Implementation and Performance Analysis of Various Techniques for Snow Cover Monitoring in Western Himalayas

Abstract

Himalayan region has very high concentration of mountain glaciers and it is completely covered by snowfall during winter period between November and April. Very few studies on monitoring of snow cover have been reported in this region. However, the demand for snow cover monitoring in this region has gained importance in order to find the solution to some of the related issues such as global warming, lack of water supply, electrification in rural areas and hazard forecasting etc. The availability of moderate resolution image data almost on day today basis from sensors such as MODIS and AWiFS has provided great opportunity for regular mapping and monitoring of snow in this region. The present thesis entitled “Implementation and Performance Analysis of Various Techniques for Snow Cover Monitoring in Western Himalayas” deals with evaluating the performance of various topographic correction techniques on the snow spectral characteristics obtained from MODIS and AWiFS data and suggesting the most suitable technique for this region. It also deals with the monitoring of snow using per pixel and sub-pixel techniques and a new statistical linear regressive relationship has been developed between normalized difference snow index (NDSI) and fraction of snow covered image obtained by Linear Mixture Modeling and Fuzzy c- mean Clustering in the Beas region of Western Himalaya. The study also been carried out to evaluate the effect of topographic corrections on the accuracy of results obtained using Improved Change Vector Analysis. Chapter 1 as ‘Introduction’ provides brief description about snow, snow cover, importance of snow cover monitoring in general and in the Himalayan terrain in particular. Basic principles of satellite based remote sensing and spectral reflectance characteristics of various land covers are also briefly explained. The layout of the thesis has been given at the end of the chapter. Chapter 2 covers a comprehensive literature survey of various models and methods of snow cover monitoring, influence of topography on snow spectral characteristics, the change detection techniques in snow cover area and the impact of topography on the accuracy of detected changes, the implementation of sub pixel techniques for improving the classification accuracy in the imagery obtained from satellites providing higher temporal resolution but medium to moderate spatial resolution, thereby, providing the motivation for this work. It also covers the problem formulation and methodology for the present work. In chapter 3, the brief overview of the ‘study area and data used’ is given. The topography of Himalaya in general and western Himalaya in particular has been described. The importance of the study area is explained. The data used for the analysis i.e. satellite imagery, map sheets, digital Elevation Model (DEM) and data pertaining to field observation used for validation is also described. Chapter 4 is devoted to the qualitative and quantitative comparative analysis of different topographic models for Himalayan terrain. The main objective of the research in this area is (i) to implement all topographic models for qualitative and quantitative assessment for comparative analysis of snow reflectance estimation using medium resolution satellite imageries forming of a part of Western Himalaya obtained from AWiFS sensors on-board IRS-P6 (ii) visual analysis (iii) validation of terrain corrected satellite estimated reflectance by all these methods with field measurements (iv) testing of topographically corrected results for snow samples on the south and north aspect (v) graphical analysis after topographic corrections and (vi) to find the best suitable method for topographic corrections for Western Himalaya. Chapter 5 deals with the models and methods of snow cover monitoring. Two sub-pixel classification algorithms such as Fuzzy C-Means Clustering (FCM), Linear Mixture Modeling (LMM), in soft form are used for the precise estimation of various classes in the moderate resolution MODIS images. Thereafter, a new Statistical relationship is developed between the NDSI from MODIS observations and the fraction of snow-cover at 500-meter grid resolution obtained by LMM and FCM for use as a universal relationship applicable for the entire study area.. Chapter 6 considers the effect of topographic corrections on accuracy of change detection in Himalaya terrain. The main focus of the work here is to analyse the effect of topography on improved CVA and carry out accuracy assessment as compared to conventional CVA. Chapter 7 covers the summary of conclusions drawn in all the three objective areas. In this thesis work it has been concluded that the effect of topography plays a major role in the accuracy of results obtained from various snow cover monitoring techniques in the rugged mountainous terrain of Himalayas. It is also concluded that none of the present satellite sensors is exclusively suitable for accurate snow cover monitoring in the Himalayan region. Therefore sub pixel analysis is required to be carried out based on the empirical results. At the end future scope of the work has been proposed. The final chapter 8 ‘List of References’ provides the details of all the references used and referred in this study and thesis work. The ‘Appendix-A’ gives a brief about the author