Assessing the sustainability of African lion trophy hunting, with recommendations for policy

Transcription

1 Ecological Applications, 26(7), 2016, pp by the Ecological Society of America Assessing the sustainability of African lion trophy hunting, with recommendations for policy Scott Creel, 1,2,5 Jassiel M soka, 4 Egil Dröge, 1,2 Eli Rosenblatt, 1,2 Matthew S. Becker, 1,2 Wigganson Matandiko, 1,2 and Twakundine Simpamba 3 1Zambian Carnivore Programme, P.O. Box 80, Mfuwe, Eastern Province, Zambia 2Department of Ecology, Montana State University, Bozeman, Montana USA 3Department of National Parks and Wildlife, South Luangwa National Park, P.O. Box 18, Mfuwe, Zambia 4Department of National Parks and Wildlife, Ecology Section, Private Bag 1, Kafue Road, Chilanga, Zambia Abstract. While trophy hunting provides revenue for conservation, it must be carefully managed to avoid negative population impacts, particularly for long- lived species with low natural mortality rates. Trophy hunting has had negative effects on lion populations throughout Africa, and the species serves as an important case study to consider the balance of costs and benefits, and to consider the effectiveness of alternative strategies to conserve exploited species. Age- restricted harvesting is widely recommended to mitigate negative effects of lion hunting, but this recommendation was based on a population model parameterized with data from a well- protected and growing lion population. Here, we used demographic data from lions subject to more typical conditions, including source sink dynamics between a protected National Park and adjacent hunting areas in Zambia s Luangwa Valley, to develop a stochastic population projection model and evaluate alternative harvest scenarios. Hunting resulted in population declines over a 25- yr period for all continuous harvest strategies, with large declines for quotas >1 lion/concession (~0.5 lion/1,000 km 2 ) and hunting of males younger than seven years. A strategy that combined periods of recovery, an age limit of 7 yr, and a maximum quota of ~0.5 lions shot/1,000 km 2 yielded a risk of extirpation <10%. Our analysis incorporated the effects of human encroachment, poaching, and prey depletion on survival, but assumed that these problems will not increase, which is unlikely. These results suggest conservative management of lion trophy hunting with a combination of regulations. To implement sustainable trophy hunting while maintaining revenue for conservation of hunting areas, our results suggest that hunting fees must increase as a consequence of diminished supply. These findings are broadly applicable to hunted lion populations throughout Africa and to inform global efforts to conserve exploited carnivore populations. Key words: carnivore; extinction risk; Panthera leo; population dynamics; trophy hunting. Introduction Large carnivores are declining globally (Estes et al. 2011), with important consequences for ecosystem structure and function (Pace et al. 1999, Scheffer et al. 2001, Terborgh and Estes 2013). These declines are wellexemplified by the lion (Panthera leo), Africa s largest carnivore, for which suitable habitat has declined by approximately 75% in recent decades (Riggio et al. 2013). Lion numbers have decreased in parallel with their habitat, and only 10 stronghold populations (stable, protected populations of more than 500 individuals) currently persist (Bauer and Van Der Merwe 2004, Packer et al. 2013, Riggio et al. 2013, 2016, Bauer et al. 2015). As with many other apex carnivores, most large, free- ranging lion populations are limited by a combination of habitat loss, prey depletion, direct conflict and retaliatory killing, wire- snare poaching, and trophy Manuscript received 4 March 2016; revised 2 May 2016; accepted 12 May Corresponding Editor: T. O brien. 5 screel@montana.edu hunting (Yamazaki 1996, Ogada et al. 2003, Bauer and Van Der Merwe 2004, Loveridge et al. 2007, Packer et al. 2009, 2011, Becker et al. 2013a, b, Rosenblatt et al. 2014). For lions and all other exploited species, trophy hunting is unique because it is intentional and legal, and thus more easily controlled than other limiting factors (Creel et al. 2015). In 2015, the U.S. Fish and Wildlife Service listed lions in Western and Central Africa as endangered and those in Eastern and Southern Africa as threatened under the Endangered Species Act. This listing decision found evidence for practices that experts have identified as undermining the sustainability of trophy hunting and included an explicit rule that body parts from lions killed in trophy hunts can only be imported from nations with a scientifically sound management program that benefits the subspecies in the wild (U.S. Fish and Wildlife Service 2015). Due to concern over population decline, poor recruitment, and a decrease in the age of lions being shot in Game Management Areas adjacent to South Luangwa National Park, the Zambia Wildlife Authority implemented a moratorium on lion 2347

2 2348 SCOTT CREEL ET AL. Ecological Applications Vol. 26, No. 7 hunting for (Rosenblatt et al. 2014). It is unusual that a limiting factor can be switched off so directly (Rutledge et al. 2010, Creel et al. 2015), but hunting is also unusual because it provides revenue that can contribute to conservation, and from this economic perspective, lions are one the most important of Africa s hunted species (Creel and Creel 1997, Packer et al. 2011, Lindsey et al. 2013, U.S. Fish and Wildlife Service 2015). Well- regulated hunting has the potential to contribute to lion conservation, but data show that overharvesting has contributed to their current decline (Packer et al. 2011, Lindsey et al. 2013, Rosenblatt et al. 2014), and the ecological, cultural, and economic importance of lions mandates careful attention to these issues (Creel et al. 2015). Empirical studies of trophy hunting effects on lions Lions, like most large carnivores, have