You are on page 1of 10


Sustainable exploitation: a review of principles and methods
William J. Sutherland

Sutherland, W.J. 2001: Sustainable exploitation: a review of principles and meth-
ods. - Wildl. Biol. 7: 131-140.

Although the main theoretical framework determining how to exploit popu-
lations was derived almost 50 years ago, overexploitation is common. I review
10 major concepts underlying the regulation of exploitation: population
increase can be exploited; density dependence is essential; quantifying densi-
ty dependence is exceedingly difficult; sustainable exploitation involves reduc-
ing population size; population growth rate is usually mismeasured; sustain-
ability has many conflicting definitions and the choice depends upon the
objectives; it is better to monitor the population than the harvest; quotas are
unstable; increasing effort is simple, reducing it is painful; exploit conserva-
tively. I then give a brief account of each of the nine main methods that are used
to determine sustainable exploitation and the uses, advantages and limita-
tions of each. The nine techniques are: surplus production models, yield per
recruit models, Robinson and Redford model, linking yield to recruitment and
mortality, adjusting to population changes, comparing demography across
sites, reducing to a fixed fraction of unexploited population size, full popula-
tion models and adaptive management.

Key words: density dependence, exploitation, harvesting, hunting, sustainable

William J. Sutherland, School of Biological Sciences, University of East An-
glia, Norwich NR4 7TJ, UK - e-mail:

Almost half a century ago the main concepts underly- exploitation there are many success stories in which pop-
ing sustainable exploitation were devised in a series of ulations are currently reasonably healthy even with in-
remarkable papers (Schaefer 1954, Ricker 1954, Bever- tensive exploitation. Examples include the moose Alces
ton & Holt 1957). These pioneering studies provided alces, the South African fur seal Arctocephalus pusil-
the framework for a series of sophisticated models and lus and most goose populations in North America.
methods of analyses that make current fisheries man- There are two objectives in writing this review. In pre-
agement a highly advanced process (e.g. Hilborn & paring this review I was struck by the fact that the lit-
Walters 1992, Quinn & Deriso 1999, Jennings, Kaiser erature is very fragmented: papers on fisheries, mam-
& Reynolds 2001). However, despite this understand- mals, birds and forestry all consider similar issues but
ing, the last 50 years have seen considerable overexploi- each uses a different terminology and seems devel-
tation of numerous species (Ludwig, Hilborn & Walters oped largely in isolation. My first objective has been to
1993, Casey & Myers 1998). Even many intensively identify the essential principles of sustainable exploita-
studied populations managed by affluent countries have tion and reviewing the main issues.
collapsed, often resulting in local unemployment (e.g. My second objective is to outline the main methods
Walters & Maguire 1996). used for determining levels of exploitation and describe
Against this generally pessimistic landscape of over- the benefits, problems and uses of each.
© WILDLIFE BIOLOGY · 7:3 (2001) 131

