PHYTOSOCIOLOGY AND LANDSAT TM DATA: A CASE STUDY FROM RIVER BEAS BED, PUNJAB, INDIA

摘要

The present work was designed to study the phytosociology and its correlation with Landsat TM data from the catchment areas of river Beas, Punjab, India, for a stretch of 63 km between the towns of Beas and Harike for four seasons, i.e., summer, pre-monsoon, post-monsoon and winter seasons respectively. Various phytosociological parameters (density, abundance, frequency, relative density, evenness and relative abundance) and diversity indices (Simpson's, Shannon's, Margalef's. Brillouin's, Chao-I and Menhinick indices) were studied. During the premonsoon and winter seasons. Beas recorded maximum abundance and density. Maximum abundance was found for Harike during the post-monsoon season. Maximum Menhinick, Maragalef's, Chao-I, Simpson's and Shannon's indices were found for the Harike during the winter season. Significant positive correlation of band ratios (G/R) was found with Simpson's. Shannon's and Brillouin's indices, whereas negative correlation of band ratios (R/NIR) existed with Simpson's, Shannon's and Brillouin's indices. Various multivariate statistical techniques Principal component analysis (PCA). factor analysis (FA) and artificial neural networks analysis (ANN)) were applied for the analysis of results. ANN models were fitted to the data. Correlation between target and output values was found to be highly significant.