Reliability-based design of structures under non-stationary climate conditions

Abstract

At present most design codes specify climate-induced loads in terms of extreme percentiles and return periods, which are estimated assuming stationary climate conditions. Further, the load and resistance factors are treated as constants and calibrated with respect to a fixed annual or lifetime reliability target. The accumulating evidence about the rapid pace of climate change and its adverse impact on infrastructure safety has motivated the development of methods for incorporating non-stationary climate conditions in the reliability-based design of structures. This paper presents a non-homogeneous Poisson process model of non-stationary changes in climate variables and illustrates its impact on various reliability measures. The paper shows that the traditional measures, e.g., return period, extreme percentiles, and the annual probability of failure, would become time-dependent variables under the changing climate. Therefore, existing design codes would need to be revised to accommodate such non-stationary changes.