to age 3 was 32-63 times larger than in unexploited areas. Sustainable exploitation involves reducing 3. McGarvey (1996) noted that pop. as sustainable fact that most data sets start relatively recently and so exploitation is dependent upon a growing population. Sauer to have persisted in the presence of sustained exploita. Wood & Stewart 1992. such as an introduced population or a population whose Goss-Custard & Sutherland 1997) but this also requires food supply has continually improved. Density dependence is essential is through behaviour-based modelling in which the There are very rare circumstances in which population decisions made by individuals are determined and in- shows a persistent increase over a reasonable period. Nt+1/Nt) against the I suggest that these are the 10 fundamental principles number in the first year (Nt) and a negative relationship of exploitation.g. However. As an example. & Otten 2000). If then the resulting increase may be exploited. The 132 © WILDLIFE BIOLOGY · 7:3 (2001) . 4. then exploited populations must be lower. population size ly difficult A common error amongst conservation biologists is to Although density dependence is fundamental to sus. Studies of den- sustainable exploitation (Ricker 1954. is a much bet- removed while keeping the population constant.g. increasing by 5% per annum then that 5% may be such as breeding success and mortality. (Caughley & Gunn 1995). White & Burnham 1998). Francis. exploitation and natural mortalities are additive then lows that density dependence is absolutely central to exploitation will reduce the population. There is whether exploitation mortality is compensatory or addi- thus not an increase that can be exploited. If there is strictly compensatory as a result of density dependence.Fundamental principles of exploitation most widely used approach is to plot the rate of increase over two successive years (i. density dependence then removal of individuals does not lation results in increases in breeding output or survival. The best means is through experimen- stochasticity (Lande. but this is obviously usually impractical. if the population is 1998). affect the numbers remaining (the 'doomed surplus'). Cappuccino & Harrison 1996). Population increase can be exploited the presence of the same variable on each axis can The essence of sustainable exploitation is the exploita. The ability of so many species Synatzske. show that exploitation has reduced a population below tainable exploitation. Quantifying density dependence is exceeding.e. Oldenburg & Guthery 1995. if there is any measurement error. thus most esti- tion of the population at the rate at which it increases mates of density dependence are flawed (Shenk et al. Hellgren. satory at high population sizes and tend towards being gellanicus persisted despite long-term intensive exploita. for naturally occurring unexploited populations not to There is widespread debate (e. However. corporated into a game theory model (Sutherland 1996.g. reducing the popu. the the population is overexploited. In prac. Harris in press) and mortality caused by tion is good evidence for the ubiquity of density depen. less excusably. & Serie 1998. often and as continuous population growth can only be change methods without evaluating the consequences) achieved by reducing populations to take advantage of and due to difficulties in isolating density dependence the density dependent increase in survival or breeding from population fluctuation resulting from habitat output. exploitation is almost always both compensatory and dence. There are four main approaches to detecting density dependence and each has considerable problems. It will generally tend towards being compen- ulations of Georges Bank sea scallops Placopecten ma. ter method but it is necessary to measure all components tice the actual rate will be below this. Schaefer 1954) sity dependence give no support for the idea of pure com- and without density dependence exploitation would be pensation (Dusek. additive once the population is at a low level (e. if tive to natural mortality. Estimating all major components of fitness. Nicols 1991) as to show increases over long periods of time. However.g. as there almost always is. result in spurious density dependence. but the norm is considerable data and understanding. it is very hard to measure as a result its unexploited level and then use this as evidence that of sampling errors (Shenk. Engen & Sæther 1995). for example due to estimate the strength of density dependence and this to the interaction with demographic and environmental is rarely practical. result in short time series (and. tal manipulations (e. then 1. additive. Bart- tion and in exploited areas the survival rate from eggs mann. comparable to mining. is evidence for density dependence. White & Carpenter 1992). even change or variation in weather conditions (e. It thus fol. Dennis when exploited sustainably. Thus. A final approach 2.

6. It will be an overestimate of the exploitation ter exchange of information between exploiters. Quinn & Deriso 1999. to exploit at a lower intensity than at MSY. there is a natural variation in the pop- starting a programme of exploitation during which the ulation size and it is often difficult to reduce agreed quo- population has to be initially reduced to provide growth tas. Such estimates are likely to produce MSY. population size and changes in exploitation methodol- fused. The second common error is to measure the maximum possible growth rate. Of course careful monitoring can pre- However this could be at very low population sizes (with vent such overexploitation but in reality monitoring is low yield). Quotas are unstable There has been considerable debate about the definitions Removing a fixed number of individuals is theoretically of sustainability (Bennett & Robinson 2000). the yield is should be no growth. Although some combinations of exploita. Furthermore. esti- principle 4 is that there has to be some reduction for mates of a sustainable quota are faulty or the quota is exploitation to occur. efficiency. Despite these criticisms of higher than zero. Munholland & Scott 1991). MSY is not the most profitable point. The growth rate at very low densities is only use. With open there is considerable confusion over what should be access the theoretical expectation is that individuals measured (Sutherland 2000). In practice r and rmax Although it is usually easier to measure changes in the are used interchangeably and r is also often applied to numbers exploited. the variance in growth rate is thus even harder. The main acceptable if the level is set correctly and the popula- definitions are: (i) that it does not significantly affect the tion is stable. Sæther & Engen 1997) as it is the population size that really matters. en quota will become an increasing proportion of those tion rate and population density result in extinction. stochas. is likely to give inflat. such as zoos. One major criticism is that incorpo- rating costs of exploitation into this model shows that © WILDLIFE BIOLOGY · 7:3 (2001) 133 . means the population is being exploited sustainably or ed results. then a giv- production. exploitation. yet it is. many remaining which can drive the population further down- combinations are in theory stable and allow regular wards (Walters 1986. better transport or bet- densities. However. Lande. Walters 1992). as a result of meas. A repeated story is for there to be a resistance to and thus by this definition is not sustainable in the reducing quotas such that once a lower quota is even- short term. This may either be the growth rate 7. populations fluctuate wild population (IUCN/UNEP/WWF 1980). A second. I believe it is still a very useful concept as it pro- estimates of exploitation that are well below the level vides an invaluable reference point as an ideal against that could be sustainable if the population was reduced. much lower than it could be and the exploiters make urement error. (for example due to varying weather conditions). The concept of MSY has been even the reduced quota cannot be caught. (iii) that the maximum sustainable yield tually agreed on the stock has collapsed so much that (MSY) is not exceeded. In practice. Determining changes in population size is tic (Dennis. then this will be excluded from the estimates of a deterministic parameter. problem concerns usually difficult. The growth rate under if the number removed per day is constant it either ideal conditions. of course. The first is to measure population growth rate als will join until the profits from the yields balance the for an unexploited population in which case there costs. (ii) that exploitation balances illicitly exceeded. ogy. less important. if illicit exploitation is taking The population growth rate is easiest to measure as place. It is better to monitor the population than the at which there is no competition (r or λ) or the growth harvest rate under ideal conditions (rmax). At this point the population is low. Sustainability has many conflicting definitions and the choice depends upon the objectives 8.5. With just one ured exploiter who has complete control over the level of total Principle 1 is that the population growth can be exploit. heavily attacked. the estimated growth rate will often be little profit (Hardin 1968). Thus rate for higher population sizes. However in reality. Population growth rate is usually mismeas. Changes in methodology may be subtle. This error will lead to exaggerated estimates it is decreasing but this is compensated for by increased of sustainable exploitation. However. exploitation as costs increase with effort it is sensible ed so it is thus important to assess growth rates. There are two common will increase exploitation levels and further individu- errors. such as ful if the population is to be exploited at the same low new paths cut through a forest. Hilborn & exploitation with the population persisting at that size. which current practice can be compared. this measure combines changes in a range of densities so this subject is horribly con. If a population declines. but measuring better for adjusting the exploitation level (Walters 1986.

