Transcript
Jaguar in Brazil ARNAUD L. J. DESBIEZ 1 , KATHY TRAYLOR-HOLZER 2 , BOB LACY 2 , BEATRIZ M. BEISIEGEL 3 ,CHRISTINE BREITENMOSER-WÜRSTEN 4 , DENIS ALESSIO SANA 5 , EDSEL A. MORAES JR 6 ,ELILDO A. R. CARVALHO JR 6 , FERNANDO LIMA 7 RICARDO L. P. BOULHOSA 5 , ROGERIO C. DEPAULA 8 , RONALDO G. MORATO 3 , SANDRA M. C. CAVALCANTI 5 AND TADEU G. DE OLIVEIRA 9 Population Viability Analysisof jaguar populations in Brazil Population viability analysis (PVA) was used during the workshop for the JaguarNational Action Plan to better understand jaguar population dynamics and simulatedifferent scenarios to understand the impact of threats and projected outcome of po- tential conservation strategies. The method is explicitly designed to broaden stake-holder involvement and enhance information sharing across disparate scientific andsocial domains. During the Jaguar National Action Plan workshop a base modelwas built for jaguars, a sensitivity analysis was run, and theoretical case studiesof questions and situations raised by the participants were developed. The focus of this work was to examine concepts of jaguar population dynamics, stimulate discus-sions on jaguar life history parameters, fuel discussion on different threats, evaluatepotential impact of these threats, and introduce participants to concepts of popula- tion viability analysis and its value as conservation planning tool. Small populations of animals are at risk ofextinction not only due to threats such as ha-bitat loss and poaching, but are also particu-larly vulnerable to the impacts of stochastic(chance) processes that can lead to extinc-tion. During the Jaguar National Action Planworkshop the simulation software programVortex (v9.96) was used to examine the via-bility of jaguar populations. Vortex simulatesthe effects of deterministic forces as wellas demographic, environmental, and geneticstochastic events on wild or managed popu-lations, making it well suited to assess theviability of small populations. Vortex modelspopulation dynamics as discrete sequentialevents that occur according to defined proba-bilities. The program begins by creating indivi-duals to form the starting population and thensteps through life history events (e.g., births,deaths, dispersal, catastrophic events), typi-cally on an annual basis. Population attributessuch as breeding success, litter size, sex ratioat birth, and survival rates are determinedbased upon designated probabilities that areestablished during the workshop based on theliterature and participant expert knowledge.Specific events that occur in the lifetime of anindividual are uncertain, so probabilities areused to determine what happens to each ani-mal in each simulated future. Consequently,each run (iteration) of the model gives a diffe-rent result. By running the model hundreds oftimes, it is possible to examine the probableoutcome and range of possibilities for the fu-ture of a population.For a more detailed explanation of Vortex andits use in population viability analysis, seeLacy (1993, 2000) and Miller & Lacy (2005).Population viability analysis (PVA) enablesworkshop participants to better understandjaguar population dynamics and simulatedifferent scenarios to understand the impactof threats and projected outcome of potentialconservation strategies. The method is expli-citly designed to broaden stakeholder involve-ment and enhance information sharing acrossdisparate scientific and social domains. Eachparticipant is encouraged to add his or herknowledge of the species and its situation,potential actions, and additional questionsto develop scenarios to be examined throughmodeling.During the Jaguar National Action Plan work-shop a base model was built for jaguars, asensitivity analysis was run, and theoreticalcase studies of questions and situations rai-sed by the participants were developed. Thefocus of this work was to examine conceptsof jaguar population dynamics, stimulate dis-cussions on jaguar life history parameters,fuel discussion on different threats, evaluatepotential impact of these threats, and intro-duce participants to concepts of populationviability analysis and its value as conserva-tion planning tool. During the workshop theviability of jaguar populations from the diffe-rent biomes was NOT examined, but will beinvestigated in the near future based on thisinitial work and the maps developed duringthe workshop (Ferraz et al. 2012, this issue). Base model Due to the potential variation of several pa-rameters among the different biomes it wasdecided to construct a general base modelfor