low adult mortality rates in the absence of human effects, making it unlikely that hunting simply substitutes for other causes of death (Rosenblatt et al. 2014, Creel et al. 2015), and data confirm that some lion populations are affected by overhunting. Loveridge et al. (2007) examined lions in Hwange National Park (Zimbabwe) in an area of 7,129 km 2, of which 83% was within the National Park where hunting is not allowed, and 17% was within adjacent communal lands, safari areas, and hunting concessions. Of 62 individuals marked on this site, trophy hunting accounted for 63% of 38 individuals who died or disappeared. The proportion of the population composed of adults males declined from 26.3% to 13.7% during a period of 4 yr over which the offtake of males doubled. Of marked, territorial males in this population, 72% were killed by trophy hunters, at a mean distance of 1.5 km from the park boundary. Thus in the Hwange region, sport hunting is largely based on source sink harvesting of lions from the national park and not on sustainable offtake of lions residing within the hunting areas themselves (a phenomenon that Loveridge termed the vacuum effect ). Rosenblatt et al. (2014) found similar results for lions on a 2,775- km 2 study site centered on the Luangwa River (Zambia), including portions of South Luangwa National Park (SLNP, which does not allow hunting) and the adjacent Lupande and Lumimba Game Management Areas (GMAs, which allow hunting). Lupande and Lumimba are considered prime hunting areas by the Department of National Parks and Wildlife. This classification is currently applied to a small number of GMAs, all adjacent to the large, protected lion populations of SLNP and Kafue National Park. Using mark recapture models fit to data from intensive monitoring of 210 individual lions in 18 prides and 14 male coalitions over 5 yr, they found that trophy hunting was the most common cause of death, with 46 males harvested from a population showing indications of overharvest that included population decline, low recruitment, low sub- adult and adult male survival, depletion of adult males (prime- aged and old adult males never exceeded 6% of the population), and a senescing adult female population. The median age of harvested males was yr old, substantially below the minimum age of six suggested by Whitman et al. (2004), and annual harvest rates ranged from 1.86 to 2.56 lions/1,000 km 2, much higher than the baseline recommendation of 0.5 lions/1,000 km 2 (Packer et al. 2011) and well above the higher value of 1 lion/1,000 km 2 recently suggested for Tanzania s Selous Game Reserve (Packer et al. 2011). As in Hwange, harvest was heavily concentrated along the national park boundary (Rosenblatt et al. 2014; see Results and discussion; Fig. 1). In Tanzania s Selous Game Reserve, which holds the largest lion population on the continent (Creel and Creel 1997, Bauer and Van Der Merwe 2004, Creel et al. 2013), lions attained densities between eight and 13 per 100 km 2, and on a focal study area of 2,600 km 2 in the Northern Sector of Selous, lions were killed at a rate of % of adult males annually between 1989 and This mortality was thought to be sustainable, while noting that lion quotas were substantially higher, equal to 10 16% of the adult male population. Offtake of the full quota was thought to be unsustainable, and the percentage of quotas filled (both in Selous and nationwide) began declining in 1988 as quotas increased (Creel and Creel 1997). Packer et al. (2011) analyzed harvest records for all of Tanzania s hunting blocks, which cover more than 300,000 km 2 (approximately 100,000 km 2 in game reserves and 200,000 km 2 in less- protected game controlled areas) and found that lion harvests declined by 50% between 1996 and 2008, with the greatest declines in the areas of heaviest harvest. Thus in Tanzania, estimated catch per unit effort has been declining for several decades (Creel and Creel 1997, Packer et al. 2011). In Zimbabwe s Gonarezhou National Park and Tuli Safari area, lion densities (as estimated from call- up surveys) were between 0% and 16% of those predicted on the basis of prey biomass, a pattern that was not observed for other large carnivores (Groom et al. 2014). Data on offtake of lions by trophy hunting and other potential limiting factors suggested that population collapse was due to direct human effects, including unsustainably high trophy hunting quotas in conjunction high levels of legal control killing and illegal killing (Groom et al. 2014). In the Bénoué Complex of Cameroon, track surveys conducted in three national parks and three hunting zones suggested that lion densities in hunting zones were only 31% of those in national parks (Croes et al. 2011). Lions attained only 27% of their estimated carrying capacity (based on prey density) in hunted areas, compared to 53% in national parks. In contrast, densities of leopards (Panthera pardus) and spotted hyenas (Crocuta crocuta) did not detectably differ between hunted areas and national parks. Relating these patterns to data on the number of lions shot and permits issued, Croes et al. (2011) concluded that in addition to other problems,

