The frequent occurrence of extreme weather and the development of urbanization have led to the continuously worsening climate-related disaster losses. Socioeconomic exposure is crucial in disaster risk assessment. Social assets at risk mainly include the buildings, the machinery and the equipment, and the infrastructure. In this study, the wealth capital stock (WKS) was selected as an indicator for measuring social wealth. However, the existing WKS estimates have not been gridded accurately, thereby limiting further disaster assessment. Hence, the multisource remote sensing and the POI data were used to disaggregate the 2012 prefecture-level WKS data into 1000 m × 1000 m grids. Subsequently, ensemble models were built via the stacking method. The performance of the ensemble models was verified by evaluating and comparing the three base models with the stacking model. The stacking model attained more robust prediction results (RMSE = 0.34, R² = 0.9025), and its prediction spatially presented a realistic asset distribution. The 1000 m × 1000 m WKS gridded data produced by this research offer a more reasonable and accurate socioeconomic exposure map compared with existing ones, thereby providing an important bibliography for disaster assessment. This study may also be adopted by the ensemble learning models in refining the spatialization of the socioeconomic data.

Against the backdrop of global climate change and rapid urbanization, island cities face increasing compound-flood risks due to their fragmented topography and limited terrestrial space, yet cross-scale modeling and high-resolution exposure assessment for archipelagic environments remain insufficient. Taking Zhoushan City, China, as a case study, this research develops an integrated compound-flood multi-receptor exposure assessment framework combining cross-scale hydrodynamic modeling with high-resolution spatialization. For hazard assessment, a one-way sequential coupling of a parametric typhoon model, FVCOM, and LISFLOOD-FP was employed to simulate storm-surge, pluvial, and compound flooding at 50-, 200-, and 1000-year design levels. Model performance was evaluated using tidal-gauge observations, historical storm-surge records, remotely sensed historical inundation, and documented urban waterlogging locations. For exposure assessment, Random Forest-based dasymetric mapping was used to generate 30 m population and secondary- and tertiary-sector GDP datasets; primary-sector GDP was spatialized using a land-cover-based approach, and 3314 critical-infrastructure facilities were compiled across eight subsystems. The results reveal the following: (1) compound-flood inundation areas were 289.95, 424.93, and 513.31 km² under the 50-, 200-, and 1000-year design scenarios, respectively; (2) under the 1000-year compound scenario, 57.5% of the population, 51.3% of GDP, and 63.3% of critical infrastructure were exposed; (3) infrastructure exposure showed strong subsystem heterogeneity, with coastal facilities having the highest exposure ratio (76%) and emergency-rescue facilities reaching 65.3%; (4) sector-disaggregated GDP exposure revealed substantial economic-geographic differences among districts/counties, highlighting the need for differentiated adaptation strategies. The proposed framework provides spatially explicit information to support differentiated coastal adaptation planning and is transferable to other archipelagos and small island cities with similar data constraints.