Working Group I discusses methodologies for estimating changes in variability from the results of GCMs (see Sections 9.3.2 and 13.4.2). Despite certain shortcomings, GCMs can provide estimates of trends in climatic variability. Using extreme events from historical data as analogs also is useful.
The frequency of extreme events is likely to change as mean values shift, even without changes in variability. Chapter 3 reviews potential changes in different climatic elements (see Table 3-10).
From the instrumental record, some regional changes in extremes have been identified, although it is difficult to say whether they are related to GHG-induced climate change. For example, there has been a recent increase in heavy and extreme precipitation in the mid to high-latitude countries of the northern hemisphere, and in several regions of east Asia a decrease in the frequency of temperature extremes together with heavy and extreme precipitation have been observed (see TAR WGI Chapter 2)
Many models that validate well for present climate conditions may not respond realistically to future climatic conditions and subsequent changes in extreme events. For some sectoral impacts, however, methods for evaluating a system's response to changes in variability change are improving. One example is estimation of changes in flooding by using 10-year return periods given by transient GCMs and applied to a watershed model (Takahashi et al., 1998).
Other reports in this collection