Estimating Aggregate Earthquake Losses to Portfolios Using Convolution Approach

Computation of expected aggregate earthquake loss to a building portfolio is a critical part of management of earthquake risks and establishing efficient risk transfer strategies. At present, simulation based approaches are often utilized for computing aggregate losses mostly due to the flexibility they provide in representing the hazard, vulnerability characteristics and the financial aspects of the problem. However, utilizing simulation based methods are become computationally very demanding particularly when large portfolios are considered. Moreover, when the portfolios to be analyzed consist of elements that are densely lumped to small regions, the capability of stochastic simulation based approach for considering intra event shaking intensity correlations becomes challenging to utilize. Here, benefits of an alternative aggregate loss computation approach based on convolution theorem is discussed and critical aspects of its numerical implementation are addressed. The results from the application of the convolution-based approach to a sample portfolio shows that it can be used as an efficient tool in computation of aggregate earthquake loss distribution and offer particular advantages in performing sensitivity analysis.