The seismic risks to which populations are exposed should be estimated reliably for mitigation and preparation of response to disastrous earthquakes. Three parameters need to be known: Population numbers, properties of the built environment, and the seismic hazard. If we focus on large cities, we can say that at least one of these is known satisfactorily, namely the population, but not the other two. In the developing world, the numbers of buildings in a city are known only approximately, their distribution into building types (resistance to shaking) has to be assumed, and the distribution of types throughout the city is unknown. Recent verification of the world seismic hazard map has shown that it is grossly misleading: Instrumental measurements of accelerations due to six earthquakes were about three times larger, on average, than the maximum likely accelerations shown on the map; the macroseismic intensities reported for the last 60 earthquakes with M ≥ 7.5 were all significantly larger than expected, based on the hazard map (by 2.3 intensity units for the 12 deadliest earthquakes); and calculations of losses of life based on the hazard map underestimate the losses sustained in the 12 recent earthquakes with more than 1,000 fatalities by two to three orders of magnitude. This means that the seismic risk in most of the approximately 1,000 large cities at risk in the developing world is unknown. To remedy this intolerable situation, models for the built environment in cities need to be constructed, using cost-effective analyses of satellite images, and worst case scenario estimates of the losses in case of the nearest maximum credible earthquake. Copyright Springer Science+Business Media B.V. 2013
