Snow Load on Structures under Changing Climate Conditions

The effect of climate change on climatic actions could significantly affect, in the mid-term future, the design of new structures actions as well as the reliability of existing ones designed according the provisions of current codes, which are derived from past observations under the assumption of stationary climate conditions. An original technique for snow loads definition is proposed, based on a non-stationary model for extreme values, able to take into account information provided by the outcomes of climate models. Temporal trends are assessed directly on location and scale parameters of Extreme Values Type I distribution, considering moving time windows of thirty years shifted ten years by ten years. The analysis are performed for the Italian Mediterranean region, suitably elaborating observed data series of daily temperatures and precipitation and climate projections of the same variables provided by Regional Climate Models for different greenhouse gas emission scenarios. The results are then compared in terms of characteristic values of ground snow load (0.98 quantile of annual extremes), that serves as basis for structural design, assessing its variability with time. Finally, an advanced Bayesian method for the definition of trend parameters is presented. This method combines climate change information present in climate models with observed temporal trends and will lead to a more trustable definition of future trend of ground snow loads.