|
Abstract:
|
There is plenty of evidence about climate change from the current period, but the only tools available for projecting the future are climate models. Information about the future climate, particularly regarding the likelihood of extreme weather, is important. Subsequently, it is important to have good information about the future climate. Therefore, it is vital to ascertain how realistic and informative climate models, or ensembles of climate models, are, especially as regards the implications for extreme weather. Climate evaluation studies thus pose challenges, though the tools used have tended to be fairly simple (e.g., comparing climatic means and second-order statistics). Challenges include: (i) the spatio-temporal context for a climate model, (ii) use of observations from the current climate to initialize future runs, (iii) the need to aggregated over long periods (e.g., decades), where current, accurate observations are only available for a relatively short period, (iv) the scale of climate model output generally differs from the observed series. This talk will explore possible solutions to some of these challenges.
|