3 October 2016 ASSESSING LION TROPHY HUNTING 2349 Fig. 1. Ranging patterns of lions in the Luangwa Valley from 2008 to 2012 and the reported locations of 21 lions shot by trophy hunters in the same years. One recorded kill site (not shown) in the western portion South Luangwa National Park was probably recorded incorrectly. No location was reported for 22 of 44 lions shot in this period. Lion kill sites are from the Department of National Parks and Wildlife records, and lion ranging patterns are shown by radio locations of resident females (Rosenblatt et al. 2014) with different colors denoting different prides. lions living in the hunting zones are strongly impacted through excessive trophy offtakes. In summary, empirical studies have detected low lion density and population declines associated with excessive trophy hunting, using a broad range of methods at sites widely distributed across the lion s range. Modelling trophy hunting effects on lions A population model applied to data from lions in Serengeti National Park suggested that hunting can be sustained simply by hunting males above a minimum age threshold (Whitman et al. 2004, 2007). Following this suggestion, age- limited harvesting has been implemented in several nations (e.g., Tanzania, Zambia, and Mozambique) in an effort to make lion hunting more sustainable. Also based on this model, a minimum permitted age of six years has been widely adopted. When considering the generality of this strategy, one must recognize that the model was parameterized with data from a wellprotected population experiencing strong, sustained growth as its prey base increased (Packer et al. 1988, Hanby et al. 1995); this limitation was carefully acknowledged by the original authors (Whitman et al. 2004). In contrast to patterns in Serengeti for the period modelled, the ungulate prey base supporting lions and other large carnivores is declining in most ecosystems (Bolger et al. 2008, Caro and Kerley 2008, Western et al. 2009), and many lion populations are thought to be declining as a result of prey depletion and related problems (Bauer et al. 2015, Riggio et al. 2016). Moreover, the structure of the underlying population model (Quadling and Starfield 2002, Whitman et al. 2004) was highly complex and process- based, requiring detailed data such as the likelihood of death in territorial disputes between males of specific ages and resident/nomad status, the likelihood of females locating open territories as a function of distance, and differences in the likelihood that dispersing males will join single males, groups of nomads, or pride- holding males. The lack of such data has limited application since the model s development (by Starfield more than 30 yr ago) to a single population in Serengeti National Park (Whitman et al. 2004), with some important parameters roughly estimated even in that application (for examples, see the supplementary data of Whitman et al. 2004). To address uncertainty about the sustainability of current hunting policies in areas more representative of harvested lion populations, we developed a simpler model using stochastic Leslie matrix projection. We then applied the model to data typical of current conditions for many lion populations, with exposure to effects of human encroachment, snaring and prey depletion, and source sink dynamics in which lions moving from a fully protected area are regularly exposed to mortality from hunting (Loveridge et al. 2007, Becker et al. 2013a, b, Watson et al. 2013, 2015, Rosenblatt et al. 2014). We used population and harvest projections over 25 yr to examine the sustainability of a range of potential policies including age- restricted harvest, quotas, and recovery periods, separately and in combination. Our results strongly suggest

4 2350 SCOTT CREEL ET AL. Ecological Applications Vol. 26, No. 7 that a combination of all of these strategies will be needed to avoid population declines due to overharvest, with appreciable risk of local extirpation. A similar model (parameterized with data from Serengeti) was recently used to examine regulation of hunting with data only on the number of days required to kill a lion (Edwards et al. 2014). As Edwards et al. (2014) note, the value of this approach is that it can be applied even if data on population size, structure, and growth rate are limited. On the other hand, indirect approaches depend on a strong and constant relationship between population size and the indirect measure of choice. In this case, the model assumed that the number of days required to shoot a lion (μ) relates inversely to population size in a constant manner, μ = 1/cN, where N is population size and c is a constant describing hunting effort and effectiveness. Thus, the model s results depend on an assumption that hunting effort and methods do not change as the targeted population decreases, which we do not consider realistic for lions. Even if constant effort and methods could be mandated and enforced, this approach relies on the additional assumption that μ has a strong relationship with population density, which to our knowledge is not supported by data from lions (see Results and discussion; Fig. 2), perhaps because a day of hunting can mean many different things. For example, the use of baits (ungulate carcasses) to attract lions is still common practice. If multiple baits are dragged and hung in areas of known use (for example, watering points with good stalking cover), the probability of attracting a lion is likely to be decoupled from changes in population size until they become extreme. Similar concerns arise for the regulation of hunting with most large carnivores, for which population monitoring is generally difficult, limiting direct data on population size and trends. Methods We modelled lion dynamics without harvest and with a range of harvest scenarios. The online supplement describes the model s structure, parameterization, and validation in detail (Appendix S1). Our basic approach was to begin with a population of defined size (180 individuals), with a sex ratio and age- distribution based on South Luangwa National Park (SLNP) and the adjacent Lupande and Lumimba GMAs. These starting conditions are typical of many of Africa s most important lion hunting areas, with a single population that occupies a National Park in which lions cannot be hunted and adjacent areas in which hunting is allowed (see Results and discussion; Fig. 1). Recent population trends for lions in SLNP have included a decrease in population size and age- sex trends indicative of excessive trophy hunting (Rosenblatt et al. 2014). Many lion populations now show trends and constraints comparable to those of Luangwa