JOHN A. MCGOWAN1
Natural populations are highly variable in space and time, so biogeochemical processes and events must also be. Population and community biologists have shown that internal dynamics in biological systems, such as those due to competition and predation, can cause some of the population changes observed. It is generally conceded (but seldom demonstrated) that climatic variations also influence population growth or decline. In addition, harvesting, pollution, and habitat destruction are important agents of change. It is difficult to separate the influences of these several causal factors, so ascertaining the effects of climate change on organized living systems (ecosystems) is not straightforward.
The prevailing approach to the determination of the causes of population variability and ecosystem change is essentially reductionist. There have been many small-scale, local studies of the natural history, physiology, behavior, or growth of single species. But species populations in nature are constantly interacting with a complex system of other species and to broad spectra of physical and chemical variables. Many ecologists have therefore chosen to make the simplifying assumption that different species can be aggregated into functional groupings of species, the dynamics of which adequately represent the state of the biotic system. The group dynamics actually studied, however, are usually the rates of transfer or "flux" of carbon, nitrogen, or calories between a few aggregated categories.
Many ecosystem models assume a steady state. But the systems seldom seem to be in a steady state when field studies are done, and concurrent measurements of "fluxes" involve only a limited number of groups. Ecosystems are organized on very large spatial scales, and the heterogeneity of the ocean makes it difficult to extrapolate upward from small scales of study. Furthermore, the roles of external forcings that are independent of population density, such as climate, are seldom evaluated empirically.
As a consequence, we have little good conceptual insight into how climate affects populations or ecosystems, and almost no predictive capability. But large-scale, long-term monitoring of both individual species populations and aggregated groups has been done. In general, when harvestable populations have been monitored, the relationship between population variations and climate is vague at best. Monitoring of unharvested pelagic systems, however, clearly shows the relationship between population and low-frequency climatic variations.