The influence of geographical resolution of urban exposure data in an earthquake loss model for Istanbul

Exposure data available to developers of earthquake loss models are often very crudely aggregated spatially, and in such cases very considerable effort can be required to refine the geographical resolution of the building stock inventory. The influence of the geographical resolution of the exposure data for the Sea of Marmara region in Turkey is explored using several different levels of spatial aggregation to estimate the losses due to a single earthquake scenario. The results show that the total damage over an urban area, expressed as a mean damage ratio (MDR), is rather insensitive to the spatial resolution of the exposure data if a sufficiently large number of ground-motion simulations are used. However, the variability of the MDR estimates does reduce as the spatial resolution becomes higher, reducing the number of simulations required, although there appears to be a law of diminishing returns in going to very high exposure data resolution. This is largely due to the inherent and irreducible spatial variability of ground motion, which suggests that if only mean MDR estimates are needed, the effort required to refine the spatial definition of exposure data is not justified.