Structure over Large Distances and Long Times

J. Yoder, A. Bakun and D. McGillicuddy

Section II. State of the Field

An emerging priority within the field of Biological Oceanography are the scientific questions involving biological processes and distributions at large scales - i.e. those ranging from meso- (10-100 km) to basin- to global scales. Furthermore, long time scales (e.g. interannual to paleo) are often closely linked to basin and global scales, and thus this range of time and space scales should be considered together. Current thinking regarding large spatial scale/long time scale questions was stimulated by early efforts to summarize biological rates and distributions on large scales (e.g. Koblenz-Miske primary production maps); by large-scale ocean models and satellite observations; by ideas and observations from those fisheries scientists who study long-term changes in fisheries populations; and by the observations, models, ideas and criticisms of geochemists. Current interests range from fish populations to biogeochemical cycles.

Recent analyses show there is growing evidence of globally-synchronized, interdecadal-scale variability in fish and other populations that suggest some type of very large-scale external forcing, most probably through global-scale climatic effects. About two decades ago, marine scientists were bemused and perplexed by the evidence from the Continuous Plankton Recorder Survey that the zooplankton biomass over a very large region of the northeastern Atlantic had apparently undergone a drastic multi-decadal decline. Furthermore, such extreme long time-scale variability may not be at all extraordinary, but rather a quite normal condition for marine ecosystems. For example, zooplankton biomass in the California Current has decreased by 70% since the 1950s, along with corresponding drastic declines in certain seabird species. At the same time, there seems to have been a doubling of biomass of zooplankton and of pelagic fish and squid in the subarctic North in the 1980s compared to the late 1950s to early 1960s period. Notably, in several parts of the Pacific, there are indications of important changes beginning in the early to mid-1970s. South of the equator, there seems to have been a very sharp 60%-70% decline in zooplankton biomass off Peru in the mid-1970s, coinciding with the collapse of the anchoveta, and depth-integrated chlorophyll in the subtropical North Pacific has doubled since the mid-1970s.

Populations of small coastal pelagic fishes such as anchoveta (anchovies, sardines, sardinellas, etc.), which support many of the largest fisheries in the world, are particularly prone to wide population swings and fishery collapse. Indeed, evidence derived from deposits of fish scales in sediments has indicated extreme fluctuations in population sizes of this type of fishes long before the advent of large scale fishing. A second very remarkable aspect is that, on these time-scales, the variations appear to exhibit a large degree of global synchrony, i.e. , when one looks very broadly at the sardine landings from the three regions of the Pacific Ocean that have supported very large sardine fisheries, a pattern of oceanwide synchrony emerges that is suggestive of oceanwide synchrony in population fluctuations. Moreover, two of the other largest coastal pelagic fish populations of the world, the Peruvian anchoveta and the South African sardine appear to rise and fall in the "gaps" of the same pattern, i.e., directly out of phase with the Pacific sardines. This pattern of increases in large important fish stocks, followed by declines after the mid-1980s, has been remarkably widespread and consistent in a number of regional ecosystems of the world's oceans. Moreover, many of the stocks that have not been in phase with the pattern have been quite directly out of phase (i.e., "mirror images"), or at least have matched some of the main inflection points. These phenomena, now evident in more and more time series of fish stock fluctuations covering the past two and one-half decades, await explanation.

More than two decades ago, interdisciplinary studies of shelf ecosystems demonstrated close coupling of physical forcing with new production, primary production and other lower trophic level processes at spatial scales of 1 to 100 kms (mesoscale) and times scales of days to weeks. In particular, these studies showed that phytoplankton productivity and other lower trophic level processes in tropical to mid-latitude shelf ecosystems responded rapidly and dramatically to wind-driven upwelling and to other physical processes affecting frequency and magnitude of nutrient fluxes to the euphotic zone. These studies also demonstrated that in many cases, plankton herbivores could not consume all of the new production, and some (much) was advected from the shelf to the open sea or settled from the water column to shelf and slope sediments. These interdisciplinary studies led to new ideas regarding the role of physical processes, such as the frequency of upwelling events, as important controllers of ecosystem trophic structure. Simply stated, regions with relatively frequent intervals between upwelling events were characterized by herbivores (e.g. sardines) which could find plumes of high productivity, or by high loss rates of water column productivity to sediments, or both. Regions with relatively low frequency of nutrient supply (or with seasonal light limitation, another relatively low frequency driver of autotrophic processes) were characterized by relatively high abundance of macrozooplankton which are apparently favored under conditions that allowed populations of these multi-stage organisms to develop at favorable food conditions. Although there are probably as many exceptions as rules for shelf ecosystems, the concept of physical forcing controlling marine ecosystem structure became an important tenet of biological oceanography, at least for mesoscale patterns and for daily to seasonal time scales.

