The spatial and temporal scales of information from GCMs, from which climate scenarios have generally been produced, have not been ideal from an impacts point of view. The desire for information on climate change regarding changes in variability as well as changes in mean conditions and for information at high spatial resolutions has been consistent over a number of years (Smith and Tirpak, 1989).

The scale at which information can appropriately be taken from relatively coarse-scale GCMs has also been debated. For example, many climate scenarios constructed from GCM outputs have taken information from individual GCM grid boxes, whereas most climate modellers do not consider the outputs from their simulation experiments to be valid on a single grid box scale and usually examine the regional results from GCMs over a cluster of grid boxes (see Chapter 10, Section 10.3). Thus, the scale of information taken from coarse resolution GCMs for scenario development often exceeds the reasonable resolution of accuracy of the models themselves.

In this section we assess methods of incorporating high resolution information into climate scenarios. The issue of spatial and temporal scale embodies an important type of uncertainty in climate scenario development (see Section 13.5.1.5).

Since spatial and temporal scales in atmospheric phenomena are often related, approaches for increasing spatial resolution can also be expected to improve information at high-frequency temporal scales (e.g., Mearns et al., 1997; Semenov and Barrow, 1997; Wang et al., 1999; see also Chapter 10).

IPCC Homepage