Projected future changes in mean and extreme daily precipitation in France conditioned on large-scale thermodynamic and dynamic drivers

In the context of anthropogenic climate change, France represents a critical study area where regional warming is observed to exceed the global average. Understanding the future evolution of precipitation is essential for water resource management and flood risk mitigation. In this work, we employ a unified statistical framework based on the generalized gamma distribution, selected for its flexibility in capturing the full range of the daily precipitation, including extremes, to assess future changes in daily precipitation across Metropolitan France. A central feature of our methodology is the integration of non-stationarity through physical covariates. We account for thermodynamic forcing using Sea Surface Temperature (SST) anomalies averaged over the North Atlantic and the Mediterranean Sea, which govern the atmospheric moisture content available for precipitation. Simultaneously, we incorporate dynamic drivers of regional climate variability by employing the North Atlantic Oscillation (NAO) index and the Western Mediterranean Oscillation (WeMO) index. These indices serve as proxies for the large-scale atmospheric circulation patterns that dictate storm tracks and moisture transport into Western Europe.

We started by training our statistical model using high-resolution reanalysis data to establish robust historical baseline relationships. We then derive the future projections by extrapolating these relationships using covariates from 15 CMIP6 Global Climate Models (GCMs). These specific models were selected based on their demonstrated ability to reliably simulate European hydro-climatic processes. We present near-future and long-term changes across four fundamental metrics: wet-day frequency, mean wet-day precipitation, mean of all-days precipitation, and the 20-year return level. Our results highlight a complex spatial and seasonally-dependant response, reflecting the competing influences of thermodynamic moistening and dynamic shifts.  Beyond the regional projections for France,  this unified framework offers a scalable approach for non-stationary weather generators. Such tools are increasingly vital for stochastic climate simulations and the development of localized adaptation strategies in a rapidly changing environment.