Recognition of policy relevant spatial patterns from high resolution multirisk data – the case of China’s water risk portfolio

Accumulation of high-resolution data relevant to water resources management and policy is fast and accelerating. The related analytic approaches to harness this data are also in rapid evolution. However, the distance form data analysis to policy making remains vast, as policy level actors usually appreciate spatially aggregated information on units such as river basins or provinces, and parallel consideration of various issues such as hazards, stressors, exposure factors, vulnerabilities, and risks, which still rarely are brought together by data scientists in scales that would readily communicate with planning units of water resources. Further, scaling from these planning units to local conditions is of great value. China, as a sizable and geographically heterogeneous country, is subject to a high diversity and blend of water related stressors, hazards, and conditions of exposure and vulnerability. We present results of the exposure and vulnerability of continental China’s eight major water stresses (variability, overuse, groundwater problems, floods, droughts, organic pollution, salinity, eutrophication). To be maximally policy compatible, this gridded high-resolution geospatial analysis employs the multiplicative risk scheme of the United Nations Sendai Framework for Disaster Risk Reduction and IPCC (risk = stress x exposure x vulnerability) and is combined with multivariate statistical pattern recognition (unsupervised learning based on eigenvalue analysis). The results unveiled five distinct zones in continental China, each with a characteristic risk profile, for both provinces and river basin planning units.