Earthquake losses due to ground failure

Ground shaking is widely considered to be the primary cause of damage to structures, loss of life and injuries due to earthquakes. Nonetheless, there are numerous examples of earthquakes where the losses due to earthquake-induced ground failure have been significant. Whereas ground shaking causes structural and non-structural damage, with associated loss of function and income, ground failure is less likely to cause spectacular structural collapses, but is frequently the cause of major disruptions, particularly to lifelines, which can lead to prolonged loss of function and income, even for undamaged areas. Those involved in earthquake loss modelling are currently presented with three choices with respect to the incorporation of ground failure: they can choose to ignore it, assuming that any estimation of losses caused by shaking would effectively subsume the impact of these secondary hazards; they can include ground failure in a simple manner, using published approaches based upon qualitative data and a large degree of judgement; or, they can opt for a detailed site- or region-specific assessment of damage due to ground failure, with the associated time and expense. This paper presents a summary of the principal features of earthquake losses incurred in damaging earthquakes over the last 15 years. Survey data are impartially analysed, considering both ground failure and ground shaking as sources of damage, and their relative contribution to overall damage in each section of the regional infrastructure is presented. There are many other variables influencing these contributions, including the size of the earthquake, the economic status of the affected region, local geology and terrain and the building stock, which have been considered. The findings of the study are discussed from the point of view of loss modelling and which components of a model should merit the most time and resource allocation. The general assumption that ground shaking is the principal cause of damage and loss is strongly supported by the study. However, there are a number of scenarios identified where the failure to appropriately include the effects of ground failure would lead to unrealistic loss projections. Such scenarios include the assessment of building losses in small zones rather than on a regional basis, and the incorporation of lifeline damage or disruption and indirect losses into a model.