However. to result in a reasonable catch per unit effort but this may Individuals often have an economic. Developments include virtual pop- tions (Table 1). Caughley & Sinclair because they have more biomass or because of their (1994) suggest a 25% safety margin and higher with vari. Hence a high effort is likely in the short term reducing the catch. Fuller & Kehler 2000). 1997). This approach can be simply tackled using a spread- sheet and data on age-specific natural morality and age-specific value and exploring the yield from differ- Main methods for exploiting populations ent exploitation strategies. Humans are usually risk adverse and the on the numbers or biomass exploited each year and the reason for this is that the utility (the perception of val. but have For a wide variety of reasons including the uncertain. Sæ. a giant tortoise hostage (Anon 2000). Increasing effort is simple. In ploitation. to determine how to exploit cohorts when in- the likelihood of driving populations towards extinction. Exploiting These include only restricting effort rather than quotas. painful than gaining the same sum is pleasurable. poor data or irregular monitoring. in which the age structure of the catch is used to back calculate the natural and exploited mortality. issued death threats. Engen & Lande 1996) and the difficulties of reduc- ing harvest levels (principle 9). The old size (Lande et al. to analyse moose exploitation (Courtois & Jolicoeur population fluctuations (Beddington & May 1977. guns. reducing it is Surplus production models painful The surplus production models (also known as surplus The history of exploitation shows that populations yield models) come in a wide range of versions but are often continue to be overexploited even when reducing all based on the original idea of Schaefer (1954). social or polit. although individuals are very happy to take advantage The simplest versions assume the population is at equi- of higher catches they are often deeply unhappy about librium. As an example the decline in catch per unit effort unless the stock has currently in the news. effort. It can be difficult to determine ical commitment to the current levels. 1993). is impossible to correct for all of these. although no one considers this to be sustainable. Hence There are serious problems with the simplest versions. burnt a research station. greater trophy value. when young produces the most individuals but delay- only exploiting populations when they exceed a thresh. but it ster quota has been increased from 50 to 80 tonnes. I will provide a brief summary of each ulation analysis (VPA) and cohort analysis (Pope 1972) with comments on their actual and potential application. roads. losing money is much more the maximum yield. Exploit conservatively ly used in fisheries (e. for example one has been used ties of the biology. setting exploitation at Yield per recruit models the level which is calculated to be sustainable is likely Larger individuals are usually more valuable. well be unsustainable. ther. and exclusion zones in either weight or value) for each recruit. rotational management in optimal solution balancing these conflicting phenom- which areas are exploited for a period and then left ena can be calculated by determining the yield (as (Myers. there are sophis- ticated models and software for analysing populations There are nine main methods of exploiting popula. the number of fishermen on the been heavily overexploited (Hilborn & Walters 1992) Galapagos Islands has doubled from 1999 to 2000 so in which case it is difficult to reverse the overex- that the lobster quota was filled in four months. blocked The other major problem with this approach is that roads. vehicles. More sophisticated versions are wide- 10. 1993). with these methods. The simplest version of the surplus production mod- el is rarely used. Beverton & Holt (1957) developed able populations. Quinn & Deriso 1999). Thus for hunt- destroyed the islands telephone antenna and even held ing there might be improvements in information. it is very difficult to standardise effort. yield per recruit models. Thus for most people. also known as dynamic pool A number of means have been suggested that reduce models. which fishing is not allowed (McCollough 1996).g.9. Such an approach requires considerable data but is widely used 134 © WILDLIFE BIOLOGY · 7:3 (2001) . ammunition or field techniques. The effort would result in a greater long-term yield (Ludwig most simple version consists of collecting information et al. As a result the lob. and then plotting the catch per unit effort against ues) decreases with each additional sum of money effort to determine the level of effort which produces owned. corporating growth and natural mortality. ing exploitation results in more valuable individuals. response the fishermen have threatened tourists. difficulties in estimating parameters. either to result in population collapses. applications elsewhere.

