Desertification characterization using predictive soil modelling and pattern recognition

Abstract

quotThis thesis presents a hierarchical methodology for land degradation mapping, newlineland use land cover classification, degradation process identification and map- newlineping using multispectral LISS-3 images. The study aims to demonstrate the im- newlineportance of remote-sensing images for various applications, both social and en- newlinevironmental. The study compares the results of different algorithms for different newlineterrains, demonstrating that Simple Linear Iterative Clustering (SLIC) segmenta- newlinetion with the random forest(RF) method outperforms CNN and pixel-based Sup- newlineport Vector Machine (SVM) with an accuracy of 85% for level 1 land cover clas- newlinesification. Vegetation degradation in forest areas is assessed in central parts of newlineGujarat, India, and land degradation in agricultural areas due to soil salinity is newlinestudied, particularly in southeastern parts of Gujarat, India. ML algorithms like newlinesupport vector machine(SVM) and RF was applied to different features to identify newlinethe degradation process. Temporal data were used to find the severity of deserti- newlinefication using the change in degraded areas. newlineFurther, it discusses soil degradation causing desertification and severely re- newlineducing potential soil productivity. The study uses machine learning algorithms newlineand an ANN-based model to predict soil properties like EC, pH, and OC, which newlineare important indicators of soil degradation. Environmental parameters are taken newlineas covariates in prediction models, including vegetation indices, terrain indices, newlinesoil parameters, spatial attributes, and meteorological parameters of the study re- newlinegion. Field soil sampling data of the study region obtained from Soil Health Card newline(SHC) for the year 2014 is incorporated in training the model. The SHC data is newlinedivided into different ratios for training and testing the model. The SCORPAN newlinemodel is considered the base approach for the development of the ANN-based newlineprediction model. Moreover, the thesis also discusses the mapping of vulnera- newlineble areas to desertification.