We can also compare predictions of future warming (as modeled with IPCC-I projections of increases in greenhouse gases in the next 100 years) with the "maximum" warming indicated in Figure 1. The IPCC-I business-as-usual (BaU) scenario for "realized" temperature is plotted on Figure 1 (open circle). The "realized" temperature for their ''B" scenario, representing a substantial reduction in greenhouse-gas emissions, is also plotted for comparison. If the IPCC-I model predictions are taken to be high by a factor of 2, as indicated by their overprediction of the past climate, then the predicted rates of increase in global temperature will be no greater than the observed increases over the recent geologic past. This further implies a climate sensitivity (increase in mean temperature for a doubling of greenhouse gases) closer to 1.25 than to 2.5, the value chosen by IPCC-I.7
But suppose that global temperature does increase at a rate greater than that observed in the past? What effect might that have on plants and ecosystems? It is appropriate in this context to compare the variations in temperature to which plants are actually exposed and the microclimatic temperature requirements of plants.
Clearly, long-term trends in climate have been reflected in ecosystem function and species ranges, and will continue to be in the future. It is important to note, however, that insofar as the physiological response of a particular plant is climatically controlled (whether it lives or dies, flowers, reproduces, etc.), it is the local energy balance and microclimate that are controlling. Indeed, the global energy balance and climate are controlled in the first instance by the sum of local climates. Thus, it is not that global climate change influences plants and ecosystems, but the other way around: Local climates, summed over the globe, constitute global climate.
There are, of course, the questions of how, and how much, local climates are modified by changes in the global energy fluxes produced by changes in such exogenous variables as the carbon dioxide concentration of the atmosphere. It is clearly reasonable to expect the predicted changes in the surface radiation balance to result in higher surface temperatures. What is not so clear is what the time and space scales of such increases might be.
It is my contention that any secular increase in the "average" temperature of a particular location would be produced by the slow change in the energy balance concomitant with an enhanced greenhouse effect, and would appear as a superimposition on the existing climate and microclimate. It is important, therefore, to investigate the actual climatic fluctuations to which plants are exposed, in both time and space, for comparison with those likely to result from a slow secular warming. It should be noted that a warming rate of 3°C per hundred years (the BaU "realized" temperature scenario) translates to a rate of 0.03°C per year for the globe, with perhaps 2 to 3 times that rate for high latitudes. But there is little empirical evidence that such secular temperature increases will be accompanied by an increase in the variability of climatic or microclimatic temperatures (IPCC-I; NAS, 1992). An increase in the mean temperature of, say, 0.03°C might thus be expected to be accompanied by a similar increase in the mean maximum temperature. Information on the specific microclimatic requirements of various species of plants provides a perspective for assessing the significance of the changes that are expected by, for example, IPCC-I.
Information on plant species' microclimatic requirements is available in the published literature, although there is less than one might expect, given the importance of environmental conditions to the growth and survival of plants. An example of this information is presented in Figure 3, which
I have assumed that the warming trend in the last century is the result of increasing atmospheric carbon dioxide and other greenhouse gases. This seems the most reasonable assumption in the absence of contrary theory. Balling (1992) states that the 95 percent confidence interval of the trend line is between 0.37°C and 0.53°C.