Not nec- expected popula. Effort needs sions available. Yield per recruit Based on age Age-specific nat. specific catch. ronmental tion lation changes. trates on popula- tion sizes. fish populations. tude and conse. available to value yield. Linking yield to Exploits below Recruitment rate. graphic esti- mates. Main methods specific growth ural mortality Full considera. game exploita- relation to popu. most successful respond to envi. errors in demo- ceeds mortality. tle data on demo- ronmental graphy. Can often measured mammals and rate. sophisticated ver. effort. quences of envi. Ideally magni. changes. Loosens/tightens Population Probably the Difficult to Widely used for ulation changes regulations in changes. per year and data). Adjusting to pop. apply to fish Incorporates Either size of populations.g. method. Does not apply to popula- tion in equilibri- um. determines con. Software results in highest variation in age. tality determines value. and age-specific tion of increase mortality and forestry. e. Requires very lit. but rarely able to apply to is. Calculates annual birth rate. precise. One ver. apply when there at wrong density. moose. Requires consid. Sophisticated. effort.Table 1. for trends or to arise from ly for mammals mortality recruitment ex. the weaknesses and the main uses Method Outline of method Data required Strengths Weaknesses Uses Surplus Determines how Simplest version Can be carried Overexploiting is Mainly fisheries production catch varies with requires data on out with very lit. yield if popula. the strengths. Main methods used for assessing exploitation with brief details of how the technique is carried out. More assume equilibri- ferent strategies.g. and recruitment. the data required. ised. age at last repro. tle data (e. able if yield < MSY. um. to be standard- Software avail. recruitment to recruitment or determine long relationship term sustainable between stock strategy. the easiest way of but also others effort and then numbers or bio.6 of duction data. in value with recruitment. strategy that sion uses annual age. No data needed Estimates likely Used occasional- recruitment and rate at which Mortality rate. Age at first Often the only Growth rate Tropical forest Redford mum growth reproduction. erable data on for fisheries and and natural mor. Concen- changes. © WILDLIFE BIOLOGY · 7:3 (2001) 135 . Only applies to populations at sustainable popu- lation sizes. are no historical Unlikely to be tion is at 0. method. mass exploited catch and weight Simple versions sequences of dif. assessing MSY. essarily sustain- tion size. Robinson and Calculates maxi. birds.

the consequences examines conse. Adaptive man. Requires esti. levels being cor- logical studies. ponents of the density depen. Thus fisheries managers often carry out surveys of eries science programme. One sensi- ble but expensive solution is to carry out surveys of the Robinson and Redford model number of young individuals so that the exploitation can Exploitation often takes place on species about which be adjusted according to the size of the recent cohorts. related with envi- ronmental fac- tors. Difficult to carry be part of any consequences of out. Can be ble. version results in across sites density or popu. almost nothing is known (Johannes 1998) and this is par- 136 © WILDLIFE BIOLOGY · 7:3 (2001) . of other changes. population size. Should doubts occur and edge of system. Sinclair & Stefánsson (1997) dard approach has been to assume that the recruitment showed that as North Sea cod Gadus morhua have will be constant. This assumption has lead to some dra. matic examples of fish population crashes. population. quences of differ- ent harvest lev- els. Rarely used but fixed fraction of lation at a pro. Uses models to Data from exper. used to examine population and dent processes. exploitation pro- these doubts. use for variable deficient popula- tion. ed population Concentrates on mate of unex. Continues forev- er improving knowledge and management. Can be used with tion. Another approach is to determine how the forestry in which growth and mortality is assessed in recruitment varies with the adult population (stock- plantation blocks or by following groups of marked trees recruitment relationships) in order that the population within natural forests. by fisheries biologists in countries with an extensive fish. portion (e. size.g. Environmental very little data. gramme. Thus Cook. ploited popula. Runs experi- ments to reduce uncertainty. tions. change) at a the short term facts resulting concerning ex- range of sites. Relatively easy. ulation size 60%) of unex. Used too infre- agement determine where iments. strength of all method. improves knowl. often part of oth- unexploited pop. Full population Creates full mod. quently. Exploitation lev. even once the adult population has been declined the recruitment has also declined leading to the markedly reduced. Difficult to Useful for data ploited popula. data if possible. er models. Full data on Probably the best Almost impossi. recruitment. Grey partridge model el of major com. out detailed eco. cult to explain. population be carried out in founded by arte. Can analysis not con. Focuses on the Need to ensure A less rigorous demography tion intensity to el and density (or population. Current population size. Continually Politically diffi. without carrying from exploitation ploitation levels. A stan. Likely unexploit. a rule of thumb lation change. potential of accelerating declines. can be managed to reduce the problems of affecting the This approach is based on a given sized cohort.Table 1 continued Method Outline of method Data required Strengths Weaknesses Uses Comparing Relates exploita. A similar approach is used in young fish. Reducing to a Maintains popu.

