Variations in the earth's climate have had considerable impact on society—particularly agriculture, fisheries, water resources, and recreation—throughout recorded history. Such natural climate variability must be identified, quantified, and understood if ways are to be found to minimize its negative consequences and maximize its positive ones. In addition, human activities could significantly alter this natural variability, and indeed may already have done so. If we are to make informed decisions about our own future, it is essential that we assess the climate's sensitivity to a variety of factors, particularly on the decade-to-century time scales that are of most concern to human beings.
Many information sources, including instrumental records, visual observations, and paleoclimate data, bear witness to substantial variability in the earth's climate on time scales from years to centuries. The notion of a stationary climate on these time scales has thus become untenable. While variability in the modern climate regime is small relative to the formidable changes that characterize transitions from glacial to interglacial periods, the rate of change is often similar or even greater.
The relatively short instrumental record of climate clearly does not represent a steady background against which future variations can be gauged. Human-induced change will be difficult to assess unless the long-term natural variability of the climate system can be characterized. Natural variations with time scales of decades to centuries may well be masking anthropogenic climate changes that have already been effected, and will continue to do so. We must be able to recognize natural variability and its results if we are to make reasoned estimates as to whether a particular climate perturbation or trend is likely to have been induced by human activities, or simply represents a natural variation.
These very basic issues give rise to four critical questions concerning climate change on time scales of decades to centuries:
Can we characterize the climate system's variability on these scales, over both space and time?
Can the causes of such climate variability be isolated?
Can such climate changes be predicted?
Can changes induced by human activities be distinguished from natural variability on these time scales?
To assess climate variability on intermediate scales, one must have some idea of what it, and the forcings, look like. Several possible forms of climate variation can be seen in Figure 1. For example, climate variability may involve periodic change (Figure 1a), which is similar in nature to a daily or annual cycle, but in this case has a cycle lasting tens to hundreds of years or longer. The climate may also undergo a sudden shift (Figure 1b) from its current state to a different state, possibly one characterized by significantly colder or warmer conditions. It may show a steady warming or cooling until a stable state is reached (Figure 1c). And last, the climate may maintain what appears to be a steady state, when characterized by a specific variable such as mean annual temperature, but variations in some other measure, such as seasonal temperature, diurnal range of temperature, snow or ice coverage, or storm frequency, may indicate