Bridging the gap between general circulation model (GCM) output and biological microenvironments

Of the various research themes that have assumed prominence in recent years, potentially the most far-reaching has been the impact of climate change upon terrestrial biota. Anthropogenic greenhouse gas emissions to the Earth's atmosphere are predicted to perturb future meteorological conditions with global mean surface temperature increasing by c. 0.3 degrees C/decade (Houghton et al. 1992). Economists, politicians and environmentalists are concerned about the consequences of these changes for crop yields, species distributions and disease epidemics. To address this issue, the international research community has recently focused resources towards a better understanding of how climate affects the form and function of the terrestrial biosphere. A range of field manipulation, mesocosm and laboratory experiments are testing the ecophysiological effects of elevated temperatures across a gradient of habitats and ecotypes (NERC 1993). However, one potential weakness of this research is its failure to acknowledge that global change predictions are based entirely upon macroclimate models. These use differential equations to describe climate on a grid-scale of hundreds of kilometres (horizontal axis) by tens of metres (vertical axis) and are effective at a height considerably above the boundary layer at which climate affects biological processes (Fig. 1). A range of mediating factors confounds the straightforward interpolation from macroclimate to microclimate, with the net result that a Delta degrees C increase predicted for a grid cell of 1000 km(2) may not translate to a Delta degrees C increase for an insect sheltering beneath a rock. This paper seeks to stimulate discussion on the validity of using macrometeorological forecasts in ecophysiological research. It is suggested that biological experiments, whether field- or laboratory-based, should take account of the differences between macro- and microclimate predictions. Mesoscale models provide one promising step in this direction, but even these have only limited relevance to microclimate conditions. Instead, greater understanding of the dynamics of simulation models is required in order that the applicability of their predictions to the biosphere may be established.