2 for long-lived species be either due to overexploitation (illustrating long-term (age at last reproduction is over 10 years).4 for short. fish and or from the field. Perez & Satroviejo 1995. It may thus be more realistic to assume that MSY occurs at 0. This seems a trivial point but actually it is crit. Walters & Ludwig 1995). One is that the estimates of den. There are a number of prob. ened if the population is declining and relaxed if it is duction and breeding per year it is possible to determine increasing. However. yet this is population size rarely measured. This can be very simple yet is extremely effective and ed. moderate and heavy). There is a need to ing the population size. age at last repro. if the population tionship between population growth rate and population drops to 110. this method has seri- have techniques that can be applied to such species ous problems. deple- tainable. which them to determine the rate of growth. In theory the is both its strength and weakness. Multiply by the growth rate goose Branta canadensis populations showed that those to give the number that can be exploited. Estimating birth and death rates with suf- and this method has been widely used (e. It is necessary to consid- lems with this approach. might be convex (Sutherland & Gill 2000). ical. As this simple example shows. 0. Fa. showed for a range of neotropical primate species that The next stage is to estimate the expected density from population size was related to hunting intensity (none. for example in soil differ in other ways. size. er whether results are confounded by interactions. type (Peres 1993). Mogaka & Fan- shaw 1995. From to population changes.6 changes (showing that there might be a need for a for very short-lived species (age at last reproduction five short-term reduction to allow the population to recov- years). The estimate is often derived from captive animals is the basis for most sports exploitation of birds. for sity from other sites will often be very inaccurate be. 1996) there are good reasons for thinking that this tinguish sustainable exploitation from eradication.6 of this level. Thus a population decline may rate is then multiplied by 0. the current level and the expected unexploited popula- © WILDLIFE BIOLOGY · 7:3 (2001) 137 . Similarly.g. Another major problem is that the sustainable yield Reducing to a fixed fraction of unexploited only applies to a given population density. 0. the sustainable yield In theory the MSY is at half the unexploited population might be calculated at 110 individuals for a population size. An obvious problem here is the population size er but no need for a long-term reduction). Thus the regulations are tight- knowledge of age at first reproduction. sensible management. Fitzgibbon.7 of the unexploited population size. Then assume that the population will be light. at which the population growth rate is being calculat. which would imply high levels of exploitation and Comparing demography across sites so result in overexploitation. and one approach has been to use the difference between This is perhaps the simplest of all the methods. and from what we know about interference.6 or Linking yield to recruitment and mortality 0. Wilkie & Carpenter 1999). reduction in exploitation) or due to environmental lived species (age at last reproduction 5-10 years) or 0. The values will thus rarely be valid mammals. For example. It has the huge advantage that it concentrates for the population at the density at which it will be on the population size and has a good track record for exploited. aging 17% (Hestbeck 1994). then removing all 110 is clearly not sus. other studies. The essence of the Adjusting to population changes Robinson & Redford (1991) model is to calculate the The basic idea is to adjust the exploitation in relation rate of growth from basic data on reproduction. comparing Canada exploited at 0. Growth rate measured Comparing sites that differ in the intensity of exploita- for field populations at equilibrium will usually produce tion is a good method for assessing sustainable levels an estimate of negligible population growth and so (Hilborn. The growth rate under ideal condition may be high. This depends critically upon the shape of the rela- size of 1.500 individuals. in the absence tion territoriality and social behaviour (Sutherland of any idea of population size it is not possible to dis. Juste. In practice. Thus Peres (2000) underestimate the level of sustainable exploitation. To take into the analysis and the better these are understood the consideration the mortality this maximum reproductive better the analysis will be.ticularly true for rain forest species.6 seems to be The recruitment rate and mortality rate can be calculated the widely accepted figure. Environmental changes will also confound the maximum possible reproductive rate. ficient precision is likely to be extremely difficult. It only needs data on calculated growth rate can be removed without chang. This is then in decline had shooting mortalities averaging 27% compared with actual numbers removed to see if the while stable populations had lower exploitation aver- exploitation is sustainable. example because areas with high hunting levels tend to cause of the variation across sites.

be far more occasions in which it is practical to adopt anced against the likely short-term loss from not har. estimating harvesting strategies more. There are. vesting in the manner thought to be most efficient (Sains. Engen. The best tainable exploitation is highly sophisticated. Swenson. A core problem of applying such models is the difficulty ties for overexploitation. dence.5 or 0. method of estimating this for stable populations in cal studies and the major sources of density depen. 1999). Tufto. It is then Ironically. Long-term studies in a range of sites have revealed mercial exploitation results in the social and financial the density dependent processes and especially the role pressures that leads both to detailed research and also of density-dependent nest predation. Studies on factors affecting the unexploited popula- standing are. used to predict the yield and population size resulting There is a case for arguing that for very many species from different percentages shot. Similarly the rent densities is probably the best ways of studying ex. Engen. especially monitoring populations and ploiters. One method is to continually reassess parameters but this may often be impractical. A major problem is that highly underused tool. tion size are thus particularly useful. cal- culating it for captive animals or animal under ideal con- Full population model ditions) or underestimate yield (the very common In some rare cases there will be considerable ecologi. Adaptive management is more likely to be ing is likely to fundamentally affect all of the methods practical where there are numerous replicated such as described in this paper (Walters & Parma 1996). Global climatic change resulting from global warm- bury 1991).e. All exploitation should involve some com. In apply- the previous approaches. As- show how yield and population size would be affected sessing effort or demographic components such as den- by altering the predator control or agricultural practice. the likely population in the absence of exploitation. yet this is very difficult to quantify. ways that will markedly overestimate yield (e. 1995) than for large dis. pling. data may no longer provide a guide to levels of sus- persed populations such as most marine fish. population growth rate is a key measure for calculating ploitation when there is almost no ecological data. this powerful technique. Potts & Aebischer and data. dence are understood and can be quantified. Exploitation is largely about density depen- ing to understand natural densities and relating it to cur.8). such the level that can be exploited. For example. simple means. tainable exploitation. if the habitat deteriorates (or improves). should it be 0. A greater problem is that the in determining the basic components of demography. then for the brown bear Ursus arctos in Norway is greatly this will change the expected unexploited level and hindered by the uncertainty in the main demographic thus also the exploited population size. Adaptive management (Walters 1986) is clearly a problems with this approach. This was also used to the simplest methods are often the most practical. which the population is not growing). There are obvious political prob- exploiters often are particularly sensitive to different rules lems in altering the regulations purely to learn more about in different areas. Bakke & Sande- considerable ecological knowledge to predict these gren 1998. many of the example of this is the extensive studies of the grey most successful schemes are based on limited science partridge Perdix perdix (Potts 1986. unexploited population will often be unknown and These are severely hindered by stochasticity and sam- extrapolating from other sites may be difficult. Swenson & Sandegren changes. is probably often the best method.g. however. ponents of adaptive management. sity dependence or population growth rates are often too difficult to provide estimates that are sufficient to pro- Adaptive management vide a basis for sufficiently accurate exploitation. These data can be to exploitation. One criticism is that the proportional reduc. It would require parameters (Sæther. Sæther. but this is a minor problem compared to other opportuni. although the science underpinning sus- possible to produce a full population model. The past lakes or forests (Hilborn et al. tions and values to discover where gaps in the under. Discussion tion is almost unknown (i. Despite these problems it is likely that attempt. The explanation is probably that intensive com- 1995).tion size. Further. use the best parameter estimates ing this method it is very useful to be able to estimate and then rerun the analysis using different assump. 138 © WILDLIFE BIOLOGY · 7:3 (2001) . yet this is measured in as for forest species. The essence of adaptive management (Walters adjusting regulations according to long-term population 1986) is to carry out the analysis using one or more of changes. as these are often perceived as unjust. the system yet the benefits are so clear that there must The long-term gains in understanding have to be bal. The There is usually a distinction between scientists and ex.