Valley lions. With these starting conditions, we projected population dynamics 25 yr into the future, without hunting and with a range of hunting scenarios described in the following section. For all cases, we used stochastic Leslie matrix projection, with linear density dependence and separate projection matrices for males and females (because they differ in patterns of age- specific survival and reproduction). Fig. 2. Indirect measures of population status showed high variability and no tendency to decline over a period of 10 yr for which direct data revealed population decline, depletion of males in prime age- classes, poor recruitment, and a decrease in the median age of harvested males to <5 yr (Rosenblatt et al. 2014). Data are from Department of National Parks and Wildlife records; recorded measurements that are not plausible were not deleted because our intention was to evaluate the signal- to- noise ratio of such data. Error bands show 95% confidence intervals. Effects of trophy hunting We modelled the effects of three methods that are currently proposed or used to regulate lion hunting, separately and in combination: (1) block quotas, (2) age- restriction, and (3) recovery periods with no hunting. SLNP is bordered by four hunting blocks (Upper and Lower Lupande GMA, Lumimba GMA, and Nyampala GMA), ranging in size from 1,296 km 2 to 4,392 km 2. Recent and proposed regulations allow either one or two males to be harvested in each of these blocks, so we modelled block quotas of one or two. SLNP is the primary source for lions shot in Upper and Lower Lupande and Lumimba (see Results and discussion; Fig. 1), but the ranging patterns of lions shot in Nyampala have not been described. Lions shot in Nyampala are likely to originate from SLNP, but if they do not, then results midway between those shown for block quotas of one and two would best represent block quotas of two for the other three blocks (see Results and

5 October 2016 ASSESSING LION TROPHY HUNTING 2351 Direct hunting mortality. We incorporated direct mortality of males due to trophy hunting by checking within each group at each time step for males above the minimum allowable age and removing them until the quota was filled. For scenarios other than continuous harvest, no males were removed by hunting in years designated for recovery. Fig. 3. With harvesting in every year, (top) population size is expected to decline appreciably, so that (bottom) the probability of local extinction within 25 years is also appreciable. For each minimum permitted age, the three curves show models with no effect on cub recruitment, moderate effects, and strong effects. discussion; Figs. 3 and 4). Current or proposed policies in Zambia, Tanzania, Mozambique, and elsewhere incorporate a minimum age for males that can be shot, though there is currently little direct guidance about the appropriate age threshold for a population facing the conditions typical for most lions. Simulations based on data from the well- protected and growing lion population in Serengeti National Park suggested a threshold of six years, but applicability of this threshold to other populations has not been tested. Thus, we modelled scenarios that varied the minimum allowable age from four to eight years. Finally, the Zambian government recently closed trophy hunting for three years to allow a period of recovery in both population size and age structure. Thus, we modelled scenarios with continuous harvest, with two- year recovery periods between hunted periods of four years (four on/two off) and with alternating three- year periods of hunting and population recovery (three on/three off). We present results for combinations of block quota (one or two), recovery periods (continuous harvest, four on/two off, three on/three off), and minimum age (4 8 yr). Super- additive effects of hunting mortality on recruitment of cubs. The social disruption created by increased mortality due to hunting often reduces reproduction in social carnivores (Whitman et al. 2007, Rutledge et al. 2010, Creel et al. 2015). For lions, an increase in male turnover increases the frequency of infanticide by males that inherit prides, reducing recruitment (Whitman et al. 2004, 2007). To incorporate super- additive effects on recruitment in a simple manner that is amenable to comparison with empirical data, we assumed that effects on recruitment will be stronger when the minimum age of males that can be shot is lower, because increased social disruption should increase the frequency of turnover. We then defined the reduction in fecundity due to hunting within the age classes exposed to this effect. As described in detail in the online supplemental materials, we examined three cases: one with no effect on fecundity due to male losses, one with strong effects on fecundity (reduction of 32 55% with stronger effects when a wider age range is hunted), and one intermediate case. These cases span the plausible range for the strength of superadditive effects in lions (Appendix S1: Fig. S4), and we show results for all three cases for each harvest regulation scenario that we modelled. Model outputs For each scenario, we ran 500 stochastic simulations over 25 yr, from which we report mean population size, the proportion of populations that were extirpated, and the total number of lions taken by trophy hunters. These data allow comparison of risks for different scenarios and allow comparison of risks to financial benefit from license sales in a manner that could readily be used to adjust license fees. Results and Discussion Lions are often hunted in areas adjacent to a fully protected source population (Loveridge et al. 2007, Rosenblatt et al. 2014), so a population model parameterized with data from such conditions is of general interest. From 2008 to 2012 in the Luangwa Valley, 22 kill sites recorded by the Department of National Parks and Wildlife show that lions were killed a median of 997 meters (standard error, 463 m) from SLNP, well within the normal ranges of lions that occupy the primary photo- tourism area of SLNP along the Luangwa River (Fig. 1; lion locations from Rosenblatt et al. 2014). The observed distribution of distances

