Heating and cooling represent a significant part of the electricity demand and this part may increase with a shift from fossil fueled systems to electrical ones for heating or with an increased demand for cooling due to the ongoing warming. Therefore, temperature is and will remain an important driver of electricity demand.
Anticipating future evolutions in the electric sector not only necessitates to project possible changes in base load, but also, and probably more importantly, in peak load. In France, peak load is linked to intense cold waves, and the last record occurred in February 2012. Heat waves induce an increase in electricity demand, but not in the same proportion. However, they may lead to tensions due to restricted cooling capacities for the thermal power plants.
In this context, information about the possible occurrence of past very cold or very warm events like the 1963 cold wave or the 2003 heat wave in the coming decades is of primary interest. A methodology for building very large ensembles of temperature trajectories covering recent past and near future has been developed and tested for the 32-station weighted average temperature in France, which is the indicator for electricity demand. The methodology is based on the combination of observed seasonalities and climate model trends with the use of a stochastic generator. Then, such a large ensemble allows an estimation of the frequency of the most extreme heat and cold waves and of the possible changes between two different time periods.
After a brief description of the production of the temperature trajectories ensemble, the presentation will focus on the analysis of the frequency of the most extreme events in different periods, in order to estimate possible evolutions in the next decades.