Herdon Virginia.J.R.Ecology 78: lapse of North Sea cod stocks. 576 pp. p.. Oxford.. White. . J. CSIRO Publishing Melbourne.G. 1997: Individual behaviour. J. . .G. S. lations to exploitation. 1992: populations. widely distributed fish. L. pp. C. H. 58: 748-756. Blackwell Science..L.IUGN Gland. & Holt. .J.). & De Barro.. Oldenburg.T. J. D. 53-64. 1995: Demography of a collared peccary population in Beddington. B-E. N. . McGarvey. & Carpenter.L.Canadian Fa. & Bennett. Jennings. A. J.R.Alces 29: 149-162. 1995: Impact Journal of Fisheries and Aquatic Sciences 57: 2357-2362. Hilborn.K. R. 570 pp.The Economist..H. R. . Floyd. 1-9. B-E. C. 1998: Effect of restric- me to such a thought provoking and enjoyable conference and tive harvest regulations on survival and recovery rates of for commenting on the manuscript. R. J.Science 281: 690-692. & Reynolds.M. 1992: Quantitative Stock Assess- Sustainability in Tropical Forests.H. D. C. . Sauer. 1998: Near extinction of a large. © WILDLIFE BIOLOGY · 7:3 (2001) 139 . D. & Sinclair. Harris.Trends in Ecology and Evolution 13: 243-246.E. & Gunn. 1995: Conservation Biology in Lande.. G. Dennis. management: examples from tropical nearshore fin fisheries.. .W.M. & Davies. pp.Blackwell Scientific. A. Oxford. 1997: Potential col. 1998: The case for data-less marine resource Casey. C. R. 432 pp.R. 2001: Marine Fish- Frontiers of population ecology. R. ter in the Atlantic flyway.R. 1994: Wildlife Ecology and American Naturalist 145: 728-745. 1968: The tragedy of the commons. ing for sustainability of fluctuating resources. R.).R. G. 1997: Threshold harvest- Cook. Chapman Press. D. R. Mogaka. .D. New York.J.J. populations in a naturally fluctuating environment. Oxford. B. tion. 2000: Hunting for the snark.C.). E. 1993: Uncertainty.J.I thank Bernt-Erik Sæther for inviting Francis. . Beverton. Sinclair.In: . (Eds. E. dependence and rainfall on abundance of San Joaquin kit Sheppard. J.M.Journal of Wildlife Management 56: 645-650. tory. . Wood.A. 459 pp. A. & Otten. & Fanshaw.W.M.M. S. J.. . 65-77. Fuller.Science 260: 17-18. Ecology and Systematics 26: 45-67.. its effects on mammal reproduction. Nicols. 1977: Fluctuating natural South Texas. & Engen.B.M.L. R. C. P.J. CSIRO Publishing.In: Robinson. 1991: Estimation McCollough. R. . vest in the presence of indirect fishing mortality. bourne..In: Floyd. I also thank Carlos Peres American black ducks. R. populations and conservation . B.Wildlife Monographs 121: 1-39. Hellgren. 2000: Joint effects of density ulations evidence of density dependence? .D. . 1991: Response of North American duck popu- Conservation Biology 9: 1107-1115. . 1996: Is long term persistence of harvest pop- Dennis. A. J. 2000: A fisheries temporal patterns of mortality among female white-tailed management strategy robust to ignorance: rotational har- deer. D.Journal of Wildlife Management 59: 153-163. . R. monitoring both exploited and unex. R. J. . . Hardin. & May. P. of market hunting on mammal species in Equatorial Guinea.Blackwell Sciences.R..J. Johannes. J. 1992: Spatial and Myers.. J. 1996: Spatially structured populations and of growth and extinction parameters for endangered spe.Journal of Wildlife Management Bennett.B. ing of fluctuating populations with a risk of extinction. Hunting for Hilborn. Acknowledgements ..Science Anonymous 2000: Fishermen’s Friends.In: Perrins.. & Scott. S.D. (Eds.Journal of Wildlife Management and John Reynolds for numerous stimulating discussions. & Serie. S. & Satroviejo.In: Krebs. 1995: Optimal harvest- Theory and Practice.D. (Eds..Ecological Monographs 61: 115-143. J. Kenya. A. & Sutherland. . - Caughley. harvest theory.J. G. & Myers. . 1993: The use of Schaefer’s Ludwig. G. dynamics and uncertainty.H. ulation ecology. W. 1957: On the Dynamics of Hilborn. J. M. P. experiments for understanding population regulation. and Fox’s surplus-yield models to estimate optimal moose resource exploitation and conservation: lessons from his- harvest and hunting effort. J. & Stewart. Caughley.B. E. 334 pp. .R. . E. 154. 1341-1350. (Eds.S. & Sæther. J. 1994: Survival of Canada geese banded in win- 197: 463-465.London. eries Ecology. & Kehler. Perez. Synatzske. .M. 62: 1544-1557.). R.Conservation Biol- ogy 9: 1116-1126. & Guthery. Kaiser. R. C. Her Majesty’s exploitation of renewable resources. & De Barro. & Walters. Behavioural Ecology: an Evolutio- References nary Approach.Annual Review of Stationary Office. 1996: Density-pertubation IUCN / UNEP / WWF 1980: World Conservation Strategy. pp. 1995: Sub- ploited populations seems the most straightforward sistence hunting in Arabuko-Sokoke forest. R.B. & Ludwig. . and means of responding to such changes. G. and Hall.D. Lande.Science Hestbeck. Columbia University ment: choice.M. Management. Lebreton.R. & Walters. Fitzgibbon. 25 162: 1243-1248 November.Blackwell Science. Goss-Custard. cies.Nature 385: 521-522. . del Val. Munholland. Dusek. Engen. A.. Cappuccino. J. pp. N.. Juste. S.E. Sæther.C.. G.. H. J.Journal of Wildlife Management 60: 1-9. R. Mel. .For many groups. & Stefánsson.B. Oxford.D. & Robinson. & Jolicoeur. Walters. Compensatory mortality in a Colorado mule deer popula.G. M. P. S. 200 pp.Safari Club International. 373-395. . Courtois.W. R. .A. in press: Biological issues in hunting vertebrate Bartmann. S.London..L. Sheppard. . & Harrison.J. F. 1995: Sustainable Exploited Fish Populations.Journal of Wildlife Management 64: 388-400. Frontiers of pop- fox.

