and 1983-1984 Californian El Niños are unambiguous and large.
The sampling and data-processing methods of the two groups of researchers were different, as were their choices of climatic indicators, but they came to the same conclusion: The largest changes in plankton abundance are clearly associated with climatic changes. Apparently smaller climatic variations have lesser or no detectable effects. We are now in the position of asking how these large variations work to change the carrying capacities of entire ecosystems. What structural changes take place? How is biogeochemical cycling affected?
It is difficult to see how short-term, ad hoc studies of the flux of energy or an element can be extrapolated upward and outward in space and time to get answers. Many such studies are driven by concepts rather than empirically derived inverse models. Realistic models of reasonably complete systems must consist of partial differentials and multiple variables which themselves are functions of other extrinsic variables (Fisher, 1988). If these are to mimic nature, close coupling between parts with vastly different intrinsic rates of natural increase (i.e., genetically programmed generation times) will be needed, including numerous feedback loops that allow external forcings to cause switches between states.
There are two cases in which time series have apparently resolved the question of anthropogenic versus natural variations. Radach et al. (1990) showed that changes in phytoplankton production and community structure were very likely due to the observed enhanced nutrient input to the German Bight from rivers, chiefly the Elbe: "Elbe freshwater discharges and nitrate/nitrite concentrations are running fairly parallel to each other." Conversi and McGowan (1994) have managed to separate low-frequency natural variability from anthropogenic, and in their study the percentage of the variance due to each may be estimated. Neither of these analyses would have been possible without time-series monitoring. Sugihara and May's 1990 analysis of Allen's magnificent time series shows that even very noisy time series have predictive value, and that given the proper set of time-series data there is a real chance of separating intrinsic, biologically driven variations from extrinsic environmental ones.
Because multiple, non-linear interactions exist in complex biogeochemical marine ecosystems, they never seem to be in steady state when we measure them. Reductionist, experimental micro-research may therefore never provide sufficient understanding of their structure and function on the critical frequency scales of decades to centuries, much less answer the questions of what keeps ecosystems stable or changes them. Even the few existing time series of observations, crude though they are, have already provided us with valuable insight into the behavior of some of these systems and how climate affects them. More observational evidence is badly needed if we are to understand the natural biogeochemical variability of the oceans and assess the consequences of anthropogenic climate change. Natural variability is the reference against which such change must be measured. There may still be time to acquire the needed evidence.
ROBERT R. DICKSON
U.K. Ministry of Agriculture, Fisheries, and Food
I am very glad to have the chance to review Dr. McGowan's paper, for many reasons. High among them is my gratification that the National Academy of Sciences has been so broad-minded as to include plankton and biology in this decadal symposium. The changes in the planktonic ecosystem are arguably the most important changes that this workshop addresses, because they have the greatest socioeconomic impact.
Although the maze of interactions in the planktonic ecosystem that Dr. McGowan showed us is absolutely correct, he was a bit more pessimistic than I would be about their use in understanding what is going on, and also a bit more disparaging about the contributions that can be made by detailed biological studies. We agree entirely on the efficiency and cost-effectiveness of planktonic monitoring, however. I have been soliciting funding for the continuous plankton recorder for the last five years, and Dr. McGowan has been doing the same for many more years for the CalCOFI data set. This issue is of key importance to this symposium because of the need to assess whether an observed change in the plankton is due to climate or to man, by which we usually mean anthropogenic nutrient inputs or eutrophication.
Perhaps you think I exaggerate the importance of answering this question. Signatories to the Ministerial North Sea Conference—and that is everyone around the North Sea—are obliged to reduce nutrient use in agriculture in their countries by 50 percent if their coastal waters are or are likely to be subject to eutrophication. One of the signs of eutrophication adopted by the Paris Convention is an increased prevalence of dinoflagellate blooms in shelf water.