• The study presents a novel, fatality-focused typhoon risk index for data-scarce regions, a key gap in risk assessment. • The study presents a framework integrating climate variability, hydrology, Niño SST indices, and socio-economic data. • Results show deadly floods can occur in sub-catchments with low exposure, highlighting vulnerability and uncertainty drivers. • GPR shows fatality risk is nonlinear, driven by key variables, stressing integration of climate and hydrology in models. Insurers and reinsurers face a structural blind spot in typhoon-exposed emerging markets as human fatality risk is material for underwriting, parametric triggers, and reputational capital, yet is rarely quantified due to sparse and unreliable data. This study introduces a Typhoon Fatality Risk Index (TFRI) specifically designed for catastrophe risk assessment in data-scarce regions. Unlike traditional catastrophe models that focus on economic loss, TFRI directly quantifies fatality risk using a data-driven, transparent framework suitable for pricing, capital allocation, and portfolio management. Climate drivers such as Niño Sea Surface Temperature (SST) indices and Deep learning (DL)-assisted, hydrological discharge values are integrated via entropy weighting to reflect hazard intensity and variability. Gaussian Process Regression (GPR) also shows that fatality risk exhibits complex, nonlinear dynamics influenced by a few dominant variables, underscoring the importance of integrating climatic and hydrological data in risk models. This framework demonstrates significant potential to improve risk assessment and capital allocation for (re)insurers and overall decision-makers operating in data-scarce regions.
Necesito et al. (Wed,) studied this question.