Assessing uncertainties in climate change impacts on resource potential for Europe based on projections from RCMs and GCMs

Abstract  An analysis is presented of the estimated impacts of climate change on resource potential in Europe under a wide range of model-based climate scenarios. Simple models and indices were used to assess impacts on the growing season, potential biomass, thermal suitability for the cultivation of crops, and potential energy demand for indoor cooling. Impacts were estimated for climate during the 1961–1990 baseline period (both observed and modelled) and projected during 2071–2100 based on outputs from a range of regional climate models (RCMs) driven by general circulation models (GCMs) assuming forcing by SRES emission scenarios A2 and B2, and from six atmosphere–ocean GCMs forced by a wider range of emission scenarios. Uncertainties in the projected impacts of climate change are assessed with respect to: (1) the direct climate model output vs. delta change approach, (2) differences in the driving GCMs and the RCM runs, (3) differences across a range of emission scenarios, (4) changes in long-term mean climate, and (5) changes in inter-annual climate variability. Future simulations show substantial changes in all analysed impact sectors, but with a relatively large spread of results attributable to uncertainties in future climate expressed by the different scenarios. Results included shifts of the northern limits of areas thermally suitable for the cultivation of soya bean and grain maize by several hundred kilometres, lengthening of the thermal growing season by 3–12 weeks, strong increases of potential biomass in northern Europe and slight decreases in southern Europe, and increased energy demand for cooling throughout Europe. Our results hint at systematic differences between RCM and GCM projections of temperature, though not precipitation, over Europe. The results also highlight the importance of accounting for inter-annual variability in estimating future impacts, through its affect on levels of risk. However, the results caution against the use of direct RCM outputs in impact models, due to biases in the representation of present-day climate. The delta change approach still appears to be the preferred option for most applications.