jaguars that could then be tailored tospecific Brazilian biomes and specific jagu-ar populations. The base population modelwas designed to investigate the viability of anon-specific but biologically accurate jaguarpopulation. Details of the parameters used inthe base model are available in Table 1.Some parameters were debated at lengthby workshop participants. For example someparticipants felt that females could have theirfirst litter at 2 years of age and males at 3,while others insisted females on averagewould have their first litter at 3 years of age,as they need to have an established territo-ry and be in good physical condition. Therewas also considerable debate about whetherjaguar reproduction is density dependent. Athigh densities some participants thought thatanimals will compete for prey, territories andmates, limiting reproduction. However, othersfelt that in solitary living carnivores reproduc-tion is not necessarily density dependent. Avalue of the PVA modeling approach is thatit helps to identify such different perspec-tives, and then allows sensitivity testing ofthe effects of alternative values without pre-judging which is closer to the truth.After some discussion, maximum age wasset at 15 years; however mortality rates af-ter 10 years were increased in the model sothat very few individuals (~5%) actually reachsuch an old age. By comparing observedpopulation growth rates to those calculatedfor different plausible values of age-specificmortality, the workshop participants wereable to come to consensus around the bestestimates of mortality to use in the models.Mortality rates were set highest for the firstyear of life, moderate during years when cubsare with the mother or dispersing, and lowestfor prime age adults, with increasing mortali-ty after 10 years of age.The base model represents the biologicalpotential of jaguars: no harvest, no increasein mortality due to road kill, disease or fire,and no catastrophes are included. This doesnot represent a realistic situation, but pro-vides the basis upon which future modelsthat include these and others threats can beconstructed.The demographic rates (reproduction andmortality) included in the base model can beused to calculate deterministic characteri-stics of the model population. These values 35 Population Viability Analysis CATnews Special Issue 7 Spring 2012 are a good initial summary of the populationcharacteristics, as they reflect the biology ofthe population in the absence of stochasticfluctuations (both demographic and environ-mental variation), inbreeding depression,limitation of mates, and immigration/emi-gration. The base model results in a deter-ministic growth rate for females of (rdet) of0.060 (λ= 1.058). This represents an annualpotential growth rate of almost 6%. Adultsex ratio is female biased and the sex ra-tio of adult males to adult females is 1:2.7.Adult individuals (sexually mature individu-als) represent 51% of the population (SOMTable 2 at www.catsg.org/catnews).Results from the base model (500 iterations)project that a population of 200 jaguars inthe absence of threats is likely to persistover the next 100 years. When N=100 andK=200 the stochastic growth rate (r-stoc,the mean rate in the simulated populations,subjected to all the demographic, genetic,and environmental uncertainty in the model)is 0.027, representing an annual populationgrowth of almost 3%, enabling the popula-tion to grow when below carrying capacity.There is zero probability of extinction (PE) in100 years, and the mean population size at100 years is 187 jaguars with 91.28% genediversity remaining (Supporting Online Ma-terial SOM Fig. 1). Sensitivity analysis of demographic rates Sensitivity analysis is a tool used to evalua-te the robustness of a model to variations inparameter values. The most sensitive para-meters require greater certainty in the inputvalues to produce more confident results. Thisalso helps to identify where further researchis needed. Sensitivity analyses using highestand lowest values for each demographic ratewere compared to evaluate the effect of mo-del parameters on the stochastic growth rate(r-stoc) of jaguar populations. Mortality rateswere increased and decreased by 25%, 1 yearwas either added or subtracted to age of firstreproduction, 4 years were added/subtractedto maximum age of reproduction, and ave-rage litter size was increased or decreasedby 25%. Results from the sensitivity analysisindicate that reproductive parameters andfemale mortality rates are the most sensitiveparameters across the values tested, whilemale demographic values are less sensitive(Supporting Online Material SOM Fig. 2).This is logical for a polygynous species, inwhich females represent the breeding poten-tial and therefore the ability of the populationto grow and recover from declines. Theoretical case studies Importance of population size To illustrate the importance of populationsize in jaguar populations, a modeling exer-cise was run during the workshop in whichthe initial population (N) was varied as wellas the carrying capacity (K) when K>N andwhen K=N. Many different values were te-sted, ranging from 15 to 200. This exercisedemonstrated that population size (both interms of N and K) is a very important factorin determining the population growth, long-term persistence and genetic diversity of ja-guar populations (SOM Tables 3 and 4). Smallisolated populations of jaguars cannot persistin the long term. However, a high carrying ca-pacity may decrease the impact of small po-pulation size on population growth, long-termpersistence, genetic diversity and mean timeto extinction, as it may allow the populationto grow to a larger size, and once it is larger,it is less vulnerable. Therefore for conservati-on purposes, protection and maintenance ofhabitat quantity and quality (cover and preybase), which determines carrying capacity,are imperative for the long-term conservationof jaguars. Impact of harvest of adult females Killing of adult jaguars through sport hunting,traditional hunting or retaliation for economicloss all have the same result: loss of adult in-dividuals (breeders) from the population. Thesensitivity analysis showed that an increasein mortality of females negatively impactsjaguar populations. For the purpose of thisexercise we modeled removal of adult fe-male jaguars. Results show that the smallerthe initial population, the higher the impactharvesting of female jaguars will be (Fig. 1).Smaller populations have higher probabili-ties of extinction and lower growth rates. Inpractical terms this means that removal of fe-male jaguars from small isolated fragmentedpopulations will have a much bigger impacton the population viability than removal of thesame number of individuals from a larger po-pulation. Overall jaguars cannot sustain highlevels of harvest given a theoretical maximumgrowth rate of only about 5%. Even if the ini-tial population size is high, jaguar populationgrowth rates will decline when harvest incre-ases, and may ultimately drive populations toextinction. Impact of fragmentation Habitat loss, fragmentation and isolation ofjaguar populations were repeatedly menti-oned as one of the main threats to long-termpersistence of jaguar populations in Brazil.Amongst its many impacts, habitat fragmen-tation reduces population size and smallerpopulations are more vulnerable to stochasticprocesses (including inbreeding) and have ahigher risk of extinction. Corridors that linkfragmented habitats are often advanced asa potential solution. During the workshopmodels were created to test the impact of Table 1. Summary of parameter input values used in the base model; EV =environmental variation, expressed as a standard deviation. Details for these valuesare given in the action plan (de Paula et al. 2010). ParameterBase value Breeding systemPolygynyAge of first reproduction (λ / λ) in years3 / 4Maximum age (in years) 15% adult males in breeding pool 90Density dependent reproduction?debatedAnnual % adult females reproducing (EV) 50 (5)Average litter size2Maximum litter size4Overall offspring sex ratio 50:50% mortality from age 0-1 (EV) (λ / λ)42(7)/ 42(7)% mortality from age 1-2 (EV) (λ / λ)17(3.5)/ 17(3.5)% mortality from age 2-3 (EV) (λ / λ)20(5)/ 20 (5)% mortality from age 3-4 (EV) (λ / λ)6(1.5)/ 25 (6)% mortality from age 4-10 (EV) (λ / λ)8(1.5) / 10(2)% mortality from age 10-15 (EV) (λ / λ)Add’l 5 % mort. ea. yr.Inbreeding depression6 lethal equivalents% of inbreeding effect due to recessive lethal alleles50 36 Desbiez et al. Jaguar in Brazil corridors. Models suggested that corridorscan either prevent or cause the extinction ofjaguar populations, depending upon the situ-ation. Metapopulation dynamics are complexand many factors come into play such as sizeof fragments, dispersal rate, demographic ra-tes in the various fragments, and the survivaland reproductive rates of dispersing animals(SOM Figure 4). Corridors to poor-quality orunprotected habitat (“sinks”) or to areas toosmall to harbor a healthy breeding populationcan further de-stabilise the overall populati-on. Further exploration and caution should beused when considering