6 2352 SCOTT CREEL ET AL. Ecological Applications Vol. 26, No. 7 Mean population size Fig. 4. Mean population size for population projections over 25 years with variation in the minimum male age permitted for hunting. For each minimum age, population means are shown for three scenarios: no effect of male loss on reproduction, moderate effects, and strong effects. Panels show (left) maximum block quota of one male shot per year, (right) maximum block quota of two males shot per year, (top) hunting cycle of four years on and two years off, and (bottom) hunting cycle of three years on and three years off. Year between kill sites and the National Park boundary did not overlap with a distribution of random locations within the GMA (P < 0.001, bootstrap randomization test). In short, the Luangwa Valley holds a single lion population with individuals regularly moving between SLNP (where they are fully protected and support photo tourism) and the adjacent GMAs (where they are exposed to hunting). We did not model the sustainability of hunting regulations based on indirect measures such as the number of days required to kill a lion. Data from the Luangwa Valley ( ) show that these indirect measures provide a weak signal- to- noise ratio (Fig. 2). Neither the number of days hunted nor reported measures of trophy quality showed any tendency to decline over a period for which capture recapture estimates of population size revealed a significant decline, which (together with changes in age- structure) was deemed serious enough to provoke a complete hunting closure in 2013 (Rosenblatt et al. 2014). More specifically, a generalized linear model (family, Poisson; link, log) relating the number of days required to kill a lion to time had a slope indistinguishable from zero (b = ± SE, z = 0.28, P = 0.78). The baseline model (a scenario with no hunting and other limiting effects at levels) yielded a deterministic growth rate (λ) of 1.022, stable stochastic dynamics, and no risk of extirpation over 25 years (Appendix S1: Fig. S3). Because the initial conditions were for a population well below carrying capacity with λ > 1, populations in the baseline model increased for roughly 6 7 years until reaching equilibrium. All scenarios with hunting produced some degree of population decline and local extirpation probabilities greater than zero, though some combinations of strategies yielded relatively stable dynamics and low probability of extirpation over 25 yr. Age- restricted harvesting Our results (Fig. 3) strongly support the prior inference that restricting harvest to males above a certain age increases the sustainability of lion hunting (Whitman et al. 2004, 2007, Whitman and Packer 2007). However, our results do not support the suggestion that restriction of hunting to males of age six or older will reliably yield sustainable offtake in the absence of other restrictions, with the conditions typified by Luangwa Valley lions. As Fig. 3 shows, with age restriction and no other limits on offtake, the probability of extirpation was essentially certain if males as young as four years were allowed to be hunted (as has recently been the case in SLNP, where of 12 known individuals shot by hunters between 2008 and 2012 the median age was 4.86 yr; Rosenblatt et al. 2014). The probability of extirpation dropped monotonically as the minimum age was raised, but fell consistently below 50% only with a minimum age of seven and remained at 5 11% with a minimum

7 October 2016 ASSESSING LION TROPHY HUNTING 2353 Proportion extirpated Fig. 5. The proportion of simulated populations that went to local extinction over 25 years with variation in the minimum male age permitted for hunting. For each minimum age, extinction probabilities are shown three scenarios: no effect of male loss on reproduction, moderate effects, and strong effects. Panels show (left) maximum block quota of one male shot per year, (right) maximum block quota of two males shot per year, (top) hunting cycle of four years on and two years off, and (bottom) hunting cycle of three years on and three years off. Year age of eight. The results suggest that (1) age restriction is an important element of sustainable hunting for lions, (2) a minimum age of seven or eight is necessary to yield a reasonably low risk of extirpation even in the near future, and (3) age restrictions must be combined with other regulations to assure sustainability. Like the baseline model (Appendix S1: Fig. S3), all of the scenarios with hunting simulated an initial population well below carrying capacity with λ = 1.022, which allowed population growth for the first few years of simulation (Fig. 3), but this transient period of growth would not be expected with less favorable starting conditions for the simulation. Lions can be aged using characteristics such as nose pigmentation, mane development, and tooth condition (Whitman et al. 2007), but even careful application of these methods will yield some errors in age estimates. Our results can be used to examine the consequences of errors in aging by recognizing that true age is being modelled. For example, if errors in age estimation cause half of the lions shot to be five- years old, even though regulations stipulate a minimum age of six, then results intermediate to those for ages five and six would describe the predicted effect on population dynamics. We note that errors in aging are likely to be reduced by application of a minimum age of seven or eight, because age- specific variation is greater at younger ages when traits are changing more rapidly (Whitman and Packer 2007). Block quotas and recovery periods Given the relatively high risk of extirpation from