Uncertainty in fisheries management: sources and consequences.
Jon Nevill 8 July 2008.
“Worse, some scientists involved in our discussions were worried about the very notion of publicly admitting uncertainty, and felt that it was important to maintain at least the appearance of consensus within the scientific community” Carl Walters (1986:343).
Introduction:
A well-known fishery scientist, John Gulland, once said that “fishery management is an endless argument about how many fish there are in the sea, until all doubt has been removed – but so have all the fish”[1]. Although factious, this comment does highlight the importance of uncertainty – and the quote unfortunately contains a certain amount of truth.
Fishery scientists provide advice to fishery managers. Fishery managers (within a governance framework created by national statute, international law, common law, and culture) provide and enforce society’s controls over fishers. Controls include both restrictions and incentives, and operate within the culture of various groups, as well as the prevailing national and international legal and economic framework.
The most basic scientific advice concerns the size of species stocks, and the effects on those stocks of harvesting pressures. In addition, predictions need to be made of the continuing ability of the ocean’s wider ecosystems to support individual populations.
Uncertainty in fisheries management stems mainly from seven sources:
- imperfect understanding of the oceanographic drivers of ecosystem function and species behaviour, reproduction and growth within ecosystems;
- imperfect understanding of species-specific biology, including growth rates and drivers, and movement patterns driven in part by feeding and reproduction;
- imperfect understanding of the behaviour of species within ecosystems;
- stemming from the above and from the practical limitations involved in expressing biological processes in mathematical form – imperfect predictive models of species biology and ecosystem function;
- imperfect stock sampling and other empirical data, which seeks to understand the size, movement, growth, mortality and genetic diversity of stocks of fishery target species, and to provide oceanographic and ecosystem-related information;
- systematic but poorly-appreciated bias in scientific advice and managerial decisions resulting from the cultures in which these groups operate; and
- imperfect prediction of fisher behaviour, including movement, fishing effectiveness, and ecosystem damage (primarily from gear damage, bycatch and discards).
A significant part of fishery science attempts to understand uncertainty, and to minimise it where possible. Where uncertainty cannot be eliminated (and this is always the case in capture fisheries) its implications for management decisions need to be understood and taken into account. Powerful tools exist for the management of uncertainty.
Likelihoods, or probabilities, can be attached to some uncertainty elements. Where this is the case these uncertainties may be termed “risks” – and this terminology is used in this chapter (although not throughout the thesis, where a more relaxed meaning of “risk” is used in line with common language). The term “uncertainty” within the chapter refers to lack of certainty where the probable bounds of the uncertainty are not quantifiable.
The purpose of this chapter is to provide a brief general discussion of the sources and consequences of uncertainty in fisheries management, to lay the foundations for the following chapter containing a detailed discussion of uncertainty management techniques.
Ocean ecosystems in a nutshell:
Almost all ocean ecosystems depend (directly or indirectly) on the energy of sunlight, which enables phytoplankton, algae and other plants to synthesize carbohydrates, using dissolved carbon dioxide and nutrients. Carbohydrates provide the energy source for resulting food chains, supporting a wide range of grazers, detritivores and predators. Photosynthesis (both marine and terrestrial) also produces oxygen, which enables the metabolic processes on which the grazers and predators (indeed all aerobic organisms) depend.
The energy of sunlight also drives the oceanographic processes which, broadly, move organisms, nutrients, and dissolved oxygen around the planet through winds and ocean currents. The primary driver for these processes is the energy differential (manifested as temperature) between the equator and the poles. This energy differential, through creating changes in density in air and water, drives the weather systems of both the atmosphere and the ocean. Ocean ‘weather’ can be every bit as variable as atmospheric weather – noting however that the time-scales on which changes are felt are orders of magnitude slower, particularly with increasing depth. Life in the deep ocean depends on dissolved oxygen ‘pumped’ from polar regions by the ocean’s thermohaline circulation. The major oceanic currents, at different depths, form ‘conveyor belts’ capable of moving water masses around the planet – connected, importantly, by the Southern Ocean. With these water masses travel organisms, nutrients and oxygen.
Oceanic surface waters, distant from continental erosion, are usually depleted in nutrients, the result of a continual ‘drain’ of nutrients to deeper water through detritus falls. Although the drain is slow, and nutrients are recycled within the uppermost few hundred metres, the lack of nutrients is often (usually) sufficient to restrict the growth of phytoplankton, and thus to restrict the abundance of organisms up the food chain. Where strong, deep, nutrient-rich currents are pushed to the surface by continental shelves (for example on the west coast of South America, and the west coast of South Africa) the resulting increase in surface nutrients can support high densities of marine organisms. This productivity may in turn change dramatically with changing ocean climate – for example the abundance of Peruvian anchovetta is strongly influenced by the La Nina / El Nino cycle (Pontecorvo 2001).
Even the deep ocean, seemingly isolated from the obvious variability of the ocean’s surface layers, can manifest striking changes in organism abundance over periods as short as a decade (Koslow 2007). The ocean is also a large place – covering nearly 70% of the planet’s surface, with a volume several times the volume of the land which rises above sea-level. By the close of the twentieth century, few reliable time-series of even basic data such as temperature, salinity and nutrient levels in the deep ocean had been collected, in spite of the importance of deep and intermediate currents in controlling ocean climate and productivity. While data collection has been expanded over the last decade, accurate ocean forecasting programs are only now being developed.
Today, while the major physical features of ocean climate and weather are understood, variations within timescales of months to years still remain, to a considerable extent, unknown and unpredictable[2].
Harvesting renewable resources:
Although is was long recognised that fishing pressures could deplete or extinguish freshwater populations, at the close of the nineteenth century is was widely believed that fishing would have little, if any, impact on marine populations. However by the 1930s many instances of marine overfishing had been observed, and fisheries managers looked to basic concepts of fish biology to guide harvesting controls.
The biological basis for all sustainable harvesting is reproductive surplus. Natural populations in favourable environments generally have the capacity to expand in size up until a point where environmental factors limit further growth – through for example competition for food or space. At this point the population, and the biomass of the population, will remain stable if steady state conditions prevail – the population has reached the ‘carrying capacity’ of the environment under the prevailing conditions.
The concept of ‘maximum sustainable yield’ (MSY) rests on the idea of finding a level of harvesting which will maximise the reproductive surplus. The surplus may be expressed as the population size balanced for births minus deaths, times an average individual growth rate, within a given period. Growth rates may be expected to increase as harvesting commences, releasing the population from the density-dependent limitations mentioned above. Most fish do not grow continuously over their adult lifespan, with the result that the capacity for physical (somatic) growth is generally most pronounced in juveniles and young adults – another factor influenced by the removal of the larger, older fish from a ‘virgin’ population. Fishing activities, under this understanding, aim to harvest the reproductive surplus at a rate which produces a maximum yield for the fish stock, or population – while not reducing the population itself below a level which insures its long-term viability (Hilborn et al. 1995).
An important element in the theory underpinning the use of MSY relates to ‘compensatory mortality’ – or loss of ‘surplus’ juveniles prior to their recruitment into the adult population. Many fish, particularly pelagic species, have reproductive strategies where females produce large numbers of eggs. Most of the resulting offspring die, either from predation or environmental factors, before reaching reproductive status as adults. Thus for many fish populations under favourable conditions, net recruitment is more or less independent of parental egg production over a considerable range of parental abundance (provided viable subpopulations remain to maintain the population’s spatial distribution). This effect initially gave fishery scientists confidence that the spawning biomass of populations could be substantially reduced without markedly reducing the rate of recruitment[3]. However confidence grew to overconfidence – for example contributing to the demise of the northern cod (Walters & Maguire 1996).
The pursuit of MSY, generally expressed in the form portrayed in Figure 1, assumed increasing prominence in fisheries management from the 1930s to the 1950s. For populations with rapid growth rates and powerful reproductive strategies (short-lived pelagic species, for example) the theory seemed at first to work reasonably well. Under basic assumptions of logistic growth, information on growth rates obtainable from the weight-at-age distributions of catches as fishing progressed, combined with assumptions on natural and fishing mortality, could be used to estimate the approximate current location of the fishery on the catch/effort curve. Logistic models of fish population growth (while admittedly simplistic) show that maximum growth is located when the virgin population has been reduced to half its original level[4]. However key assumptions within the theory are not reliable in practice[5] (more below). By the late 1950s the limitations of the MSY approach were apparent, resulting in moves towards an ecosystem-based approach (Beverton & Holt 1957) at least amongst fishery scientists if not fishery managers. Unfortunately the absence of straightforward and palatable alternative methods for establishing catch limits has seen the MSY concept continue in use long past its use-by date. Another factor which probably contributed to delays in abandoning MSY was that managers could actually understand the theory – while the more complex modelling methods remained obscure, and untrusted.
FIGURE 1. The catch / effort curve. From Bonfil (2002)
An important difficulty in estimating the MSY of a given population relates in part to obtaining accurate information on catch and effort. Much of the immediate action within a fishery is beyond the direct observation of fishery scientists.
Given the inaccuracies in basic data, and the limitations of deductive, observational and experimental methods, “if we wish to learn how intensively a population can be exploited, we can find its limits only by exceeding them… This is perhaps the central problem with MSY: you cannot determine it without exceeding it.” (Hilborn et al. 1995:54).
In summary, sustainable harvesting of a renewable resource rests on harvesting the population’s reproductive surplus. This surplus does not exist in a natural population under steady-state conditions, and the population must be ‘fished down’ to release the surplus. Fishing down to find the point of MSY inevitably results in over-exploitation of the population.
Over-exploitation of populations has historically resulted in the depletion of many stocks by an order of magnitude or more (MEA 2006), and the extinction of some subpopulations (Sadovy & Cheung 2003, Casey & Myers 1998, Tegner et al. 1996). Yet, until recent decades, the effects of these changes on the ocean’s ecosystems has been all-but ignored by fishery managers.
Sources of uncertainty – oceanographic drivers:
Coastal, island and estuarine marine ecosystems are manifestly complex, with highly variable seafloor topography and biology-mediated habitats, such as coral reefs and seagrass meadows. However even the seemingly simple ecosystems of the open ocean are complex.
Several key determinants of biological function vary with depth, diurnal and seasonal cycles, and ocean dynamics: sunlight, temperature, dissolved oxygen, and nutrient levels. Radiant energy falling on the ocean varies with latitude, season and cloudiness, and the clarity of the water effects the degree to which sunlight penetrates below the ocean surface, thus influencing the photosynthetic abilities of phytoplankton. The open ocean is far more complex, physically, chemically and biologically, than it seems.
The ocean has its own climate and weather, just as the atmosphere does. While the key features of the atmosphere vary over hours and days, and more slowly over years, the larger features of ocean climate vary over weeks and months, and more slowly over years and decades.
Currents change in speed and direction, and create eddies at different spatial and temporal scales. The larger eddies, ocean gyres, are often semi-permanent features of the open ocean, varying however in exact location and strength. Gyres may create upwellings or downwellings, also a feature of major fronts created where large water bodies, driven by the major ocean currents, meet. These features may deplete or enhance nutrient levels, with consequent implications for phytoplankton and micro-organisms which form the base of ocean food chains.
The seemingly featureless open ocean also responds to seafloor topography. Seamounts in the path of ocean currents create downstream eddies, which can also move nutrients vertically, creating patches of enhanced productivity. These areas act as attractors to the larger inhabitants of the ocean – a fact long recognised by fishers.
The physical and chemical properties of the open ocean are variable in space and time, demonstrating both systematic and stochastic patterns. Ocean ecosystems respond to these variable drivers. Massive phytoplankton blooms come and move and disappear. Schools of ocean grazers and predators track these blooms, aggregating and dispersing over different spatial and temporal scales. Some, like jellyfish, move largely with the ocean’s currents, while more mobile creatures display massive latitudinal and vertical movements. What has been dubbed the world’s largest migration – the vertical movement of marine organisms between the epipelagic and mesopelagic zones – occurs on a daily cycle. All these movements are primarily focused on feeding, avoiding predators, or reproduction – activities which are central drivers of the behaviour of most marine organisms.
Both the open ocean and shallower seas (the continental shelves and coastal bays and estuaries) are complex even in terms of their simple physical and chemical parameters, a complexity which is magnified through biological processes. This complexity, poorly measured and understood, introduces much uncertainty into scientific predictions – and thus to management decisions.
Sources of uncertainty – species biology:
The theories underpinning fisheries management over most of the twentieth century depend on assumptions which seldom hold reliably in practice: (a) fish populations may often exist in unfavourable environments, with subsequent impacts on growth, recruitment and mortality; (b) even in relatively favourable environments, the logistic growth function does not hold, particularly with larger, slow-maturing fish; (c) fish populations are seldom in steady-state within their supporting ecosystems; (d) spawning stock and recruitment levels may be positively correlated within the range of fishery-induced population reductions, and (e) measurements taken from sampling commercial catches may not accurately reflect the state of the population at large.
The breakdown of core assumptions supporting methods commonly used to set single species harvest targets results in considerable uncertainty. Estimates of growth rates and mortalities (natural and fishing) are thus subject to considerable error, which is often difficult or impossible to quantify. Consequently estimates of current spawning biomass, virgin biomass, and ‘safe yield’ also contain substantial error margins, which are often not fully acknowledged by fishery scientists themselves, and consequently not properly taken into account (or even ignored entirely) by fishery managers.
A single species within a large marine ecosystem (and perhaps within the jurisdiction of a fishing nation) may contain considerable genetic diversity. This diversity is often (usually) spatially distributed. The population may be made up of relatively isolated local populations, or (perhaps likely in purely pelagic stocks) a single dispersed population. However, for coastal species (and species with coastal life-stages) it seems likely that the stock will be made up of distinct sub-populations which together comprise a larger metapopulation[6] (Kritzer & Sale 2006). Even for purely pelagic species, adherence of sub-populations to ocean topography or oceanographic features could produce a metapopulation structure.