Abstract:

aims at evaluating the impacts of climate change on pastoral resources located along the Apennines chain. To this end, random forest machine learning model was first calibrated for the present period and then applied to future conditions, as projected by HadCM3 general circulation model, in order to simulate possible spatial variation/shift of pastoral areas in two time slices (centred on 2050 and 2080) under A2 and B2 SRES scenarios. Pre-existent spatial database, namely Corine land cover map and WorldClim, were integrated and harmonised in a GIS environment in order to extract climate variables (mean seasonal precipitation, mean maximum temperature of the warmest month and minimum temperature of the coldest month) and response variables (presence/absence of pastures) to be used as model predictors. Random forest model resulted robust and coherent to simulate pastureland suitability under current climatology (classification accuracy error=19%). Accordingly, results indicated that increases in temperatures coupled with decreases in precipitation, as simulated by HadCM3 in the future, would have impacts of great concern on potential pasture distribution. In the specific, an overall decline of pasturelands suitability is predicted by the middle of the century in both A2 (–46%) and B2 (–41%) along the entire chain. However, despite alarming reductions in pastures suitability along the northern (–69% and –71% under A2 and B2 scenarios, respectively) and central Apennines (–90% under both scenarios) by the end of the century, expansions are predicted along the southern areas of the chain (+96% and +105% under A2 and B2 scenarios, respectively). This may be probably due to expansions in pastures dominated by xeric and thermophiles species, which will likely benefit from warmer and drier future conditions predicted in the southern zone of the chain by the HadCM3. Hence, the expected climate, coupled with an increasing abandonment of the traditional grazing practices, will likely threat grassland biodiversity as well as pastoral potential distribution currently dominating the Apennines chain.

Keywords: