PART II ELABORATION AND DEVELOPMENT OF SOME KEY CONCEPTS RELEVANT TO ECOSYSTEM MANAGEMENT

The fact that in harvesting ecosystems we are tapping into a flow of energy as opposed to a static pool is evident, and Golley (1972) notes that:

“The living portion of the planetary ecologicl system requires energy to maintain the thermodynamically unstable condition of life.”

and clearly ecosystems thrive where there is a high flux of energy through the environment. Thus Golley notes:

“It is well known that the energy which drives the ecosystems on the plant Earth comes from the sun. At the outer limits of our atmosphere, 1.94 gram calories of solar energy per square centimeter are received per minute. Not all of this energy reaches the surface of the earth; 35 percent is reflected and 17.5 percent is absorbed by the atmosphere and clouds. This leaves only about 47.5 percent of the solar radiation effective at the level of the biosphere.

“How is this energy used? About 30 percent is reflected as long wave radiation, 49 percent is exchanged by conduction of heat to the air through movement of wind over the surface of land and sea. The energy expended in evaporation and condensation and in heat conduction drives the circulation of the atmosphere and oceans, which truly tie the biosphere together into a one-world ecological system.“…”Atmospheric circulation acting on the oceans produces the water movements that erode coastlines. Evaporation from the oceans drives the hydrologic cycle, in which water is evaporated from the oceans, precipitated onto the land and returns to the oceans in rivers, effecting the downcutting of the continents in the process. These are the patterns of wind and water which make up the environments in which living organisms act out their existence.”

1. TEMPORAL CHANGES AND STABILITY OF FISHERIES SYSTEMS

The major source of energy that drives fisheries systems comes from fixation of carbon dioxide into organic compounds by marine plants, and as we have noted, this process tends to show peak production in rather localized regions, such as (e.g., Figure 3) in areas of high vertical mixing of water masses; i.e., areas of upwelling, in estuaries and over the continental shelves of high latitudes; also further offshore, along horizontal and vertical boundaries between ocean current systems. Highly productive areas nearer inshore in addition to estuaries are lagoons, salt marshes, seaweed beds and mangrove swamps. Dispersal of organic material in the form of organic detritus, plankton, invertebrate biomass, and (actively) fish from these centres of production goes on in space and time from the original date and location of organic synthesis. This occurs both passively by means of eddies and currents, and more actively by means of migration of living organisms: all of the time the synthesized organic material forming the bodies of plants and animals, moves “upwards” in the food web, and-at each step towards the size range of interest for commercial exploitation, loses roughly nine - tenths of its biomass and chemical energy in the form of metabolic heat, work, excreta (soluble organics), and detritus. As illustrated earlier, some of the latter two categories are cycled back into the food web by means of marine bacteria and other unicellular organisms, scavengers and detritivores (e.g, filter feeders). Despite what is believed to be the generally low productivity of clear water tropical areas, coral reefs also act as centres for production, concentration and elaboration of organic material, although as for other complex eco-systems, there is a great deal of recycling of the original synthesized material within the system.

The limits to transport processes and in particular, nutrient exchange, in stratified oceanic water between bottom water, rich in nutrient salts, and surface waters where such minerals are severely depleted, must largely account for the generally lower productivity and standing stocks in central oceanic areas, distant from the main centres of production. This is especially true for offshore waters in the tropics, where stable thermoclines may restrict high production to near-shore areas. Peak production elsewhere is usually rather restricted in time, in most regions of upwelling, in north and south temperate phytoplankton blooms, and in seaweed and eel grass beds. This means that the herbivores feeding on plant material, and the primary carnivores feeding on them in turn, tend to be often (but not always) small, short-lived (annual species) with big population blooms and crashes on a seasonal basis. This seasonal peak in production is also true in the tropics and sub-tropics in those areas where seasonal monsoons drive upwelling zones; for example, in the northern Arabian Sea and off Peru, but not so obvious in other tropical areas where a more consistent level of production continues throughout the year.

Generally speaking then, seasonal effects apply throughout the world's oceans, even in clear waters tropical areas and in the arctic, where fluctuations in water temperatures are less marked seasonally. Food web studies (cf. Figure 16), need to take these short-term fluctuations into account. Many of these seasonal changes in diet are associated with transient (short-lived or migratory) species that are frequently voracious feeders, and can significantly affect directions and quantities of energy flow in the ecosystem when present: (e.g., Figure 24). Many fisheries can also be viewed as seasonal predators and show a strong seasonality of activity, which is generally related to the temporal availability of different species to capture (e.g., Figure 17), as well as to the often associated fluctuations in market price.

Figure 16 Degree of spatial coincidence between a prey species (A) and its resident (B) and migratory (C) predators, expressed as four seasonal Venn diagrams in which the area and relative location of each circle represents their geographical extent and the degree of spatial co-occurrence. The shaded segments of the predator circles represent the percent of stomach contents (diet) consisting of the prey at that season of the year

As we have noted, one way to visualize the flow of food energy in time and space is to view the food web as a dissipation structure, where input energy from the sun, now in chemical form, moves up through the food web, at each step losing a large fraction as heat in the course of work performed by individual organisms in carrying out the various metabolic processes necessary for the maintenance of life. At the same time (and this is a function of the size scale of human harvesting) the value of the individual energy package generally increases, so that a small fraction of the original (low unit value plant material) becomes often high value fish tissue (e.g., tuna, cod-fish, etc.), which is energy-rich, harvestable in a cost-effective way, and preferred for human consumption, and usually commands a higher market price.

One of the features of natural systems that is now becoming better understood, is the degree to which fluctuations over the medium to long term in climatic factors are a normal feature of global environments and act as forcing functions in determining annual levels of recruitment, production and annual yield, acting through the level of primary production of the ecosystem.

Figure 17 Illustrating the seasonal distribution of fishing activity by species, type of boat and gear, typical of many fisheries (From Lamson and Hanson, 1984)

Figure 19 shows one of the few series of oceanographic data that began to be collected prior to the Second World War, that shows the order of magnitude of year-to-year changes associated with fisheries production. Solar, volcanic, as well as human impacts on the climate are all involved in long-term fluctuations (see e.g., Gilliland 1982). These events may be synchronized over very wide areas (e.g., the E1 Niño phenomenon, which appears associated to climatic changes on a Pacific-wide basis, and is now referred to as ENSO: the El Kino/Southern Oscillation phenomenon) -(Rasmusson, 1984).

In placing any natural system along the stability scale from close to equilibrium to wild fluctuations or periodic population crashes, the time scale of production is evidently very important: the flow of energy through an ecological system, especially close to the base of the food web is unlikely to closely approximate stability if production is seasonal or even worse, erratic. The chances of managing a system are likely to be improved if there is a relatively continuous stable production, and fortunately even in very seasonal (but regularly fluctuating) systems, higher level organisms have often evolved life history mechanisms to ensure that their production ‘evens out’ somewhat the seasonal cycle by adjusting their diet at different times of year, or during different life history stages, to seasonal changes in prey abundance, or by migration to locate other areas of peak production.

In fact, a whole new group of life history strategies come into play once evolution of multi-age group forms is widespread, and clearly, new strategies for dealing with the “feast and famine” situation caused by seasonal productions cycles had to be evolved, and as indicated above, these include active swimming, seasonal migration, separate feeding and spawning areas, ‘omnivory’ (feeding at more than one trophic level), and the whole range of associated morphological and behavioural characteristics that make up contemporary fish species.

The big evolutionary advantage of a longer life span that results when the above problems are solved, is that in evolutionary time, these species have survived instabilities in the environments by adjusting the life span to be roughly equal to the average interval between favourable spawning years, so as to survive a poor spawning year that might otherwise be disastrous for localized stocks of an annual species. The life history mechanisms shown by annual species (e.g., squid, penaeid shrimp) also tend to confirm that the environment that most species inhabit is unstable: compensatory mechanisms here include production of many eggs, several separate groups spawning at different times of the year, and a fairly wide geographical range for single stock species (e.g., the Pacific squid, Todarodes which is found throughout much of the North Pacific), thus ensuring repopulation of a stock from the centre of its range if the peripheral populations are wiped out by an unsuccessful spawning.

Figure 18 Not all species compositions in a marine community change gradually with fishing (or environmental change), but can on occasions change catastrophically. This diagram, freely adapted from Kerr (1977), shows how the abundance of a species (here a Perciform fish), can change abruptly with various types of environmental stress. Thus, an increase in effort (or a change in some environmental variable) from A to F might lead to species abundance following trajectory A-B-C with an abrupt fall to E-F, but in the opposite direction return via F-E-D-B-A (see Kerr, 1974). These types of phenomena are generally discussed under catastrophe theory (See e.g., Saunders, 1980)

The existence of multiple age groups is itself then, with multiple spawning, a life history adaption that reflects the difficulty of ensuring species survival in an unstable physical environment, as evinced by perturbations or irregularities in annual recruitment. Recent studies in the North Atlantic on long-term recruitment trends in cod and haddock, show that the very occasional big year classes play a major part here in supporting the fishery in most years; the rest of the time the current level of annual yield would not be sustainable from current production by recruitment, i.e., heavy reliance is placed for a number of years in a row on fishing survivors from previous good year classes. For these North-West Atlantic species, annual landings exceed annual recruitment about 70–80% of the time; from only 10% of the bigger year classes came 24%, 33% and 37% of the total yield for cod, mackerel and haddock, respectively (see e.g., Hennemuth and Autges, 1982). In such a fishery, it seems likely that a series of investment waves also occured, following closely behind good year classes in these fisheries. This is likely to prove typical of what happens in other industrial fisheries elsewhere (Caddy, 1984). Some of the man-induced changes in a fishery that can result from periodic waves of investment, are illustrated in Figure 20. These periodic fisheries investments can be initiated by a particularly good year's recruitment to the fishery, and should be taken into account in socio-economic studies of the fisheries sector. Further approaches to classification of fisheries that take the above points into account, are given in Caddy and Gulland (1983).

Figure 19 Long-term yearly anomalies in sea level, air temperature, sea surface temperature and diatom abundance at Scripps pier (Redrawn from Horne and Platt, 1984)

2. SYSTEM STABILITY AND MANAGEMENT

The environment: It goes without saying that fish populations, their abundance and species composition, are very responsive to changes in key environmental variables that are not under control of the fisheries administrator. In many cases these variables may act on the stocks in a way that is poorly understood, despite a large literature on effects of a range of environmental variables on growth, survival and reproductive success. Convincing evidence of the impact of environmental change, even without fisheries, has been provided by recent studies of, for example, the occurrence of fish otoliths in stratified bottom sediments from productive upwelling zones; these showed wide fluctuations in abundance and frequent switches in dominant species at specific sites over hundreds and thousands of years prior to human exploitation (e.g., Devries and Pearcy, 1982).

The simplest mathematical representation of a two species system that has been put forward to explain regular fluctuations in linked predator and prey populations, is that referred to as the Lotka-Volterra model and this has been one basis for the modelling of oscillating resources (e.g., Larkin, 1963). it seems likely however, that the extension of such an approach to more than two species will result in a proliferation of population parameters at a greater rate than it will be possible to collect the necessary data to estimate them. This model nonetheless indicates in simple form that an assumption of ‘steady state’ may have to allow for a degree of oscillation in the abundance of components of a (simple) food web around the mean, even without extrinsic influences. Such oscillations in recruitment - hence population size, in otherwise stable populations, can infact be shown to be initiated by human responses to changing abundance (Caddy, 1984; Allen and McGlade, 1986).

It has also been observed that the behaviour of dominant and subordinate species in an eco-system may be modified in relation to environment, if they change places in relative abundance in the system. Skud (1982) noted for example, that if all species when dominant respond favourably to increases in temperature, they may show the opposite reaction when subordinate, since they are now suffering population pressures from the dominant species which is more successful with its higher biomass in taking advantage of the changed environment. The point of this observation is that environmental effects, if not understood or taken into account, will in some marine systems make for considerable errors in the yield forecast from a given expenditure of effort, or alternatively, will mean that the effect of a given management measure will be very variable.

Analysing the relative impacts of system instability and intensive fishing is a problem area that has greatly concerned fisheries workers in recent years (e.g., Sharp, 1980b; Sharp and Csirke, 1983), and there is growing evidence that changes in species dominance may be accelerated by fishing. An example is the apparent partial replacement of a sparid dominated community in the 1920s by a community dominated by cephalopod species (Caddy, 1983). Such processes may best be modelled in a discontinuous fashion, and the theory of catastrophes has been proposed as one way of describing this abrupt switching of community dominance (see Figure 18 for an example of a “cusp catastrophe”and Saunders (1980) for some simple theory.

Figure 20 An idealized fisheries “cycle” showing trajectories of some important variables (From Caddy, 1984)

The general conclusion that has been drawn from the phenomenon of temporal variations in marine environments and fish ecosystems, is that there is already a substantial degree of ‘elasticity’ built into most marine systems during their evolution, and that it is this elasticity that permits harvesting by man on a continuing basis as long as the stresses imposed are not excessive, or do not drive the system into a different state.

This concept of system “elasticity” or resilience does not necessarily mean that ecosystems return directly to their former state after the disturbance. Odum (1969) recognized for terrestrial communisties that the diversity of a commmunity increases in time along a gradient of successional stages as a new (or depleted) habitat in recolonized. It is not clear to what extent this concept of “successional stages” or increasing degrees of community maturity, applies to the marine ecosystem however; given that some highly productive areas (e.g., upwelling systems), seem to be “arrested”at an early, simple stage of community development. However, the progressive colonisation of marine surfaces (marine “fouling”) shows some similarities, with a succession of organisms predominating. As Odum notes, estuaries and intertidal zones are maintained in an early, relatively fertile stage by the tides, which provide for rapid nutrient recycling. In complex food webs on land, however, the bulk of biological energy follows detritus pathways, a situation that has some parallels in marine ecosystems described later in this document. Clearly tapping the detritus food web (e.g., by shellfish culture) can lead to some very high levels of marine production. Odum (1969) contrasts the main features of early and late successional stages of ecosystems in the following manner:

With some qualifications, this rather oversimplified generalization applies in the marine environment also, although as noted by Holling (1973), the more obvious antithesis to stability in the above scheme is resilience, which is not notably a feature of most mature ecosystems, but should be in Holling's opinion, a characteristic of schemes for ecosystem management. In managing a mixture of successional stages, we should, in his words, be prepared to “view events in a regional rather than a local context”, and “emphasize heterogeneity”.

3. SPATIAL CONSIDERATIONS: MAPPING FISHERIES RESOURCES

The importance at an early stage of a fisheries resource investigation of mapping the spatial distribution of key components in the biological system to be studied is emphasized by most ecologists, but has perhaps been neglected in fisheries investigations, perhaps because of the difficulties of actually determining the distribution of ichthyofauna, and of related environmental and substrate (e.g., sediment type) characteristics, below the intertidal zone. A renewed interest in this procedure has been evident in recent years (Butler et al., in preparation), and in tropical clear water areas, mapping of key features of the shallow sublittoral by remote sensing (satellite) technology has made this more feasible than formerly. From the perspective of analyzing multi-species fishery systems, it is likely that a better understanding of community structure and its response to fishing for demersal fish assemblages, would be one consequence of more precise spatial information on fish distribution and catches, as well as on the areas fished seasonally by the commercial fleet (Caddy and Garcia, in press).

One of the important components of resource mapping is thus the characterization of the principal marine communities or ecological complexes, and their geographical extent. This may be done on a very broad scale initially - e.g., Garcia (1982) for West African exosystems: (see Figure 21 for an example of a mapping of the distributions, migration and stock units for one important pelagic resource); and Baisre (1985) for some generalizations on the distribution of Cuban shelf ecosystems. Baisre divided Cuban fishing grounds into three main complexes, and since most Cuban fisheries are fully or close to fully exploited, fish productivity estimates for each were then possible from separate estimates of their spatial extent. The three complexes are as follows:

1. The estuarine littoral, including coastal lagoons, estuaries and bays where land effects are important, and where food chains are characteristically short due to high, irregular environmental stress caused by intermittent (seasonal) inflows of fresh water, nutrients and sediments. These areas have the highest productivity; estimated in Cuba at 1.47t/km² of shelf, they include such coastal ecosystems as mangroves, discussed later, as well as supporting high value fisheries, especially for penaeid shrimp. In Cuba, this area is estimated at 8 500 km².

2. The second ecological complex in Cuban waters is associated with coral reefs. Coral reef species are of considerable importance, but many of the important fish species here also forage at night on adjacent turtle grass beds, which Baisre (1985) includes in the same complex. This complex is typified by long food chains due to high environmental pre-dictability, and an overall productivity estimated at 0.58 t/km², and 45 000 km² in extent: spiny lobsters are the most valuable species.

3. Cuban territorial waters overlap an area of much greater extent, namely the fisheries complex characteristic of oceanic waters, with a fisheries yield of some 0.24 t/km² tuna species (especially skipjack and blackfin tuna). To an undetermined extent, some of this production here results from outflows from the preceding two zones.

Trophic diagrams showing the interchange in Caribbean waters between these major ecological zones are given in Figure 48.

Portraying Events in Time

Two types of approaches to the portrayal of the time sequence of events in ecosystem analysis might be mentioned here, which can be graphically represented as an aid to research decisions. The first of these, path analysis, represents an attempt to express the functional and temporal relationships between abundance of food web components, and the physical and biological factors that influence them. Such an approach (shown in Figure 22) for investigations on the causes of stock fluctuations in flying fish), is a prerequisite for time series analysis, and has been formalized under the name “path analysis” This is a procedure for explicitly examining hypotheses, and developing causal models linking time series of correlated variables. A good example of this approach is given in Coelho and Rosenberg (1984).

Flow Charting or Scheduling Research Activities

Recognizing that it is important to represent events sequentially in time, this approach to representing points in a complex train of events is widely used, and may be represented in the case of food web analysis by Figure 8. Such an approach also has obvious applications for planning research programmes (see e.g. Welcomme and Henderson, 1976) and Figures 22 and 23.

Monitoring seasonal changes

Information on seasonal distribution and relative abundance will be needed to distinguish key migratory species from the resident species making up the main fish assemblages (e.g., Tyler, 1971), and in this case, separate distribution maps may need to be drawn for each main season of the year. This information should be supplemented where available, by a comparison of annual time series of abundance or landings of each species. Here, the records of local fish companies and markets, and of buyers and fishermen, may provide useful clues to past and current variability of species catches. Temporal coincidence (common or contrary) and changes in relative species abundance, may be useful clues to possible trophic interactions or competition, if supported by other biological information.

Six types of questions can usefully be asked in the course of a fisheries mapping exercise:

1. What is the appropriate spatial scale of the investigation?

2. What is the spatial distribution and relative abundance of those species which are now economically important to fishermen, on a seasonal basis, and throughout the year?

3. What information is there (in records of fish buyers or the memories of other fishermen) on the extent of past year-to-year variations of the important species?

4. Which species occur locally year round, and which are transient, occasional or migratory?

5. What information is there on the diet, behaviour and ecological interrelationships, of the main species in this area or in similar (adjacent) areas?

6. Are there any obvious seasonal changes in food preferences of the main species of commercial interest? (see e.g., figure 24 for an example of seasonally varying food linkages).

A degree of spatial coincidence between those species that share a common habitat for at least part of the year would thus seem to be a necessary precondition for an important species interaction to take place. Records of the areas of distribution of important species can be provisionally mapped on a seasonal basis from e.g. earlier resource surveys or fishermen interviews. Bearing in mind that this latter kind of information is likely to be less than completely objective, it may still be worthwhile to try to assemble it, and compare distribution ranges for key species especially at an early stage in investigation.

Figure 21 Mapping of seasonal and long-term changes in migration and distribution of two pelagic species associated with the West African upwelling system (From Garcia, 1982)

Figure 22 Suite of factors to be considered in evaluating interannual variability of flying fish (From Mahon, Oxenford and Hunte, 1986)

Figure 23 Flow chart of decision-making in management of fisheries (From Welcomme and Henderson, 1976)

D.O.M.: Dead organic matter

Figure 24 Simplified representation of the feeding relationships of a cod stock showing differences in principal prey at two times of the year (January to June and July to December) (Redrawn from Armstrong, 1982)

4. FISH PRODUCTION PER UNIT AREA ESTIMATES

The final major group of methods that avoids to some extent the questions of biological detail dealt with elsewhere in this work, is the use of past experience in exploiting ecosystems to suggest what is the expected fisheries productivity of the area in terms of yield per hectare or square kilometre. A summary of some typical levels of productivity is given in Marten and Polovina (1982) and substantiates the impression given elsewhere, namely that productivity falls off along the gradient from estuary, coastal lagoon, or upwelling area, to high seas; although a rather wide variance of these estimates around the mean for each habitat type is typical. Thus, for the shelves of many Caribbean islands, the annual harvest varies from 0.4–0.5 t/km² to as high as 18 t/km² for specific habitats in the Indo-Pacific (or even potentially 35 t/km² for American Samoa - Munro, 1984). This wide range undoubtedly reflects regional differences, but contains potential biases from at least two sources:

1. Multispecies resources of a similar basic productivity may differ in yield depending on the state of exploitation. This should be taken into account by comparing yield/area with fishing intensity (effort/area) over the same area: (Munro, 1977; Caddy and Garcia, 1983). The yields expressed at some standard fishing intensity may then be more comparable.

2. Because of imperfect habitat mapping, the extent (area) of productive habitats is normally not correctly measured, and is likely to a greater or lesser extent, to include (less) productive areas in the hinterland.

Both of these sources of error can be partially avoided by specific attention to mapping of habitats and/or ecosystem or assemblage distributions, collecting where possible, data on production by ecosystem areas, or by Assemblage Production Units, APUs, (Tyler, Gabriel and Overholtz, 1982), and also, by considering such ecological units as the ‘building blocks’ of the fishery management system for shelf fisheries. This approach leads to a consideration of production per unit area as a function of harvesting rate (Figure 25).

Figure 25 Production per unit area as a function of fishing intensity for a coastal fishery with seven main ports (Modified from Caddy and Garcia, in press) (^fmsy is a rough estimate of the fishing intensity providing Maximum Sustainable Yield under “average” conditions)

5. COLLECTING AND ANALYSING DATA ON FEEDING PREFERENCES OF COMMERCIAL FISH SPECIES

The ability to carry out any serious resource investigation needs, as a first precondition, familiarity with the main taxonomic groups present in an area, and their local common names: (bearing in mind that these latter may vary from area to area). The FAO fish identification sheets offer the most direct approach to identification of fish species and are available, or will soon be available, for many tropical areas (Table 1). These sheets can be used over a period of time to build up a fauna list from species identifications in a given area. If separate files are maintained for each key commercial species, to these can be added direct observations on occurrence by area, season and depth, charts of the main fishing and spawning areas, and other relevant literature. From these source materials, a more or less comprehensive picture of the state of knowledge (and gaps in knowledge) of the species in question can eventually be assembled. The importance of such a “mapping and filing system” for resources is emphasized in Caddy and Garcia (in press) and Butler et al. (in press); and is also a logical basis for development of a statistics monitoring system (Caddy and Bazigos, 1985).

^a Note: A variety of species Synopses and Catalogues summarizingbiological information by taxonomic group are also available onrequest from FAO

Elucidation of trophic relationships

Four main approaches to elucidating of trophic relationships exist, namely:

1. Direct observation methods.

2. Experimental studies.

3. Examination of morphological adaptations of component species.

4. Stomach content analysis.

The first of these methods is particularly feasible in clear water areas (e.g., coral reefs) where it is of growing importance, (e.g., Hobson, 1974). The second class of methods ranges from field experiments involving removal of predators from closed and natural systems and observations on subsequent perturbations (e.g. Paine, 1969), to experiments in field or laboratory conditions on choice of foods. The laboratory approaches will not be described here, involving as they do a significant investment in laboratory facilities and extensive time to achieve results. They are also to an unknown extent problematical, particularly in relation to food preference, which is often difficult to extrapolate to nature, given that the relative availability and abundance of prey is not representative of natural conditions (see however Chapter 12 Part II for an attempt to use existing data on quantities consumed in feeding studies). Despite the significant investment in time and effort necessary for this kind of experimental studies in the field, experiments on caged areas of a bottom community, excluding predators, (e.g., Hancock and Urquhart, 1965; Young, Buzas and Young, 1976), look like being useful in estimating the contribution made to natural mortalities by the various predators, and their role in the food web (Figure 26).

Figure 26 Illustrating an approach to investigating food webs by means of cage experiments: comparison of faunal abundance inside and outside the cage may help to differentiate the impact, for example, of crabs and predatory fish or benthic invertebrates (From Young Buzas and Young, 1976)

The third approach is one which intuitively allows us to distinguish, at least to a first approximation, the main feeding habits of the particular size range of the species in question. This approach relies on an examination of the body configuration and size, and the presence of various morphological adaptations in the species; in particular shape and size of mouth and dentition, form of the gill rakers and intestinal tract, in order to allow a preliminary classification to be made into (for example) planktivores, bottom feeders, grazers or browsers, piscivores, etc., although this type of approach rarely substitutes for method 4 in determining trophic inter-relationships.

Stomach content analysis will of course present more complicated problems of identification because of the small size and poor preservation of the stomach contents than direct sampling of the prey, but may provide a more extensive species list and size ranges than that obtained from the commercial catch. This is because a number of species and size will be encountered, possibly for the first time, that are not commercially important or catchable with fishing gear. The problem of species identification can be approached initially by identifying unfamiliar species first to genus or family, and, secondly, in a first (and probably qualitative) approach to food web analysis, only attempting more specific identifications on well preserved and documented specimens after consultation with (museum) taxonomists who are experts in the various groups. In the meantime, an empirical key must be developed for unknown but readily characterized taxa that will at least allow the work to continue (see e.g., Wildish and Phillips, 1974).

Stomach content analysis, even in its most casual and anecdotal form, can yield incidental but immediately valuable information, since predators are often better sampling devices than most commercial fishing gears. The presence of a number of valuable resources (e.g., squid, crustaceans) not yet commercially harvestable from existing gear, have been first identified in an area from this type of observation.

Although entire food webs have been elucidated based on gut-content analyses (e.g., Roger and Grandperrin, 1976), caution should be exerted against over-enthusiasm for large-scale systematic studies of this kind which can be particularly expensive in using up scarce and valuable manpower, research vessels, laboratory workspace and storage facilities, as well as valuable computer time. This appears to be one of those types of activities that very soon leads to rapidly diminishing returns with time. A great deal can however be achieved over a period of time if this type of observation is fitted in with other field work, and we may note that such an approach is a necessary precondition to an understanding of species interactions. Observations on freshly-caught individuals of known species, sex and size can be accumulated over several years when opportunities permit, and the information stored on file cards or in folders marked by species (predator or prey) names, allowing a picture of the trophic interrelationships to build up gradually, which can eventually be abstracted and described. Each individual record should note, in addition to the species of prey and their size and numbers present, the place, date and time of capture, depth and other observations of possible interest; cross-referenced to other sources or records of information collected at the time - e.g., catch rate and size composition of the commercial catch, depth, bottom type as well as fishing gear used, as well as to any museum or otolith specimens that may have been retained.

A description of the mechanics of analysis of stomach contents is given by Bagenal (1978). Four major problems should be recognized before collecting data so that precautions in sampling can be appropriately taken:

1. use collection methods that minimize regurgitation of food (such as, for example, caused by gillnets);

2. avoid holding fish for protracted periods (e.g., long lines, traps) before removing and histologically fixing stomach contents;

3. given that most fish species have a relatively restricted period of feeding, timing of capture should ideally take into account the diurnal feeding cycle and also seasonal variations in diet, where these occur;

4. the rate of digestion (or more relevant, the rate of loss of identity of the food item) is more rapid for soft-bodied prey than for those with pronounced exo- and endoskeletons, thus making greater problems in identification, and exaggerating the contribution of the latter type of prey species to the diet if these are retained longer in the gut.

Gravimetric, or if not possible, volumetric evaluation of the gut contents in order of exactness, is the recommended approach in most texts. We may note, however, that for a preliminary investigation of the type we are suggesting here which is not principally concerned with the energetics of biomass transfer, enumeration and measurement or estimation of length ranges of fish prey species is largely adequate, and may be of more importance than total species weights in terms of recognizing critical stages in predator life histories. The first thing that may be evident and of potentially commercial importance, is to recognize that certain changes in feeding preferences occur seasonally and in the life history, and to note the periods and sizes at which these occur. Any study of the species may then recognize these size categories in field sampling. A more direct application is to recognize that in certain areas and seasons, a species may be available to capture in a fairly limited zone (or season) where a particularly characteristic and locally available food species occurs. Information on such (potentially commercial) prey species may not always be evident from experimental fishing or survey results.

Preliminary food web analysis

As we have noted, studies of trophic interactions in northern latitudes are based upon a long and rich scientific literature on life history, food preference and behaviour for a system where relative abundance and size and age structure of most fish populations are rather well known. Under these conditions, the construction of complex interspecies models (e.g., Anderson and Ursin, 1977; Laevastu and Larkins, 1981) is a logical extension of a rich data base, as is the extension of other conventional methods of single species analysis to multi-species use (e.g., phalanx analysis; Pope, 1980). These rather detailed approaches to multispecies assessment are less likely to have any early application in data-poor situations, however, (and we may note that these models are not usually applicable on a routine basis, even in better-studied areas). Larger models with many parameters, in fact pose serious problems to the fishery scientist of a philosophical nature: if all the information is included in the model, what data should be used for testing it? Hence we may have to be content with some understanding of the structure and main interactions of the system, and in consequence, expect to gain at best, some qualitative guidelines for multispecies management.

At first approximation, it may be convenient to proceed directly from a preliminary prey-predator contingency table (e.g., Figure 42) to construction of a tentative food web (Figure 43), even though estimation of the rate of feeding of predators in absolute terms requires a considerable knowledge of the ingestion, digestion or egestion rates, both as a function of environmental factors and of type and abundance of food. These latter factors will be touched on later, but for the moment the assumption is made that the relative volume of identifiable food contents per predator body weight is an index of their fraction in the total diet. This assumes that the sample of stomachs is properly weighted both for diurnal and seasonal factors, and for the distribution and relative abundance of the different size groups under investigation.

To some extent, a preliminary trophic classification of this kind is possible by reference to the literature, although information is surprisingly sparse on the diets of many marine fish species.

“Stomach content analysis” (including intestinal contents) is the principal method available to the field worker with limited facilities for intensive laboratory or field studies. This methodology, and the data it generates, has been inadequately utilized to date, mainly, we believe, because of uncertainties in the interpretation of the large volume of data that can be rapidly obtained by this method. To a significant extent this is also because methods of population analysis until recently have placed inadequate emphasis on multispecies approaches and species interactions.

6. EQUILIBRIUM CONCEPTS AND THE FLOW OF ENERGY THROUGH A SYSTEM AND ITS OPTIMAL UTILIZATION

6.1 The equilibrium concept in fisheries

It may be useful when discussing ecological stability, to follow Johnson (1981) in contrasting the concept of an isolated system coming to a unique thermodynamic equilibrium, with the term equilibrium as it is normally used in the fisheries literature.

In thermodynamics, the flows of energy and/or materials at equilibrium across a boundary between zones A and B: i.e., A⇋ 2B, are equal and opposite; implying a state of rest exists, where there is no net transfer of energy and materials.

In population dynamics, the rate of flow X of energy from the rest of the food chain, to the component of population biomass (B) being subject to fishing, is shown in the following simplified scheme:

The rate of flow of chemical energy in the form of biomass tapped by man, and by natural predators are respectively the (ponderal) rates of fishing (F) and natural (M) mortality as discussed in the literature on fisheries assessment. The flow of materials of immediate concern to the fishery is from left to right in the above scheme and although, eventually of course, there is a reverse flow as shown by the dashed lines, this does not constitute the “equilibrium” referred to here, which is said to occur when a constant yield can be removed from the population biomass per time interval, without prejudicing the systems' ability to sustain this yield. The yield can then be considered “surplus” in that it does not affect the stability of the population. Of prime interest to fisheries ecologists then, are those factors affecting the rate of flow (X) of energy and materials from the food web into the exploited system, and how variations in the energy entering the system, and the structure of the food web itself, affect the yield from the system. Clearly this yield cannot be considered as being strictly at equilibrium with biological production unless both production and harvesting rate are constant or unless the latter is closely responsive to the former. Any “equilibrium” that may come to exist, will therefore not be passive, but will depend on a good system of monitoring of both production and fishing effort, and on a constant M.

Before looking at an ecological system from the point of view of its energetics, we could begin with an analogy, by examining a better known system which is currently under intensive discussion, namely the different uses of energy by modern industrial societies.

Figure 27 shows in a diagrammatic form how society makes use of the various potential energy sources available to it, by in most cases transforming raw materials into another form that is most suited to the many and varied applications modern industrial society has found for energy. Several things are noticeable there.

Figure 27 A stylized “spaghetti chart” showing uses of energy in modern society (After Louins and Louins, 1981)

1. Figure 27 is in fact an “artificial” dissipation structure in which energy is transformed from a less acceptable form into one that is more easily used by society; noting from the first law of thermodynamics, that although energy can be transformed from one form to another, it cannot be created or destroyed.

2. A large part of the input energy is “lost” during both conversion and use by man, and (a consequence of the second law of thermodynamics, but not shown here), still more energy must be expended in order to rectify some of the side effects of the use of the energy by man. These result from the need to dispose of the physical by-products resulting from consumption of fuels, (i.e., chemical and physical pollution), and of waste energy: (so-called thermal or heat pollution). These ‘down- stream’ effects of human interventions also apply in natural systems (see later).

3. Not all sources of energy are equivalent in “quality” in our modern society - in fact chemical energy (oil derivatives) are “high grade” energy sources and less easily substituted for in their applications than, for example, coal. In other words, in order to use coal for running a car, it is necessary to first transform it physically or chemically, and in doing so, heat (energy) is lost, and a variety of costs incurred; i.e., the energetic efficiency of the overall system is reduced.

Each of these three features of the above system are equally valid for optimal utilization of a food web - also properly regarded as an energy dissipation structure. This transmits the (still largely unusable) solar energy falling on the ocean into chemical energy, i.e., phytoplankton, by photosynthesis into sugars and other carbohydrates, and subsequently to fats and proteins. This energy is large in quantitative terms, but low grade (i.e., a greater portion is unusable in quality) and/or dispersed in space and time so as to be difficult or costly to harvest (as phyto-plankton). Each step upwards in the food web results in a significant energy loss - of the order of 70–90 percent per step - of that initially available. After 1–3 more steps, the food energy may in some forms be considered “higher grade”, i.e., more acceptable or desirable as a human food item. Equally important, it may be harvestable in a cost-effective fashion (i.e., occur in larger unit “packages”) which, especially in the case of schooling fish, or species that can be aggregated by bait or other attractants, can be harvested yet more cost effectively as a group or school. Obviously there are many exceptions to the idea that level in the food chain is directly related to perceived quality in human terms. For example, in terms of marketability (“quality”), tuna and abalone are higher-priced products on the fish market. Each however, has several different features ecologically:

1. Abalone is a herbivore and therefore near the bottom of the food pyramid. It is also easily harvested as well as locally vulnerable to overfishing, but suffers from a major disadvantage in that its substrate or ‘niche’, and hence its distribution, abundance and biomass, are very limited.

2. Tunas are highly mobile apical predators in their harvestable stage, but with a habitat or niche that is very large in extent. Obviously only a small fraction of the chemical energy originally synthesised in the oceanic habitat as biomass is available for harvest by man as tuna, and even then, only with considerable investment in technology and effort. On the other hand, the ecological system supporting tuna is very large, diffusely patchy, and the energy needed to harvest and utilize most of the lower trophic levels leading to tuna, would be prohibitive with current technology, fuel and labour costs and market prices. Hence, harvesting a large schooling fish such as tuna becomes an acceptable alternative to harvesting the less available and lower priced, but potentially more productive trophic levels lower - down, (see Chapter 13 on the oceanic tuna system).

Systems concepts of this kind offer us messages that are potentially of great interest to fishery managers, which will be touched upon in greater detail elsewhere in this paper, but can be summarized as follows:

1. In deciding to harvest one or more components of a system, one should explore first the present uses made of the chemical energy by the other components of the unexploited ecosystem and by man. Viewed as a dissipation structure often evolved over millions of years, eco-systems contain an elaborate systems of feed-back loops that means that no interference with the system outputs will go entirely unnoticed by the system, (although a generally high level of system “damping” is the case). Neither will these perturbations necessarily be “harmful”, especially if the man-induced perturbations lie within the natural dynamic range of the system's capability to respond to normal variations in environmental or biological variables.

2. When looking at a system before harvesting, we should consider each component from the point of view of (a) its volume (biomass), (b) its availability and vulnerability, and cost of harvesting, (c) its marketability (value) or ease and cost of transformation into alternative more acceptable or desirable products, and last (but far from least), (d) the impact of harvesting any component on other components of the system, or on the system itself.

3. An exploration of how the system components interact will not give you hard and fast answers to all these questions, but will allow you, or a “manager”, to identify the risks of any particular action.

4. One of the main uses of a knowledge of trophic interactions will be that ‘down-stream’ effects of human interventions on the ecosystem may be better visualized in advance (although much more information will be needed before these can be evaluated quantitatively).

Benefits from harvesting an ecosystem

Expressed in simple terms, the benefits from any given strategy of harvesting an ecosystem consisting of i = 1, 2, 3, ..... n species, is the sum of the benefits from each of the components, i.e.:

The last term in equation (A), the Interaction term, although difficult to quantify, is essential for generality: it covers those incidental losses, benefits to society or down-stream effects, that will sometimes result from a harvesting strategy, but which are not contained within the fisheries subsystem being considered. Thus, for example, the commercial harvesting of individual components of an ecosystem (e.g., using coral heads for building material in the Indo-Pacific or commercial gillnetting of Atlantic salmon), may each be cost effective when considered as isolated activities, but will also have incidental effects on local revenues from the tourist, sports angling, or sport diving industries. In the first of these two examples, it may also result in deleterious ecological changes, since coral is itself a substrate species supporting reef fisheries, as well as a natural protection against coastal erosion, so that coral harvesting will affect the viability of the system as a whole. Similarly, for a marine trawl fishery to develop, provision of freezer of harbour facilities for the fishing fleet may be a necessary cost for developing the resource, but may also offer other benefits to the food industry or to coastal transportation (see also Chapter 7 on the mangrove ecosystem). This Incidental Component can sometimes (but not automatically) be discounted, but should not be ignored, since it may be considered to represent the linkage between the fishery system, the ecosystem as a whole, and the rest of society, its economy and objectives. Approaches to understanding the economics of natural systems within the constraints imposed by the biosphere, are now beginning to enter the literature (e.g., Passet, 1979).

No theoretical tools yet exist that permit an a priori determination of the optimal level of harvesting of a whole ecosystem; but if broad enough data are collected, we can consider the impact of a given change, assuming that the change in harvesting strategy is sufficiently marked and has been in place long enough for a new more or less steady-state condition to be reached. The impact of a change in harvesting of any component (i) on the marine sector can then be given by the following expression:

Obviously decreases (-) can be substituted for increases in the above expression with a change of sign on each of the boxes. Implicit in the above equation is the realization that decisions taken with respect to harvesting of one component will have impacts on the other components of the fishery system, and could even have economic impact outside the fisheries sector, as for example, if an increase in coral fish abundance had increased revenues from tourism. Even if we are rarely in a situation when we can define these impacts quantitatively, it is possible under some circumstances to give an idea of the nature of the incidental effects likely to ensue, even if not of their extent.

An example of such a check list of renewable uses and impacts is given in Table 2 for coral reefs.

7. QUALITATIVE CONSIDERATIONS IN ECOLOGY - THE MANGROVE ECOSYSTEM AND CORAL REEFS

In its most fundamental form, a knowledge of how the various species in our environment are functionally linked together has been part of the human heritage since prehistoric times. The limits on production of the systems they live in are often enshrined in the fundamental religious beliefs of hunter-gatherer societies (and in fact encompass the idea that man is part of nature also). Unfortunately inhabitants of our modern technological society are not automatically aware that the interrelationships which exist between components of an ecosystem are not infinitely elastic, and societies still act in many cases as if these matters pertained to systems extrinsic to their own lives and local micro-environments, whereas it is clear that all environments have a limited capacity to withstand extrinsic stresses.

A good example of the complex and far-reaching impacts of such perturbations is the mangrove ecosystem - a highly important coastal environment common to tidal and brackishwater areas of tropical and subtropical areas of the world (see Tables 3 and 4). Figure 28 from “Conservation Indonesia” gives a simple illustration of some of the typical interdependencies existing within mangrove systems, showing the natural pathways of material from leaf detritus and other plant production inside the mangrove forest, and its subsequent dissipation outwards and upwards in the food chain. This material enters the aquatic system as leaves and sticks, and later in the form of detritus and the excreta of herbivores, and is used as a substrate for an abundant bacterial flora. It then passes in sequence through the bodies of detritus feeders (shellfish and other invertebrates), primary (small) carnivores, and secondary and higher (large) carnivores. Also briefly illustrated here are the multiple usages of various components of this food web by the human inhabitants of the coastal ecosystem. Just in terms of natural products harvested by man, the timber itself provides fuel and a diversity of construction materials, including a number of valuable and inexperience components for the construction of fishing gear; plus a wide variety of foodstuffs, folk medicines agricultural and other products (Saenger, Hegerl and Davie, 1983); not to forget the key role of this system in the life cycles of other components of the coastal ecosystem, including man, e.g., Figure 29. Kapetsky (1985) estimates some 460 000 artisanal fishermen live around mangrove associated estuaries and lagoons (Table 4). Beginning first with the production of plant material, Christensen (1978) estimated for a stand of Rhizophora in southern Thailand that annual production of organic matter above the ground was equivalent to 7 tons of leaves and 20 tons of wood per hectare per year. This amounts to 10 tons of plant production that is ultimately not used by man (twigs, leaves, fruit, etc.). This “deciduous” component of production per m² of mangrove forest is roughly five times the phytoplankton productivity of the same area of coastal zone without mangroves, and Kapetsky (1985) estimates very high figures for production from mangrove-fringed estuaries and lagoons (91 kg/ha/yr). World-wide, this is roughly 925 000 t of fish and crustacean products coming from coastal and estuaries areas where mangroves are important components of the ecosystem. This almost certainly implies (as for other coastal macrophyte production systems, e.g., kelp and sea grass beds), a large net export of detrital material to the food chains of the continental shelf. Mann (1972), for example, estimates that roughly 40 percent of the detrital material produced by macroalgae in high latitudes is dispersed offshore to adjacent shelf areas and eventually to groundfish food chains.

Figure 28 Interdependencies and dissipation outward of biological production in the mangrove environment (Modified from “Conservation Indonesia”, Vol.5, No. 3, December 1981)

The other roles of the mangrove complex in the ecosystem and life history of the associated fauna are well marked: close to 30 species of fish, 200 species of crustaceans, over 200 species of birds, and a wide variety of species from other taxa make this a highly diverse community with a complex food web that is much interlinked with adjacent ecosystems, and would therefore (like other highly diverse environments, such as the tropical coral reef), be impossible to portray as a detailed food web. Some degree of compartmentalization of the system into herbivores, detritivores, primary predators, etc., is necessary before going much further in describing such a system. However, even here changes in trophic level during the life history are common (Table 5), and make this objective far from easy to attain. The detailed description of such a complex system with many external linkages as a ‘closed’ food web, may not therefore in itself be very helpful, especially bearing in mind that the coastal, and especially the intertidal, zone is a transition area between terrestrial and aquatic systems, and as such is completely “open-ended”. The mangrove system receives inputs from terrestrial systems in the form of runoff of freshwater containing dissolved nutrient salts, organic material and growth factors, bacteria and often, pollutants either in suspension or adhering to silts, clays or detrital material. The system provides similar outputs to offshore areas, as well as providing habitats (nurseries) for many coastal or even off-shore species of commercial importance at one or other stages of their life history.

It would be necessary therefore, in modelling such a system trophically, to show the major routes whereby the contributions to, and losses from such an open system occur, if a food web is to be of much predictive value in this case. The economic importance of such ‘transition areas’ as estuaries and lagoons to fisheries, is documented in various places (see Table 3), and is considerable. For example, it is estimated that as much as 90 percent of the U.S. commercial catch, and 70 percent of the recreational catch in the Gulf of Mexico, is made up of fish, shellfish and crustacean species that spend all or a vital part of their life history in estuarine areas (where mangroves form an important part of the habitat). It has even been demonstrated (Fuller, 1979), that the length of shoreline in estuarine areas is a controlling factor in shrimp production: something to consider when “rationalizing”shorelines by landfill in coastal wetlands.

Here we may specifically mention the penaeid shrimps, some species of which, although spawned offshore, spend their critical inshore juvenile stages in coastal lagoons and mangrove dominated areas (see Garcia and Le Reste (1981) for a review). Some indications have emerged of an apparent regularity between the declining extent of mangrove areas and the decreased shrimp catch offshore from each coastal area, and a similar relationship between mangrove area and offshore shrimp production from different regions of the coast of the Philippines is suggested in Martosubroto and Ndamin (personal communication).

It seems as if it is necessary to consider all potential usages of these habitats and, to look at the linkages to adjacent offshore environments when considering major changes or alternate uses of these areas for (say) building, landfill disposal, forestry, “wild” fisheries, salt pond and fish pond use, etc. Similarly, on the terrestrial side of the land-se interface, mangroves perform a useful function as windbreaks for tropical hurricanes, help in flood control, and prevent soil erosion. All of these factors and more are discussed in Christense, (1978), and also need to be considered before major ecosystem changes are made: (see cover design).

Figure 29 Some functional relationship of mangrove with artisanal fisheries, shery resources and aquaculture (From Kapetsky, 1985)

In conclusion, although this system is one of the more complex in the world in terms of species diversity and interactions, we must take a holistic view of the problem before focusing on management decisions relating to any single usage. It is also clear that economic modelling based on one or more social, economic or biological subsystems taken in isolation, will be completely misleading in assigning net values to one particular strategy or another. In this case, the interaction term which expresses the impact of any one utilization on other resources (Chapter 6) is likely to be larger than the benefits from any one potential usage taken in isolation.

Although actually calculating net benefits from different usages of the mangrove littoral is not easy to do, the most feasible approach is probably to consider for each strategy the net revenues, either per hectare of forest, or per kilometer of coastline, for any change proposed. This calculation should take into consideration the rotation time of the elements of this unique system and then integrate these over the whole area. In general, the optimal strategies of usage seem likely to be those that follow a “zoned” approach to utilization, based on a categorization of each section of the coastline in terms of its particular suitability for each potential usage. Figure 29 after Kapetsky (1985) presents a useful schematic in these cases.

The coral reef system and fisheries

The problems of fisheries management of complex ecosystems is perhaps best epitomized by coral reef resources, which Pauly (1981) suggests may contribute up to 10% of the world's annual fish catch. The extremely high diversity of these ecosystems and their very complex food webs frankly complicates the problems of their scientific management on a food web basis, although as suggested elsewhere in this document, the approach of cropping off the apical predators should probably be abandoned in favour of a strategy of fishing all trophic levels so as to keep the community ‘balance’ as similar as possible to an unexploited state. Paying attention to the obvious need to conserve the ‘substrate’ species (here referring to the wide diversity of encrusting organisms that form the bulk of the reef) is a top priority. This would mean for example avoiding destructive practises such as use of bleach or dynamite for fishing, and avoiding siltation and domestic waste runoffs in sensitive areas. This question and some suggested strategies are well described in Kenchington and Hudson (1984) who discuss a range of practical problems relating to conserving reef resources.

From the perspective of research in support of fisheries management, it is unfortunate that a great deal of the scientific literature deals with various specialized features of life cycles of coral reef fauna, and only quite recently (e.g., Munro, 1982), with fisheries management needs. This is probably inevitable, given the biological complexities of the system (see Figure 36 for a very simplified view), but does not augur well, except in qualitative way, for the use of food web information directly in management decision making, as we suggest for some simpler systems. What seems indicated is that a conservative strategy for using complex systems be adopted that equates to cautious use of all components, so that the whole ecosystem or “dissipation structure” continues to maintain its productivity. Nor is it obvious that detailed single species analysis, e.g., of the yield per recruit type, is a practical option for each of the many components of the system. Three possible approaches which take into account trophic interactions have, however, been suggested. The first, is to pick a limited number of indicator species (e.g., a herbivore, a grazer, a small predator, a detritivore, a large predator), and monitor and analyse their state of exploitation as indicators of the state of health of the ecosystem as a whole. The second, is to identify and concentrate on one or two “keystone predators” (e.g., Plectropomus leopardus on the Great Barrier Reef - Goeden, 1982). A “keystone predator” (Paine, 1969), is one which because of its economic importance, high fraction of the predator biomass, and diverse feeding habit, will integrate and reflect any adverse changes occurring at the system level. Removal of such a key predator (making up with other plectropomids of the Great Barrier Reef), about 50% of the predator biomass and 30% of the reef catch, apparently led to the wide fluctuations in biomass of other predators observed by Goeden (1982).

Figure 30 Structure of a traditional West African small-scale fisheries community (Redrawn from FAO/DANIDA Project Document RAF/171/DEN)

Figure 31 Landings by gear and port/markets and by eventual product type in the Bay of Fundy (New Brunswick and Nova Scotia) Herring Fisheries (Redrawn from Lamson and Hanson, 1984)

Figure 32 The structure of choice and constraint relationships in Newfoundland fisheries (Based on Lamson and Hanson, 1984)

8. DIAGRAMMATIC REPRESENTATION OF LINKAGES WITH SPECIAL REFERENCE TO FOOD WEBS

Diagrams that illustrate how components are interconnected are of great utility in understanding complex interrelationships, both in aquatic food webs, and in understanding the human components of the fishery ‘dissipation structure’ that results in the fish reaching consumers (see, e.g., Figure 30). Such diagrams can also be helpful in visually integrating social, economic and biological factors, such as the hierarchy of events in the secondary and tertiary components of the fisheries subsystem (e.g., Figures 31 and 32). These go beyond just those aspects of the biological production and harvesting of a fishery system.

Concentrating for the rest of this chapter on the marine ecosystem and its representation, we may note that methods of displaying species relationships in this way by means of food webs have not yet been standardized, largely because there are different levels of information content in our knowledge of species interactions, and different objectives for information transfer. Thus, as in Figure 3, we may simply find a representation of a flow of energy or material from the prey to the predator symbolized by an arrow in the direction of energy flow. It is also possible (e.g. , Figure 33) to roughly quantify the biomass of each component using the sizes of the symbols (circles) surrounding them in food web diagrams, and to estimate the rate of flow of material from predator to prey by linkages of different thickness (Figure 1). Even more elaborate representations may involve a knowledge of the quantity of materials consumed, or the rate of consumption of materials, and their calorific values, organic carbon, protein or organic nitrogen content. These more elaborate representations result from quantitative (usually laboratory) experiments on the rate of feeding of predators, and a knowledge of how the energy content of prey species is sub-sequently partitioned (e.g., Figure 34). These latter diagrams are one of the principal tools of the study of population energetics, and have been formalized in various ways (e.g., Figure 35A) with the objective of identifying not only the direction of flow of energy and other interractions in the food web, but also its quantification in terms of an energy budget for individual species, and for the selected food web components being considered (see e.g., MacDonald, 1983).

We may mention here another type of trophic diagram often shown in the literature: the compartment model, of which Figure 33 is an example. This may be regarded as a condensed form of the food web where subcategories of the community (e.g., benthos, zooplankton, herbivores, etc.) consisting of more than one species, are grouped together for convenience. Although this is a useful simplification, it should be used with care, since the categories of species represented in for example benthos, may be highly diverse, ranging from herbivores to secondary carnivores, so that the true impact of any species may be minimized or misrepresented in the process of simplification; also, as noted elsewhere, many marine species change their effective trophic level in the course of their life history. One other requirement for their use in quantitative calculations, is that no food transfers should be concealed within any compartment. We may note here that exchange of materials between components other than directly by predation can also be represented: dead organisms (as carrion, detritus, dissolved organic matter or nutrient salts) may supply other components, often “lower down” in the web, or be “stored” for indefinite periods in biologically inactive (e.g., anoxic) parts of the environment such as in bottom sediments. Similarly, contributions from unspecified components outside the food web may be portrayed. Although the food web components modelled are usually chosen to be largely self-contained and receiving little material from outside components,as we have noted earlier, realistically this is never completely feasible. Interactions other than trophic ones may also be shown by means of a dotted line implying transfer of influence but not of material (e.g., Figure 36). Non-trophic competition between species for substrate or niches are examples of such influences.

Mention should be made here of the new field of ‘loop analysis’ (e.g., Saila and Parrish, 1972; May, 1974; Li and Moyle, 1981) which considers solely the functional linkages between components (species), and the direction of flow of material and influence between them in making predictions. Applications (such as in the last reference quoted), include evaluating the impact of introducing an exotic species into a food web.

The influence (positive or negative) of one organism on another can be expressed graphically by this methodology (e.g., Figure 35B); arrows indicating the direction of interactions with ‘loops’ connecting species in one or more closed circles. Positive and negative feedacks are identified: the latter processes tending to stable system behaviour and vice versa. The problems with the use of such an approach with other than very simple food webs seem to be threefold:

1. without quantitative information it is difficult to distinguish important from less important linkages;

2. problems of multiple feeding strategies cannot be handled; and

3. determining the stability of the whole system as complexity increases is difficult.

Figure 33 Ecosystem structure in terms of the nitrogen content of the main components (proportional to diameter of circles; directions of flow shown by arrows) (From Jones and Henderson, 1980)

Ex: Flow of excretory products in inorganic nitrogen pool

Do: Flow of dead material to dead organic matter pool

Figure 34 Average partitioning of dietary energy for a carnivorous fish. A ration of 100 calories is equivalent to about 2% dry weight/day for 2-kg fish. Non-fecal energy is mostly excreted as ammonia and urea. I = M + G + E; (where I = rate of ingestion; M = metabolic rate; G = growth rate; E = excretion rate). Amounts marked with an asterisk total 100; figures in parentheses indicate the possible range of net energy distributed to metabolism and growth (Redrawn after Brett and Groves, 1979)

Figure 35A Some of the symbols used in the energy circuit language of H.T. Odum (Redrawn from Platt, Mann and Ulanowicz, 1981)

A: Stable prey-predator system

B: One predator with two preys in which one prey and the predator are self-regulating; the other prey is not

C: A system into which a destabilizing predator has been introduced

Figure 35B Loop analysis of three simple systems. Negative feedback (i.e., causing a decline in population size, or controlling the size of the component pointed to), is shown by broken lines and minus signs, while positive feedbacks is shown by solid lines and plus signs. Negative feedback generally results in stability, and positive feedback in increasing amplitude of oscillation (Redrawn from Li and Moyle, 1981)

Figure 36 Summary of information flow through a community of reef fishes exclusive of food pathways. Dashes lines indicate symbiotic associations; double lines indicate space supplies; shaded lines indicate potential food competition between boxes (From Smith and Tyler, 1973)

Nonetheless, this approach points the waytowards a method of at least illustrating some aspects of system behaviour in a concise visual fashion.

The more elaborate type of food web, (such as that in Figure 36 for coral reef fish) cannot be constructed without a great deal of experimental data which, for practical reasons, cannot usually be collected in the earlier stages of fishery investigations that principally concern us here.

Conceptually, it helps to integrate the ecological aspects of an exploited system with the fishery it supports, if we view both from the perspective of the common ‘dissipation structure’ which radiates out from those areas where energy production tends to be concentrated, and where it is passed upwards in the food web, as well as dissipated over a wider area. It is then logical to view the linkages or transactions the biomass is subjected to, from capture until it is consumed, as part of the same structure. Such an approach has the advantage that other alternative or conflicting uses of the same area can be compared directly with each other.

Interestingly enough, when the fishing industry is viewed from this perspective, the number of linkges or transactions that follow upon capture, confer added value to the product, as well as providing employment in the secondary and tertiary sections of the industry. In human terms therefore, adding to the complexity of the resource ‘dissipation structure’ seems to be a worthwhile activity. Figures 30-32 can therefore be viewed as components of the upper levels of the ecosystem dissipation structure and its fishery that supports this human activity.

Finally, we should note that the functional relationships between components of the fishing industry, and the constraints that apply to the industry ashore as well as at sea, also affect the fishing pressure and hence the biomass, and age and size composition of a species, as well as its resistance to competitors, ecologically speaking (e.g., Figures 40, 41 and 61). The flow of materials to man, the principal apical predator in most contemporary marine ecosystems, could thus be usefully followed through in a more or less continuous fashion from the primary production to the harvesting and consuming sectors of the fishing industry.

Portrayal of life history stages in food webs

In an ideal world, we should attempt to construct food webs that are valid for all life history stages. However, the multiple linkages (or even reversal of the direction of flow of materials) between different components in different stages of their life histories) are difficult to represent graphically. A few suggestions can be made however, which may have some value if we can regard those interactions that occur in the larval or premetamorphosis stage as separate from those occurring when a species enters its adult habitat(s), where it eventually becomes vulnerable to commercial capture. These suggestions do not belittle the importance of interactions at the larval or early juvenile stages (in fact there is growing awareness that events in the early life history control the recruitment to the adult fishable stock -see IOC Workshop Report No. 28, and the reports of the Costa Rica meeting: Sharp and Csirke, 1983, for a review of recent work on this topic). What is quite clear, however, is that:

1. learning about species interactions in the larval, or even juvenile stages, especially in the tropics, will require quite different and very extensive commitments in research effort to elucidate them; only the first of these will be a major new problem of species identification for these early life history stages;

2. in situations where plankton diversity is high, the biomass of larval or juvenile individuals of a given species that is included in the energy budget of a predator is likely to be almost impossible to measure, and significantly less important than the impact of the predator on the prey recruitment size;

3. these interactions may occur over a relatively short period of time, in patchy distributions scattered over extensive areas, which makes their quantification difficult; and

4. finally, there is a significant body of opinion suggesting that starvation, due to “mismatch” in time and space of larvae with patches of abundant food, rather than predation, may be a key factor in determining recruitment success (see however the Annex).

These points suggest that although individual (larval) prey items can contribute significantly for relatively short periods to the diet of predators, they should perhaps be regarded in many cases as being taken incidentally to generalized feeding on the planktonic community, by carnivores which extract their food from the plankton in an opportunistic fashion for most of the year.

It may be suggested as a convention therefore, that the impact of the morphologically “adult” stages (i.e., those stages occupying the adult niche(s) in the ecosystem, even if they are not sexually mature), on larval stages of other organisms, be represented by a dashed line. Thus, although adult cod feeding on herring would be represented by a solid arrow on a trophic diagram, the latter also act as predators on cod larvae between the time of cod hatching and metamorphosis, which would be represented by a dashed line. This clarifies that the concept of a food web as a way of illustrating trophic interrelationships can be usefully extended to include information flow of all kinds between participants in an assemblage and not just movement of biomass. This is also shown in Figure 36 from Smith and Tyler (1973), which is a compartment model illustrating that species may be grouped by mode of feeding instead of food type; (i.e., adaptations for exploiting the same food in a different form exist, and are one of the reasons for the complexity of some marine food webs). From this figure we can see that other interactions, e.g., symbiotic associations and competition for space, may also be of importance, and can also be readily illustrated.

Compartment models attempt to simplify the complexities of food webs by including all elements occupying the same place and having the same function in the food web, in a common ‘compartment’, so that the same transactions are supposed to take place for elements combined in this way. Figure 37 from Jones (1982a) shows some simple examples, which in this case are intended to bring out those elements of interaction which are common to planktonic food webs. Used in this way, diagrammatic representations are obviously able to communicate a great deal of information in a very economical fashion.

The use of compartment models in simulating ecosystems has, however, led to criticisms (e.g., Mann, 1982), particularly since if individual species are condensed down into broad categories, e.g., “browsers”, ‘secondary carnivores’ etc., this can minimize the member of linkages. This, in turn, directly affects ecological efficiency estimates and also because the time and space scales and the frequency of nutrient cycling, are so different at different levels in the system, as we have discussed elsewhere in this report.

omnivory in zooplankton

An important adaptation is omnivory Omnivores have the capacity to switch feeding from one trophic level (A) to another (B) Seasonal variations in Omnivory could lead to significant variations in the grazing pressure on the herbivore component of the zooplankton.

Omnivores as herbivores

As long as omnivores act as herbivores (Flow A) there is a maximum of grazing pressure on the phytoplankton. This depicts a situation of maximum grazing pressure on phytoplankton and maximum tendency for zooplankton biomass to increase.

Omnivores as carnivores

If omnivores act wholly as carnivores (Flow B) grazing pressure on the phytoplankton should be reduced to a minimum. This depicts a situation of minimum grazing pressure on phyoplankton and minimum tendency for zooplankton biomass to increase.

Figure 37 From Jones (1982a) showing how various aspects of trophic interaction can be easily represented in diagrammatic form

Despite this critique, compartment models are coming to be used for preliminary investigation of the likely influence of a change in fishing strategy on potential yields, and for organisms that do not differ too greatly in size, scale and turnover rate, they perhaps represent a reasonable way forward at present. For example, Sheridan (1984) investigated the relative impacts of partially eliminating discarded (dead) fish and small shrimp by-catch on the penaeid shrimp biomass, compared with elimination of fish and small shrimp (live) on bottom by the use of selector trawls. In attempting to weigh the loss of food for shrimp in the first case, against the effect of increased fish predation on shrimp in the second case, the model allowed some broad generalizations to be made that would certainly be of use to managers.