Ideas evolving from physical and biological coupling at mesoscale and at daily to seasonal time scales may extend to some extent to larger spatial and longer time scales. Testing of hypotheses related to physical forcing of lower trophic level processes at these scales, however, is quite difficult because of sampling problems. Satellite observations of vector winds, sea level, sea surface temperature and of various data products derived from ocean color observations are providing one important new source of data which can be used to address some questions on large-scale coupling of physical and biological processes. For example, the 1982/83 El Nino had dramatic effects in ocean color imagery from many regions of the eastern Pacific including the Galapagos, Baja California, California Current, and the coastal waters off Peru and Chile. The 1997/1998 El Nino is presently under study by investigators using a full suite of new satellite systems, and results will be forthcoming in the next several months.

A. Satellite Observing Systems

With the launch of the ADEOS satellite in August, 1996, we began the era of concurrent global satellite observations of ocean currents and eddies (from altimeters), vector winds (from scatterometers), sea surface temperature (from thermal IR radiometers) and ocean color and surface chlorophyll (from satellite ocean color scanners). Firm plans are in place by NASA, NOAA and other international space agencies to sustain these measurement sets well into the next millennium. In addition, computer hardware, software and networks are now sufficiently advanced that global satellite data sets can be rapidly and efficiently processed, distributed and archived. Mapped, binned and contoured data products are accessible via the Worldwide Web, and they can be scientifically analyzed using desktop computers available now to most oceanographers, including graduate students. Thus, beginning in the mid 1990s, the hardware and software are available for biological oceanographers to effectively incorporate satellite observations in studies and models of biological distributions and productivity at meso- to global-scales.

Satellite-derived data products probably of most interest to biological oceanographers include vector winds, phytoplankton chlorophyll, detritus and (in combination with other information) primary production. The two major strengths of satellite data products are their synopticity, i.e. capability to rapidly (minutes-days) acquire measurements over large areas of the ocean and secondly, the potential to build multi-year time series. The main weaknesses of satellite ocean color measurements is that they provide estimates of only a few very crude indices (e.g. phytoplankton chlorophyll) related to biological processes; they are not very useful (to date) for studies of estuaries and other relatively small bodies of water; they cannot resolve vertical distributions; and measurements cannot be made through clouds.

B. Models

Interconnection between the physical, biological and chemical aspects of oceanographic ecology stems from two distinct sources: (1) redistribution of water column constituents by currents through advective processes, and (2) mediation of biological and chemical reactions by environmental fluctuations. The duality of this linkage poses a difficult challenge to understanding observed distributions as differentiation between the sources of variability requires accurate assessment of property distributions in space and time in addition to detailed knowledge of the processes by which ambient conditions control the rates of biological and chemical transformations. Logistical constraints make a purely observational approach to such problems impractical. However, the coupling of models with observations offers an alternative which provides a context for synthesis of sparse data with articulations of fundamental principles assumed to govern functionality of the system. This combination of observations with models via data assimilation facilitates the construction of optimal estimates of real oceanic fields that are simply not accessible by any other means. These space-time continuous representations of the ocean are ideally suited to the investigation of intermittent processes as they provide a framework for detailed term by term diagnosis of complex behavior.

Ecological models developed for these purposes have taken many forms. Zero- and one-dimensional frameworks [Note: "Zero-dimensional" models refer to those which do not represent horizontal or vertical structure. Mixed layer formulations which accommodate changes in layer depth have been referred to as "half-dimensional."] have been used extensively for simulation at specific locations throughout the world ocean. Regional models provide explicit horizontal resolution but require specification of boundary conditions which are often difficult to obtain. Basin- to global-scale models can in some ways minimize boundary condition problems at the expense of making contact with local and regional data sets more difficult. Each of these types of models has utility in ecological problems, and all will probably play a role in future progress. The nesting of models provides an exciting challenge for integrating these various approaches.

