Optimisation of strategies using spatial approaches to manage flood risk in Thailand

Flooding poses a major and escalating threat across Thailand, particularly within the Chao Phraya River Basin where complex hydrological processes interact with rapid urbanization, socio-economic inequality and long-standing governance challenges. While GIS-based flood risk assessment is widely applied internationally, limited research has examined model transferability, data sensitivity and the integration of expert knowledge in data-constrained environments. This thesis develops an optimized GIS-based flood risk assessment framework intended to support evidence-based decision-making for Thai local authorities. The research integrates three components: a UK pilot study to examine model behavior in a data-rich environment, a full transfer and localization of the model to Thai conditions using publicly available hazard, exposure and vulnerability datasets, and an expert-elicitation phase involving practitioners from key national and provincial agencies. Statistical methods including Principal Component Analysis (PCA), Ordinary Least Squares (OLS), and Geographically Weighted Regression (GWR) were applied to diagnose variable behavior, minimize multicollinearity and identify robust predictors of flood risk. The results show that flood frequency is the strongest hazard indicator, while vulnerability variables, especially education, income and savings, consistently outperform exposure factors in explaining spatial patterns of flooding. Expert consultations support many of these findings while also highlighting aspects of long-term experiential knowledge that are not fully captured by statistical relationships alone. Comparative mapping shows that expert weightings tend to broaden mid-range classifications, whereas GWR produces sharper and more localized clusters of risk, indicating stronger analytical reliability. The research contributes to the academic literature by demonstrating the transferability of GIS-based flood risk modelling across contrasting international contexts and by providing a practical framework for managing uncertainty and integrating expert knowledge within quantitative spatial analysis. Overall, the research provides a transferable and data-efficient flood risk modelling approach suited to the constraints of Thai local governance and offers practical guidance for sub-district planning, early warning development and climate-resilient flood management.