corridors as a conser-vation measure. Translocations and re-introductions During the workshop jaguar translocationsand re-introductions were discussed. Due tologistical difficulties, potential risks and highcost, there is a lot of controversy surroundingthis conservation option. As an exercise du-ring the workshop we tested some re-intro-duction/translocation scenarios (SOM Table5). The modeling showed that there are manyaspects to be explored in order to guide andformulate a re-introduction program, such asage, sex, number of animals, reintroductioninterval, re-introduction time period, survivaland fecundity rates of reintroduced animals,and many more. The modeling exercise sho-wed that re-introductions need to be wellplanned and part of a comprehensive programto be effective tools for conservation of jaguarpopulations. Real case studies During the workshop we also investigated se-veral real case studies. The impact of huntingon a jaguar population from the Tapajós-Ara-piuns Extractive Reserve, Central Amazoniawas investigated using data collected by Eli-ldo A. R. Carvalho Jr. (ICMBIO Parna GrandeSertão Veredas). Modeling was first used toevaluate the importance of some of the datagathered during the field study. For example,the model showed that data on the sex ofanimal hunted had a significant impact on thefinal outcome of the model, while knowledgeof the age class of jaguars hunted (adults orsub-adults) had less impact. Modeling wasthen used to predict the impact of huntingand source-sink dynamics between the reser-ve and surrounding areas. The viability of theMinas Gerais Espinhaço jaguar population inthe Cerrado was investigated, and a wide ar-ray of conservation measures to protect thepopulation was tested using data collectedby Edsel Amorim Moraes Jr and Rafael LuizAarão Freitas (Instituto Biotropicos). TadeuGomes de Oliveira (UEMA & Pro-Carnivoros)modeled the importance and impact of pro-tecting the Nascentes Parnaíba Jaguar popu-lation in the Cerrado. The long-term viabilityof jaguar populations from each biome is cur-rently being investigated and will be used tomake specific conservation recommendationsfor each biome. Conclusion Using computer models such as Vortex duringa workshop helps to integrate detailed dataon species biology, genetics, and ecology withestimates of human-based threats to evalua-te the risk of wildlife population decline or ex-tinction under alternative future managementscenarios. One of the advantages of usingVortex in a workshop is that it is a participato-ry exercise that helps participants understandlong-term impacts, threats and probable tra-jectories of animal populations. It also helpsexperts determine the state of knowledge onthe species and pinpoint areas where futureresearch is needed. It helps to extract impor-tant data and knowledge from all participantsand advance knowledge of the species. A mo-del cannot make value decisions about whatto conserve and why, nor can it guaranteethat the needed actions will be implemented,but modeling does empower participants bygiving them a scientifically sound methodto integrate their knowledge into a compre-hensive picture of population dynamics, riskanalysis, and assessment of options. Thus, itcan serve as an excellent tool for scientistsand wildlife managers to work together intheir quest to make better decisions aboutconservation. References de Paula R. C., Desbiez A. L. J., Cavalcanti S. M.C. 2010. Plano de Ação para a Conservaçao daOnça-Pintada ( Panthera onca ). Serie EspeciesAmeaçadas. ICMBio. Brasilia.Lacy R. C. 1993. Vortex: A computer simulationmodel for population viability analysis. WildlifeResearch 20, 45-65.Lacy R. C. 2000. Structure of the Vortex simulationmodel for population viability analysis. Ecologi-cal Bulletins 48, 191-203.Miller P. S. & Lacy R. C. 2005. Vortex: A StochasticSimulation of the Extinction Process. Version9.50 User’s Manual. Apple Valley, MN: IUCN/SSC Conservation Breeding Specialist Group.Supporting Online Material SOM is available atwww.catsg.org/catnews 1 Royal Zoological Society of Scotland; IUCN/SSCConservation Breeding Specialist Group Brasil<
[email protected] 2 IUCN/SSC CBSG 3 CENAP – ICMBIO 4 IUCN/SSC Cat SG 5 Pró-Carnivoros 6 ICMBIO Parna Grande Sertão Veredas 7 Instituto de Pesquisas Ecológicas 8 CENAP/ICMBio and Pro-Carnivoros 9 Universidade Estadual do Maranhão and Pró-Carnivoros Fig. 1. Probability of extinction (within 100 years) of jaguar populations in re-lation to initial population size and percent harvest (H) of adult female jaguars. 37 Population Viability Analysis