A. Marine Populations

The apparent strength and potential adverse global impact of the 1997/98 El Nino, the extent of publicity associated with it, and the significance of the El Nino signal for interannual and large-scale processes are reasons for making the effects of El Nino a high priority for future study. Furthermore, the El Nino-Southern Oscillation (ENSO) system of the Pacific must be in one way or the other involved with the fluctuations in marine populations discussed in Section II, either as the driving engine or at least as an integral component of the mechanism. Of all the sources of short-term global climatic variability, ENSO is by far the strongest, completely dominating any other mode of variability in the Pacific Ocean. Moreover, because the Pacific is so large, its effect on the earth's climate system is global. El Nino episodes generally last only about a year but there are decadal-scale modulations in intensity and frequency . .In particular, the period from the early to mid-1970s to the mid-1980s which has seemed to synchronize long-period fish population variations occurring in many different regions, was a period of prolonged long-term drop in the "low-passed" Southern Oscillation Index, i.e., a sort of decadal-scale analog of a standard El Nino episode. Thus, the coupled ocean-atmosphere system of the equatorial Pacific appears to have undergone a long-term relaxation during this particular period. Correspondingly, judging from time series of indicators of atmosphere-ocean mechanical energy and momentum fluxes many peripheral regions of the world's oceans, including the eastern boundary upwelling systems as well as the subarctic Pacific (dominated by Aleutian low/N.Pacific Oscillation), appear to have undergone long-term intensification during the same period.

It seems unlikely that a set of separate autonomous ecosystems, each dominated by its own internal chaotic dynamics, could somehow "self-organize" themselves to generate mutual synchronous population variability on a global scale. Thus, if these populations are indeed varying synchronously, the conclusion seems to be that they must not be functioning primarily as independent chaotic systems, at least on the time scales of the synchronized variability. Moreover, since biological models representing anything but the simplest of marine trophic interactions are characterized by chaotic behavior, the implication of global synchrony would seem to be that the biological dynamics involved must be very simple: i.e., a rather direct effect of the external physical forcing acting either on the organisms themselves at some sensitive life stage in their complex life cycles, or at most directly on a primary food source. It must not be, for example, an effect working its way through a rather complex planktonic food web. That is, if the evident synchrony is not merely fortuitous, the scientific problem that we face would appear to be much simpler than it might very well have been. In such a case, there would appear to be a realistic hope of success in gaining on a reasonable time frame a real understanding of, and resulting rudimentary predictive capability for, the most prominent factors determining reproductive success and resulting population dynamics. This reasoning might provide a basis for outlining a reasonable "model" for an interdisciplinary process-oriented research efforts: i.e., the "probabilities" that: (1) the global climatic signal that may have acted to synchronize fish population variability in recent decades is transferred locally through the sea surface skin and (2) the physical transfer includes the mechanical factors, energy and momentum (which should leave patterned "signatures" of important physical ecosystem structures that should be evident in satellite imagery) leads obviously to a research framework, based on the several special ocean-oriented satellites expected soon to be in orbit simultaneously. This framework would lend itself to investigating, using a wide range of biological oceanographic and other disciplinary specialties, the biological consequences of variable patterns of important physical ecosystem structures.

There are now available some very interesting and promising tools that might be conveniently and economically employed. Fish eggs may be sampled, identified, and counted in a recently developed underway "egg pump" sampling system. Adult nektonic behavior with respect to ocean features can be readily studied during the sonic biomass surveys that are becoming a standard fish stock assessment methodology around the world. Recent nutrition and growth progress might be gauged by biochemical analysis of the tissues of a modest representative sample of larval or early juvenile fish. When satellite images, etc., may indicate short-term variability in hypothetically important ocean features, one may later sample juvenile fish and count daily otolith increments to determine birthdate frequencies of surviving individuals (i.e., to develop a time history of variation in early life survival) and daily growth histories.

Satellite images and within-season birthdate and growth rate information constitute quite short time-scale descriptions, while the time-scales while the time scales that appear to be most involved in radical population biomass variability are the multi-year and inter-decadal time scales. The key in such a case is to manage to "telescope" the findings from the studies of short-time scale variability that can be encompassed in research projects to the much longer time-scales which may be of greatest societal interest. For example, the active agent in the atmosphere-ocean transfer of mechanical energy and momentum is the action of the wind on the sea surface. Moreover, wind intensity and direction is largely a linear function of the atmospheric pressure distribution which serves therefore as a "stream function" for linearly averaging wind data to extend it in an undistorted manner from shorter to longer time scales (either as a simple mean or as a more complex parameterization of the structure of shorter-term variability). Consequently, once one defines on short operational time-scales (e.g., using satellite images and at-sea studies as suggested above) the relative favorability for a given species of observed variations of a (hypothetically) biologically-crucial physical ecosystem "structure" and the corresponding variations in large-scale wind pattern that control them, one can then use the much longer available record of large-scale climatic data to hindcast the descriptions over previous decades for which historical data on population variability may be available, and thereby attempt to verify the findings on the interdecadal scale.

B. New Opportunities and Reasons too Study Mesoscale and Intermittent Processes

Some of the most exciting opportunities for progress in our understanding of oceanic ecology may arise from improved abilities to sample intermittent processes. Emerging technologies in towed and autonomous vehicles, in addition to moored instrumentation, are beginning to facilitate biological and chemical measurements on the same space and time scales at which physical properties can be measured. Early results from these novel systems have revealed an unprecedented level of detail in property distributions, which appear to be tightly coupled to ephemeral mesoscale (and finer) physical structures. The following are some specific examples: (1) spatial survey data from the Video Plankton Recorder has shown that the abundance of several zooplankton species is highly organized with respect to hydrographic characteristics of the water column; (2) high resolution time series using moored in-situ chemical analyzers on the Bermuda Testbed mooring has revealed temporal variations in nitrate concentration which implicate mesoscale eddies in nutrient supply to the euphotic zone. These relationships among properties could never have been resolved with traditional oceanographic methods.

Although sampling systems such as these are opening a new realm of observational capabilities, the problem of synopticity is likely to remain. Our ability to sample the ocean in-situ is ultimately limited by the speed at which measurement systems can move through the water. Therefore it is unlikely that we will ever be able to achieve perfectly synoptic realizations of the desired water column properties at all the relevant scales. However, the combination of observations with models provides a framework in which the impact of sparse measurements can be maximized. Moreover, continued advances in interdisciplinary ocean prediction and data assimilation will facilitate the development of techniques to optimize sampling strategy. Real time nowcasts and forecasts of oceanic fields will provide the information necessary to deploy these novel sensors "where the action is" such that intermittent processes can be observed directly. Synergistic conjoining of remotely sensed and in situ data with numerical models offers a powerful approach to ocean ecology which is likely to provide new insights and detailed process understanding, perhaps not achievable by any other means. Using satellite data to support in situ observations and numerical models for regional to global studies of upper ocean biological processes has been much discussed during the past decade, but the lack of satellite ocean color and other key satellite data sets has delayed implementation of this research strategy, except in a few test cases. With new satellite systems in place, and with increased computer capability for modeling and data assimilation, sophisticated analyses that combine multiple sources and types of data and models are now possible leading to new observational strategies (real-time applications), as well as better retrospective analyses. These new observational techniques and strategies could be applied to many different scientific questions, but the highest payoff could come to those studying coupling of mesoscale physical and biological processes and the seasonal to interannual time scale.

Oceanographers have struggled for almost two decades with the question of nutrient supply to the oligotrophic waters in the main subtropical gyres of the open ocean. Geochemical estimates of new production far exceed that which can be sustained by traditional mechanisms of nutrient supply (convective mixing and turbulent diffusion). This remains a central question in biological oceanography, and is mentioned in the NSF chemical oceanography planning group report as well. There is mounting evidence that mesoscale processes may be at least part of the answer (modeling studies, moored observations, satellite data, in situ surveys, etc). Coordinated research strategies composed of all the elements just listed, in addition to some deliberate tracer release experiments, could perhaps resolve this issue in the coming decade.

C. Ecological Significance of Atmosphere/Ocean Chemical Exchanges at Regional to Basin Scales.

Assuming appropriate coordination between and among federal agencies, the timing is excellent to study relations between upper ocean biological processes and the atmospheric transport of important trace and major plant nutrients. Furthermore, upper ocean biological processes may act as important sources and sinks of interesting compounds to/from the lower troposphere. For example, Equatorial Pacific, North Pacific and Southern Ocean productivity is probably limited by the trace nutrient iron. Recent studies (e.g. Ironex) demonstrated that ecosystem productivity is enhanced, and ecosystem structure changed, following iron enrichment of surface waters in the Equatorial Pacific. Recent studies also indicate the importance of iron as a control on the rate of nitrogen fixation, and hence productivity, in oligotrophic ocean regions. One of the major sources of iron is iron oxide in dust carried on winds from African, S. American, and Asian deserts. The atmosphere may also be an important conduit of major plant nutrients from the land to the ocean. Recent studies show that atmospheric nitrogen deposition ranges from 0.056 to 0.15 Gt of nitrogen annually, which is equivalent to the nitrogen flux to the ocean from rivers. Most of this nitrogen originates from terrestrial sources (e.g., from dust and other aerosols), although human activities (e.g., combustion) may have increased this N source to the ocean by as much as 0.12 Gt per year. These results imply that burning and other human activities could be contributing to an atmospheric N flux potentially resulting in a 10-20% increase in overall ocean primary production, and this N flux may increase with increasing industrialization. Ocean biological processes are potentially an important source of interesting compounds to the lower troposphere, including DMS and methyl bromide. The latter is an important chemical affecting the fate of tropospheric ozone.

Return to OEUVRE Topics Page

National Science Foundation