age restricted harvesting, particularly with the commonly used threshold of six years, we examined scenarios with two restrictions applied in conjunction with age limits: maximum block quotas and recovery periods. A block quota of two examines a situation in which total harvest (maximum of eight individuals killed/yr) is slightly lower than reported harvests (X = = 8.8 individuals killed/yr) from 2008 to 2012, and a block quota of one examines the consequences of halving this rate. Recent observations suggest that South Luangwa lion dynamics have responded to a moratorium on hunting that began in 2013, so we also modelled recovery periods of two or three years out of each six- year period. Results for these scenarios (Figs. 4 and 5) show that block quotas and recovery periods both considerably increased the sustainability of hunting and that combining a block quota of one with a recovery period was the only scenario that consistently yielded a risk of extirpation below 20%. This scenario yields maximum annual mortality due to hunting of substantially <0.5 lions/1,000 km 2 (see Results and discussion: Number of males shot), which has been proposed as a reasonable threshold for sustainability (Packer et al. 2011). The value of 0.5 lions/1,000 km 2 is a valuable general guideline, but it was based on estimates of population growth rates taken from coarse estimates of population size, which is notoriously difficult to estimate for

8 2354 SCOTT CREEL ET AL. Ecological Applications Vol. 26, No. 7 Males shot Fig. 6. Differences between hunting scenarios in the number of males legally harvested in each year. As population size and age structure changed through time, harvest rates declined for all scenarios, with the smallest decline for a maximum block quota of one, minimum permitted age of seven or eight, and recovery periods of three years between three- year periods of hunting. Year lions without intensive monitoring (Creel et al. 2013, Rosenblatt et al. 2014, Riggio et al. 2016). As expected, scenarios that included recovery periods and block quotas were less sensitive to effects of minimum hunted age than more liberal scenarios (Figs. 4 and 5). This result provides potentially useful guidance by identifying regulations that are expected to be sustainable in the face of uncertainty in the aging of lions or difficulty in enforcing age restrictions. Strategies that incorporate recovery periods and block quotas can promote sustainability despite problems with accurate aging or enforcement of age restriction. This combination of strategies is also inherently safer than harvesting solely on the basis of age with a points system that reduces quotas if underage lions are shot, because such systems adjust quotas after the effects of killing underage males have already occurred. Finally, because very few males reach old age classes under most harvest scenarios (as in empirical data for SLNP), differences in outcomes for age thresholds of seven and eight years were somewhat unpredictable unless aggregated across a very large number of model iterations. This result is logically realistic; a change in policy that affects a small number of individuals in a stochastic process will have rather stochastic outcomes. Number of males shot Figure 6 shows the number of males legally shot in each year for each of the harvest scenarios described previously. These data reinforce the conclusion that a combination of strategies is needed to promote sustainability, because a clear decline over time in the number of harvestable males is apparent for most scenarios. This decrease in the availability of prime- aged males is minimized by the combination of a block quota of one, a three- on/three- off cycle of hunting and recovery, and a minimum hunted age of seven or eight years. While this scenario yields the lowest risk of extirpation, even this combination is predicted to produce a decline in harvestable males with the baseline conditions of this model, even with no super- additive effects on reproduction. Moreover, the survival rates in our base simulations include effects from poaching and human encroachment at their current levels and then assume that these problems will not worsen, which is not necessarily correct (Becker et al. 2013a, b, Watson et al. 2013, 2015, Rosenblatt et al. 2014). If other negative effects on lions are not controlled, it is unlikely that trophy hunting at any level will be sustainable. Implications for sustainable license fees Figure 7 shows the mean number of males killed by trophy hunters over 25 years for each of the policy scenarios discussed previously. The number of harvested males ranged from a maximum of 37 to a minimum of 7. As Fig. 7 shows, many scenarios yield similar total offtake over a span of 25 years, parallel to results above (Figs.4 & 5) showing that many scenarios yield similar risks of extirpation. This result is important

9 October 2016 ASSESSING LION TROPHY HUNTING 2355 Fig. 7. The total number of adult males shot over 25 years for scenarios varying the minimum age permitted, the duration of cycles between hunting and population recovery, and block quotas of one or two. because it implies that, among alternatives with similar expected outcomes, the one that is most easily and effectively enforced should be favored. Figure 7 can also be used to assess what changes might reasonably be made to lion hunting fees to promote sustainability, while maintaining current revenue for conservation. Forty- four males, or a mean of 8.8 males/yr, were taken from this population by trophy hunters between 2008 and 2012, which yielded population decline, poor recruitment, a population with few old males, and a decrease in the age of harvested males (Rosenblatt et al. 2014). Figure 7 suggests that scenarios yielding sustainable offtake would require offtake to decline by roughly an order of magnitude. Thus, a clear policy recommendation is that the fee for