins can result in optimum operation of reservoirs for conservation and flood control, for planning hydropower regulation and hence optimum use of water. '

A number of models have been developed for simulation and forecasting of snowmelt runoff. Data requirement of these models is quite high and conventional ground measurement in snow covered areas can not meet these requirements because of various shortcomings like frequency of observation, point measurement not representative of large areas, hostile climate conditions and inaccessibility of areas. Since last two decades, remote sensing techniques have offered excellent synoptic and repetitive overviews in various spectral channels of electromagnetic spectrum which has served as a spatial data base for snow related studies.

The Indian Remote Sensing Satellite, IRS-1C was launched in December, 1995. The LISS-III sensor of this satellite has higher spatial resolution, i.e. 23.5 m. In addition, a spectral channel in the wavelength range 1.55 -1.7 µm has been added in the sensor while the channel below 0.52 µm has been dropped. The new channel is—highly useful for discriminating snow cover areas from clouds. In the present study, the information of the new channel has been utilized for determining the snow cover area in the Spiti sub-basin of the Sutlej river basin. A procedure, developed by Dozier, has been utilized for automated mapping of snow cover area in the basin. It has been found that the Dozier's algorithm can be utilised for snow cover estimation using the remote sensing data of LISS-III sensor. However, because of the deficiency of satellite information below 0.52 Am with this sensor, it is not possible to directly identify snow in shadowed areas and under the clouds. An indirect way of predicting snow in shadowed areas and areas under cloud cover has been suggested. This method is based on the combined use of digital image processing techniques (Proximity analysis) and the topographic details. For getting the topographical details, Digital Elevation Model (DEM) for the area was developed using a GIS software, ILWIS. Topographic information such as slope and aspect for each pixel was derived from DEM and used as data for the proximity analysis.