Clustering high dimensional meteorological scenarios: Results and performance index

The Réseau de Transport d'Electricité (RTE) is the French main electricity network operational manager and dedicates large number of resources efforts towards understanding climate time series data for purpose energy optimization. A key challenge at core being able to detect common patterns between temperatures series, choose representative scenarios simulations, which in turn can be used We addressed this using provided by RTE, comprised 200 different possible on a grid geographical locations France. first show that choice distance clustering has strong impact meaning results. Depending type used, either spatial or temporal prevail. Later we discuss difficulty fine-tuning distances with dimension reduction procedure propose methodology based carefully designed index.