Loss of life modelling for a GIS framework to analyse the impacts of dam breaks

In recent years, advancements in dam management and construction have greatly contributed to improved dam reliability. However, failures still occur due to natural disasters, human error, and structural deterioration, often leading to severe flash floods. This research presents a GISbased framework aimed at assessing the impacts of dam failures, with a primary focus on loss of life (LOL) estimation. To achieve this, an improved LOL estimation model was proposed using a dataset of historical dam failures, divided by flood severity and evaluated through multivariate regression analysis. Key influencing factors—such as flood depth, population vulnerability, warning time, and evacuation conditions—were systematically weighted using entropy-based methods, allowing for a more accurate representation of risk levels in different failure scenarios. The framework integrates GIS spatial analysis with hydraulic modelling to evaluate potential damages resulting from dam failure. Results from model validation against historical cases demonstrated that the proposed LOL model achieved a strong fit to the observed data with a coefficient of determination (R²) value of 0.99 for both medium-high severity cases and for low-severity cases, outperforming conventional models, such as Graham's model. By integrating LOL prediction with spatial analysis, this framework provides assessment that can be incorporated into hazard mitigation plans, emergency response plans, and can also be used to inform insurance policies and risk mitigation strategies.