Improving Coastal Adaptation: Climate Change Impact Modelling for Resilient Communities

Abstract

Continuous change in climate highlights the vulnerability of coastal communities, there will be a lot of worries among the coastal communities with the risk of rising sea levels, storm surges, and extreme weather conditions, The factors that affect flooding in low-lying coastal regions are urban expansion, deforestation, and climate-driven changes. Machine Learning and geospatial tools are used to create accurate flood susceptibility maps and help in identifying high-risk zones and mitigating disasters. This study helps in identifying the risk of flood using two machine learning algorithms Random Forest and Support Vector Machine for predicting flood-prone areas. The experiment was performed on the Permanent Service for Mean Sea Level (PSMSL) dataset. The random forest algorithm accuracy of calculating Flood situation is 78%, and that of the Support Vector Machine it is 71%, also random forest shows high precision around 74% which is greater than the Support vector machine which is 71%, the comparison of recall is about 86% for random forest and that for Support vector machine it is 71%.

The ability of machine learning to accurately identify high-risk flood zones and provide reliable predictions for disaster preparedness is demonstrated by these results. In addition, this study investigates ML-driven methods for conserving cultural heritage and enhancing community-based climate adaptation. The study concludes by emphasizing the crucial role of ML-driven predictive models in climate resilience planning, particularly in improving flood risk assessments and disaster mitigation, and suggests integrating real-time big data analytics, cloud computing, and nature-based solutions for stronger coastal protection and sustainable adaptation