S. pp. .. R.). E. Hunting for Sustain. Columbia University Press.F. & Hirons. . . 1996: Fixed exploitation strate- gren. Sutherland. C. 1954: Some aspects of the dynamics of pop- Peres. J. G. .Oxford University Press. . J-J. Sæther.J. pp.MacMillan. London.H.). & Sande. Oxford University 193: 301-320.Collins.G. Society. London. (Eds.Oxford University Press. (Eds. Bulletin 1: 27-56.C.H. 1993: Structure and spatial organisation of an ulations important to the management of commercial marine Amazonian terra firme forest primate community. . J. . F.. . Press.G. .L.Ibis 137: S16-S28.A. Swenson. 1999: Harvesting strategies for conserving minimum eries Research Canada 11: 559-623. 1998: Assessing the viability of Scandinavian brown gies for coping with effects of climatic change.. B-E. & Sandegren. pp.). ogy and Fisheries 6: 125-137. K. Journal of Fisheries and Aquatic Sciences 53: 148-158. Sæther. R. 1954: Stock and Recruitment. Behaviour Potts. 604 pp.B.R. 2000: Predicting the conse- Potts. F. Sutherland. Ricker.R. & Aebischer. 2000: Evaluating the impact and sustainability of Shenk.Oikos 83: 403-416. . & Lande. London Series B. 498-525. . G. the Exploration of the Sea.P. Wilkie.In: Robinson. .B. (Eds. Conservation. - Quinn III. Cambridge University Press. meter estimates. G. Bakke. 1996: Lessons from stock assess- dence and optimal harvesting of fluctuating populations.E. W. 1972: An investigation of the accuracy of virtual Ecology.A. J. & Deriso. L. 51-64. & Sutherland. (Eds. Swenson. population analysis. 1995: Population dynamics of and Conservation.S. .W. gation. Engen.E. & Parma. Oxford. K. J. Resources. modelling and management. J. 1986: Adaptive Management of Renewable University of Chicago Press. J. Walters.. K. ly uncertain dynamics. Dynamics. D. pp. 31-56. Engen. 266: 961-967. Bird Population Studies: Their rele. J.. N. 1996: Density-depen. 1999: Bushmeat hunting in the Sainsbury. Walters.ICNAF Research Bulletin 9: 65-74. C. W. 278 pp.J. Predation and quences of human disturbance from behavioural decisions.G. New York.M. M. White.G. viable populations based on World Conservation Union cri- Robinson. Neotropical Wildlife Use and Conservation. 1991: Sustainable harvest for teria: brown bears in Norway. . T. .E. W. Sæther. & Maguire.Reviews in Fish Biol- Oikos 76: 40-46. S. Marine Science Symposium vance to conservation and management. 1986: The Partridge: Pesticides. 274 pp.Journal fisheries.A.Biodiversity and Conservation 8: 927-955.Sainsbury International Council for 140 © WILDLIFE BIOLOGY · 7:3 (2001) . Engen. & Redford. G.M. K. 1991: Application of an experimental approach Congo Basin: an assessment of impacts and options for miti- to management of a tropical multispecies fishery with high. . . ment from the northern cod collapse.Proceedings of the Royal neotropical forest mammals.. & Gill.J. 415-429..J.M. B-E. A.Canadian bear Ursus arctos populations: the effects of uncertain para. the grey partridge Perdix perdix 1793-1993: monitoring. 1999: Quantitative Fish Blackwell Scientific.Ecological Mono- ability in Tropical Forests. C. Peres. graphs 68: 445-463. Schaefer.J. 2000: The Conservation Handbook. Tufto.Inter-American Tropical Tuna Commission of Tropical Ecology 9: 259-276. C. Cambridge. 542 pp.J. C. B-E. & Carpenter. & Redford. .. .).In: variance effects on detecting density dependence from Robinson.J. Sutherland. Chicago.In: Gosling. Oxford.J. temporal trends in natural populations. 1996: From Individual Behaviour to Population Pope. J. & Bennett. .Journal of Fish. Oxford. Ø. Oxford. 213 pp. W. J. & Burnham. Walters. 1998: Sampling- subsistence hunting at multiple Amazonian forest sites. S. T.