A Developed Hybrid Integrated Framework with Combined Analytical Approaches in mitigating the Flood and Drought Risk on River Severn Basin

Floods and droughts are among the most devastating natural disasters, significantly impacting environmental and socio-economic systems. With climate change exacerbating these risks, it is crucial to develop robust frameworks in assessing and mitigating the risk. This thesis aims to develop a hybrid integrated framework for flood and drought risk assessment, combining multiple methodologies and modern predictive techniques. The research employs a combination of Interpretive Structural Modelling (ISM), Causal Loop Diagrams (CLD), and network theory to build the framework. Statistical and machine learning methods are used to calculate and test the framework, ensuring a comprehensive analysis. The integrated framework effectively identifies and assesses key risk factors and their interdependencies. The spatio-temporal mapping revealed significant trends in flood and drought occurrences. Despite the presence of flooding risk partly due to more intense rainfalls, the risk of drought coexists on a river basin scale. Validation using Receiver Operating Characteristic (ROC) curves demonstrated the model's accuracy. Sensitivity analysis highlighted critical variables such as community resilience, precipitation, access to transportation networks, and reservoirs, which contribute significantly to the variance of predicted risks. Other parameters aid in the accuracy of these predictions, while factors like elevation and slope assist with the spatial distribution of the risks. The developed framework has shaped and enhanced substantial understanding of flood and drought risks, providing a robust basis for future research. Future work should focus on integrating more diverse datasets and exploring long-term climate impacts to further refine and improve the assessment process.