lion hunting should be increased comparably, if it is expected to yield ecologically and economically sustainable trophy hunting. Such changes might usefully be paired with longer leases on hunting concessions (so that hunting operators could plan for required periods of recovery) and staged fees so that the full amount is only charged if a lion is shot. The economics of supply and demand imply that any scarce and declining resource will become more expensive. If treated simply as a commodity (an approach that is questionable on many fronts), lions have declined dramatically over recent decades (Bauer and Van Der Merwe 2004, Packer et al. 2011, Riggio et al. 2013). Declines have been larger in the areas with heavier hunting, and hunting was the strongest predictor of population declines for lions in Tanzanian protected areas (Packer et al. 2011). Our results strongly suggest that lion hunting fees have been too low to promote truly sustainable hunting. Throughout their range, the conservation of free- ranging lion populations has been demonstrably constrained by budgets that do not provide sufficient protection (Packer et al. 2011, Creel et al. 2013). Thus, any policy change that reduced funding for necessary conservation actions, such as patrolling and snare removal, would be expected to have negative effects on lion populations by exacerbating widespread and accelerating problems such as wildlife depletion due to snaring and encroachment by local people (Watson et al. 2013, 2015). At the same time, substantial recent data demonstrate that trophy hunting at unsustainable levels has contributed to lion population declines (Loveridge et al. 2007, Packer et al. 2009, 2011, Rosenblatt et al. 2014). Collectively, these results suggest the following. 1. For trophy hunting to be sustainable under the conditions that most lions now experience, modifications to policy must reduce total mortality. 2. Age-restricted harvesting is effective to increase the sustainability of trophy hunting, but is probably not sufficient to yield sustainability by itself. The widely suggested minimum huntable age of six years (Whitman et al. 2004, 2007) was based on data from a growing lion population in Serengeti, with more favorable conditions for lions than now exist for most populations that support trophy hunting. Our results suggest that a minimum age of seven or eight is substantially more effective for the conditions typified by Luangwa lions and should be combined with other policies to limit total mortality. 3. The acceptable risk of local extirpation will have to be confronted. Our results show that a risk of extirpation that might be considered acceptable (e.g., <10%) is likely to be accomplished only with a combination of

10 2356 SCOTT CREEL ET AL. Ecological Applications Vol. 26, No. 7 methods. This inference is generalizable to other exploited species. 4. Policies that combine age restriction, conservative block quotas, and recovery periods are more likely to promote sustainable harvest, particularly with uncertainty about field estimates of age or difficulties with enforcement of regulations. A general lesson from this result is that policies that incorporate a combination of strategies to limit negative effects of hunting on population dynamics provide a margin of safety. 5. As lion populations (and particularly prime aged males) decrease, the price of licenses must increase to halt declines linked to overharvesting, to accurately reflect a decrease of supply (in the economic sense), and to ensure that hunting contributes to effective conservation of areas in which hunting is allowed. The value of lions that are not directly linked to short-term economics (e.g., ecological and cultural importance, as well as value to non-consumptive tourism) should not be ignored and where lions are hunted in protected area complexes, efforts should focus on restoring and increasing lion populations and habitat for GMAs now classified as secondary or understocked rather than finding means to increase offtake on prime hunting blocks. More broadly, low hunting license fees for carnivores promote aggressive offtake of these rare and ecologically important species (Creel et al. 2015; e.g., a wolf license for out-of-state hunters in Idaho costs only US$32, with individuals allowed to take up to five, while licenses for deer and elk, which outnumber wolves by several orders of magnitude, cost US$302 and US$417, respectively). 6. The data used to parameterize our model includes impacts on survival from poaching and human encroachment at their current levels, thus assuming these activities will not worsen, which is unlikely given burgeoning human pressures across Africa s remaining lion populations. If the current acceleration of other negative effects on lions is not addressed and controlled, then it is unlikely that trophy hunting at any level will be sustainable. These conclusions are likely to generalize to other African large carnivores, particularly the leopard. Acknowledgments We thank the Zambia Department of National Parks and Wildlife for permission to conduct this research and for their support and cooperation. This work was supported by grants from the National Science Foundation (IOS ), WWF- Netherlands, Mfuwe Lodge/Bushcamp Company, Painted Dog Conservation, and the National Geographic Society Big Cats Initiative. Literature Cited Bauer, H., and S. Van Der Merwe Inventory of free-ranging lions Panthera leo in Africa. Oryx 38: Bauer, H., G. Chapron, K. Nowell, P. Henschel, P. Funston, L. T. Hunter, D. W. Macdonald, and C. Packer Lion (Panthera leo) populations are declining rapidly across Africa, except in intensively managed areas. Proceedings of the National Academy of Sciences USA 112: Becker, M. S., F. G. Watson, E. Droge, K. Leigh, R. S. Carlson, and A. A. Carlson. 2013a. Estimating past and future male loss in three Zambian lion populations. Journal of Wildlife Management 77: Becker, M. S., R. McRobb, F. Watson, E. Droge, B. Kanyembo, J. Murdoch, and C. Kakumbi. 