PREPARING COASTAL EROSION VULNERABILITY INDEX IN ODISHA APPLYING GEOSPATIAL AND MACHINE LEARNING TECHNIQUES

Abstract

With the changing climatic and anthropogenic conditions, the natural ecosystem especially the coastal zones is at great risk. The ocean is engulfing the land through the process of coastal erosion and this is becoming a great threat to coastal communities by forcing them to relocate from their homes and destroying their livelihoods. To understand the severity of the coastal erosion by keeping the usability aspect in mind, this study has considered the whole Odisha coast as the area of interest and has studied the effect of coastal erosion as a statistical analysis to evaluate the Net Shoreline Movement (NSM), End Point Rate (EPR), Linear Regression Rate (LRR) with the help of Digital Shoreline Analysis Tool (DSAS) at the blocks level for all the 22 coastal blocks of the 6 coastal districts. The area that has altered in the past has been determined to better comprehend the impact of the erosion process in the state. The future trend for the coastline position for 2030 and 2040 has been forecasted which showed the estuary positions are going to be affected most. Considering the Landsat satellite data to manually delineate the shoreline position from 1973 to 2020, the analysis showed average NSM, EPR, and LRR values of -84.95m, -1.81 m/year, and -0.36 m/year respectively with 72.47 sq. km of the eroded coastal area against 42.83 sq. km of newly formed landmass by the coastal dynamic process. Furthermore, this study has tried to use machine learning algorithms for the first time to find the probability of the vulnerability associated with this hazard along the coast of Odisha state of India using a total of 32 factors involving environmental and socio-economic conditions. A total of 2500 locations have been used to create Support Vector Machine (SVM), Random Forest (RF), Shallow Neural Network (SNN), Deep Neural Network (DNN), and Convolutional Neural Network (CNN) models. Various accuracy metrics have been calculated which showed the RF model outperformed all with an accuracy score of 0.96. iv This is followed by CNN (0.93), DNN (0.91), SVM (0.88), and SNN (0.88). Further, factor importance analysis by RF has been performed at state, district, and block levels to understand the influence of various parameters in this disaster. The study showed that mitigation and preventive measures are of utmost importance for the coast. This novel method will broaden the approach which we use to analyze this calamity and serve as an aid in the decision-making process of the concerned authorities.