Localising global urban development; simulating local exposure to natural hazards in the global 2UP model

Future population growth is expected to concentrate in urban agglomerations that are already exposed to numerous natural hazards. It is difficult, however, to assess this increase in risk as natural hazards are often concentrated in space and population growth scenarios tend to be defined at much coarser scales. By combining recently released high-resolution spatial data on land use, population density and natural hazards with a novel, computationally effective simulation approach we analyse global increases in local exposure to two important natural hazards: flood risk and landslides. We develop global spatially explicitp rojections of population change and urban expansion using a land-use and population allocation model. The model is developed in the Geo Data and Model Server (GeoDMS) modelling framework, that also underlies Land Use Scanner and several other operational models of land-use change developed for individual countries, larger river catchment areas and the territory of the European Union. The model disaggregates scenario-based national-level population estimates to a high resolution spatial grid (30 arc seconds). It simulates local population development and urban growth on a global scale. The main steps include: 1) compiling current global population and urban land use data layers; 2) developing projections of future population and urban area growth; 3) defining suitable locations for future development following a logistic regression analysis explaining urban patterns around the globe; 4) allocating future urban area development and population change; 5) assessing exposure to natural hazards. We conclude that on global scale urban development is likely to strongly increase exposure to both floods and landslides. In almost all world regions urban growth during the coming decades is larger in hazard-prone areas than in non-exposed areas. This is especially prevalent for countries in Sub-Saharan Africa and South Asia. In developed countries growth rates are much lower and show far less variation between exposed and non-exposed areas. In our presentation we will discuss the functioning of the model, its calibration and validation and the most interesting outcomes. We will briefly reflect on its usefulness for policymakers, suggesting that the model is best applied in fast developing regions where model-based risk assessments were hitherto impossible because of a lack of data.