Spatially-resolved emulations of droughts and fire weather using MESMER-X

Climate extremes are among the most impactful consequences of climate change. Moreover, droughts and fires may also hinder envisioned solutions to mitigate climate change, by reducing the efficiency of bio-energies with carbon capture storage and afforestation. Though, investigating such issues would benefit from a tool allowing fast computation of spatial climate extremes, for coupling to other models and exploration of scenarios. Here, we present an approach for emulating such extremes that are based on extensions of the spatially-resolved climate model emulator MESMER-X. In particular, we consider four annual indicators of the Canadian Fire Weather Index and the annual mean soil moisture derived from the Climate Model Intercomparison Project phase 6 for training and emulation. To emulate these indicators, we consider extensions to the framework that include the Gaussian and the Poisson distributions, non-linear evolutions of the parameters of the distribution and lagged effects. We show that the emulator reproduces the trajectories and the statistics of these annual indicators for fire weather and droughts accurately. By doing so, we show that the theoretical framework of MESMER-X can be applied for a large number of annual indicators of climate extremes.