2013b. Evaluating wire- snare poaching trends and the impacts of by- catch on elephants and large carnivores. Biological Conservation 158: Bolger, D. T., W. D. Newmark, T. A. Morrison, and D. F. Doak The need for integrative approaches to understand and conserve migratory ungulates. Ecology Letters 11: Caro, T., and G. I. H. Kerley Decline of large mammals in the Katavi- Rukwa ecosystem of western Tanzania. African Zoology 43: Creel, S., and N. M. Creel Lion density and population structure in the Selous Game Reserve: evaluation of hunting quotas and offtake. African Journal of Ecology 35: Creel, S., et al Conserving large populations of lions: the argument for fences has holes. Ecology Letters 16:1413 e3. Creel, S., et al Questionable policy for large carnivore hunting. Science 350: Croes, B. M., P. J. Funston, G. Rasmussen, R. Buij, A. Saleh, P. N. Tumenta, and H. H. De Iongh The impact of trophy hunting on lions (Panthera leo) and other large carnivores in the Bénoué Complex, northern Cameroon. Biological Conservation 144: Edwards, C. T., N. Bunnefeld, G. A. Balme, and E. J. Milner-Gulland Data- poor management of African lion hunting using a relative index of abundance. Proceedings of the National Academy of Sciences USA 111: Estes, J. A., et al Trophic downgrading of planet Earth. Science 333: Groom, R. J., P. J. Funston, and R. Mandisodza Surveys of lions Panthera leo in protected areas in Zimbabwe yield disturbing results: What is driving the population collapse? Oryx 48: Hanby, J. P., J. D. Bygott, and C. Packer Ecology, demography and behavior of lions in two contrasting habitats: Ngorongoro Crater and the Serengeti Plains. Pages in A. R. E. Sinclair and P. Arcese, editors. Serengeti II. University of Chicago Press, Chicago, Illinois, USA. Lindsey, P. A., G. A. Balme, P. Funston, P. Henschel, L. Hunter, H. Madzikanda, N. Midlane, V. Nyirenda, The trophy hunting of African lions: scale, current management practices and factors undermining sustainability. PLoS ONE 8:e Loveridge, A. J., A. W. Searle, F. Murindagomo, and D. W. Macdonald The impact of sport- hunting on the population dynamics of an African lion population in a protected area. Biological Conservation 134: Ogada, M. O., R. Woodroffe, N. O. Oguge, and L. G. Frank Limiting depredation by African carnivores: the role of livestock husbandry. Conservation Biology 17: Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell Trophic cascades revealed in diverse ecosystems. Trends in Ecology and Evolution 14: Packer, C., L. Herbst, A. E. Pusey, J. D. Bygott, J. P. Hanby, S. J. Cairns, and M. Borgerhoff Mulder Reproductive success in lions. Pages in T. H. Clutton-Brock, editor. Reproductive success. University of Chicago Press, Chicago, Illinois, USA.

11 October 2016 ASSESSING LION TROPHY HUNTING 2357 Packer, C., et al Sport hunting, predator control and conservation of large carnivores. PLoS ONE 4:e5941. Packer, C., H. Brink, B. M. Kissui, H. Maliti, H. Kushnir, and T. Caro Effects of trophy hunting on lion and leopard populations in Tanzania. Conservation Biology 25: Packer, C., et al Conserving large carnivores: dollars and fence. Ecology Letters 16: Quadling, H. S., and A. M. Starfield Exploiting object- orientated programming structures in the quest for an individual- based lion population model with an attractive user interface: Starfield Festschrift. South African Journal of Science 98: Riggio, J., et al The size of savannah Africa: a lion s (Panthera leo) view. Biodiversity and Conservation 22: Riggio, J., T. Caro, L. Dollar, S. M. Durant, A. P. Jacobson, C. Kiffner, S. L. Pimm, and R. J. van Aarde Lion populations may be declining in Africa but not as Bauer et al. suggest. Proceedings of the National Academy of Science USA 113:E107 E108. Rosenblatt, E., M. S. Becker, S. Creel, E. Droge, T. Mweetwa, P. A. Schuette, F. Watson, J. Merkle, and H. Mwape Detecting declines of apex carnivores and evaluating their causes: an example with Zambian lions. Biological Conservation 180: Rutledge, L. Y., B. R. Patterson, K. J. Mills, K. M. Loveless, D. L. Murray, and B. N. White Protection from harvesting restores the natural social structure of eastern wolf packs. Biological Conservation 143: Scheffer, M., S. Carpenter, J. A. Foley, C. Folke, and B. Walker Catastrophic shifts in ecosystems. Nature 413: Terborgh, J., and J. A. Estes Trophic cascades: predators, prey, and the changing dynamics of nature. Island Press, Washington, D.C., USA. U.S. Fish and Wildlife Service Endangered and threatened wildlife and plants; listing two lion subspecies. Federal Register 80: Watson, F., M. S. Becker, R. McRobb, and B. Kanyembo Spatial patterns of wire- snare poaching: implications for community conservation in buffer zones around National Parks. Biological Conservation 168:1 9. Watson, F. G., M. S. Becker, J. Milanzi, and M. Nyirenda Human encroachment into protected area networks in Zambia: implications for large carnivore conservation. Regional Environmental Change 15: Western, D., S. Russell, and I. Cuthill The status of wildlife in protected areas compared to non- protected areas of Kenya. PLoS ONE 4:e6140. Whitman, K. L., and C. Packer A hunter s guide to aging lions in eastern and southern Africa. Safari Press, Huntington Beach, California, USA. Whitman, K., A. M. Starfield, H. S. Quadling, and C. Packer Sustainable trophy hunting of African lions. Nature 428: Whitman, K. L., A. M. Starfield, H. S. Quadling, and C. Packer Modeling the effects of trophy selection and environmental disturbance on a simulated population of African lions. Conservation Biology 21: Yamazaki, K Social variation of lions in a maledepopulated area in Zambia. Journal of Wildlife Management 60: Supporting Information Additional Supporting Information may be found online at: Data Availability Data associated with this paper have been deposited in Dryad: