Climate projection for the Mediterranean and implications for the forest fire risk (with a stress on Sardinia)

Future climate projections are commonly based on Global Climate Models (GCMs). As the GCMs do not realistically represent the site specific climate, the GCM output must be downscaled (in space and time) to prepare weather/climate data for the climate change (CC) impact studies. One of the available downscaling methods is based on a weather generator (WG), whose parameters are derived from the present climate observational weather series and then modified according to the GCM based CC scenario. To account for uncertainties in the future climate projection, we employ multiple scenarios which represent uncertainties related to emission scenarios, climate sensitivity and choice of GCM. The CC scenarios are derived using the pattern scaling method, in which the scenario for a specific future is defined as a product of change in global mean temperature (estimated by simple climate model MAGICC for several combinations of climate sensitivity factor and emission scenario) and standardised scenario (= changes related to 1K rise in global temperature, which are derived from individual GCM simulations). The presentation will consist of two parts: (1) Construction of multi-GCM CC scenarios for the Mediterranean (with a stress on Sardinia). We will (i) present maps showing the probabilistic (based on all GCMs included in IPCC-AR4 dataset) projection of temperature, precipitation and PDSI index, and (ii) discuss options for choosing a representative subset of GCMs from all available GCMs. (iii) In addition to commonly used climatic parameters (changes in monthly means of temperature and precipitation), the scenarios will include changes of climatic characteristics, which affect high frequency variability, e.g. changes in probability of wet day occurrence and variability of daily values. (iv) The M&Rfi weather generator, whose parameters are modified according to the above scenarios, will be used to estimate changes in -advanced? climatic characteristics with a stress on extremes (including changes in extreme temperature and precipitation, and changes in lengths of wet/dry/hot/cold spells). (2) Assessment of possible impacts of climate change on wildland fire risk. The main weather parameters affecting the wildland fire occurrence and propagation (air temperature, air relative humidity, wind speed, and precipitations) will be used to calculate the Fire Weather Index (FWI) on annual and seasonal basis. Finally, the possible implications of the projected changes in climate on the forest fire risk will be analyzed at local and regional scale. Acknowledgements: The contribution brings outputs of the PRASCE project (project IAA300420806 funded by the Grant Agency of ASCR), CNR-ASCR bilateral project, and the Proterina C project (co-funded by the European Regional Development Fund (ERDF) under the Italy-France Maritime).