1 ; lCES CM 19971P:1O Diadr:imous Fish EXtinction: 1fueats on LOCal and Global Scales . .~. ~ , , , RISK FACTORS IN EXTINCTION AND OVEREXPLOITATION RUSSELL LANDE (l) RiskS ofpopulation extinctiori from a variet)r ofdetenninistic and stochastie ~ ...... factors are eategonzed. Anthropogenie factors generany eonstitute the primary eauses ofendangennent and extinction..The prim:iry anthropogenie factors have :rn.muying • ecological änd genetie effects. All factors 3.ffecting extinction risk are expressed. änd ean be evaluated. through their operation on population dynamies. (2) Results ofa quantitative population,viability analysis ofspring ehinook saltnori in the South Umpqua River, Oiegon, indicate that linder ewrent habitat conditions tlle wild population has a high probability ofperSistence for 200 years. However, ifhabitat ' degradation contimies at historical rates, tlle population will altnost'certainly become extinet withm 100 years. (3) Threshold harvesting strategies for sustamable eXploitation offluctuating populations are diseussed in comparisori with commonly . used flshenes practices. Based on recent analytieaI results itis coricluded thllt the thooretieal justifleation for constant harVest rate strategies is extremely we3.k in eomparison to thai for optinial threshold strategies. Ke)rwords: extinetion, overexploitation, population viability analysis; stochasticity, thn~shold harvesting, uneertainty RussezlLtiluie:Dept. ofBiology. University ofOregon,'Eugene OR 97403-1210 USA L- • [tel: +1'. 54i 346 2697.fax: +1 541'.3462364, email: [email protected]öregon.edul. • , • j',. • ,~. "- ANTHROPOGENIe, ECOLOGlCAL AND GENETIC FACfORS IN EXTINCfION RISK A varieiy ofdetenninistie and stochasrle factorS contlibute to extinction risk. The folloWing list orders risk faetors by descending general iniponance aInong and within categories, although the fankirig offactors differs 'among speeies arid should be evaluated on a ease-by-ease basis. AnthropÜgeme factors generally eonstitute theijrirriary eauses ofendangerment and extinction. The Priiriäi:Y anthrOpogenie factors have raInifying ecolpgieal and genetic effects. All factors , affecting extirierlon risk are expressid. and ean be evaJuated. through their operation on population dynamies. With sufflcient data on aparticular popUlation or species. a quantitative PopuIätion Viability Analysis (PVA) ean be peiforined. inc1uding stochastic factors affecting the dynamics of small populations. A PVA estimates the probablliiy ofpopuhition extiriction orcollapse.WithiIl ä specified ume (Sh3.ffer 1981, Gilpin an~ Soul6 1986). For species with insufflcient daci to ..--------------------------------- 2 perfonn a quantitative PVA, objective population-baSed criteiia can be employed to qualltatively categorize extinetion risk (Maee and Lande 1991, IUCN 1994). Foradditional details and referenees see Lande (1998). ANTIIROPOGENIC FACfORS Land development: habitat destruetion & fragmentation Overexploitation Unregulated competition Eeonomie diseounting Species translocations & introduetions . Pollution . ',"' ECOLOGlCAL FACfORS Environmental fluetuations & eatastrophes Metapopulation dynamies Oocal extinction & eolonization) Small population size • Allee effect Edge effects Demographie stochasticity GENETIC FACfORS J Hybridization with nonadapted gene pools Seleetive breeding and harvesting Small population size: " Inbreeding depression . Loss ofgenetie variability Fixation ofriew mutations ~ 1. >, POPULATION yrIABILITY ANALYSIS OF SPRING CHINOOK SALMON IN THE SOUTH UMPQUA RIVER, OREGON A comprehensive qualitative assessment ofPaeific sahIlon by Nehlsen et al. (i991)"showed. • that about one-quarter oftlie histOIical salmon runs are now extinet, an equal number are at high risk ofextinetion, and most ofthe rest are in serious decline. Totest the validitY ofthe qualitative assessment methods used by Nehlsen et al (1991) and proposed by Allendorfet al. (1997), we . performoo a quantitative PVA on a particular run for whieh relatively extensive popUlation data were available. Run sizes ofspring chinook salrrion in the S01ith Umpqua Riverin Oregon have deeliried dramatieally sinee the early part ofthis eeritury. Habitat degradation is though to be an important factor in eontributing tothe deeline ofthis stock, arid qualit3.tive assessment suggests thai the stock is at moderate risk of extinction~ Data from this and similar stocks were used to develop an age-struetured, density-dependerii mOdel ofthe population dymimies thiu incoi-porates both demographie and envirOnrriental stochasneity. Under the assumption ofrio fuIther habitat degrad3tion, the population is prewctoo 10 have a greater than 95% probabilityofpersistence far 200 years. However, sensitiVity analysis for the derisHy-dependenee estimated from. historieal run-retUrn data shows that substantially lower predicted viabilities are also statistieally consistent with the data..A model that simulates eontinued habitat degradation results in almost certain 3 J exiiricrlon witlun 100 years. This firiding supports the conclusions oftheeariier qualitative, , assessment (Nehlson et al. 1991) thai this stock is at niOderilte risk ofextinction. For additional details see Rattler et al. (1997). " , 1HRESHOLD HARVESTING FOR SUSTAINABILITY OF FLUcruATING RESOURCES Although temPoral varlabÜitY is a ubiqwtous feature ofrenewable resoUrce dynam1cs, it has Teceived serious attention in harvesting theory onlydUrlng the past two decades. Coefficients of variation in änrimil abimdance for unexploited vertebrate populations typiCaily are in the range of 20% to 80% ormore (Pimm 1991). Exploited.populations also are highly variable (Myers et al. ' 1995), not onlydue to natural environmeriiaI stochasticity, but because exploicition usually reduces popUlation stability (Beddington and May i977). Ifharirested populations beCome small, they aTe • vulrierable to demographie stochasticity arising from chanee eventS ofindividual. niortality and reproduetion (Lande 1993). Environmental and demographie stOchastieity together guarantee that extirietion is the eventUal fate ofall species (reg3.idless of whether they are harvested), and evidenee from paleontology indieates thai the vast majority ofspecies that ever existed are now extinet. Unfortunateiy, exploitation cari aceelerate this seemingly ineVitable march to extinction. in pärticular, modem technology arid rapidly iricreasing reoource demandS from the growing hunian population, coupled with madequaie management and regulation, have causect rnany eonmiercially importlnt species to be overexploited to the Pointofdepletion; and occasionally 10 extinction (Ludwig et al. 1993, Rosenberg et al. 1993, Groom1Jridge 1992). AbOut half ofthe ,commercial fisheries in the Uniied States and Europe were receritly dassified,as overexploited (Rosenberg et al. 1993)., Teiresirial habitat alteratioris, primarily by forest deaiing and agrlculture, have ., , threaiened a large proportion ofveriebraie species in develojx;d eountiies; approxim3.tely one-third ofthe endangered birds and one-half of the endangered m3.mmals ofthe wofld are threatened with extinetion by hwiiing and international trade (Grocimbridge 1992, Redford 1992).. ' Despiie these observations, imtil receritly, little effort has beeil made tri understand how , • stoehastie Population dynamies affeets the risk:. ofresource collapse or extinction (Getz and Haight 1989). Laride et al' (1997) review recent developments toward a unified 'theory ofoptiriial härvesnng that explicitly iricorpcirates the risk ofdepletion or extinction offluctuating resources. Tbey focus ori analytical results pertaining io biologiCal criteria oflong-term sustainability, eombming approaches from conservation biology arid ti-aditiorial harvesiing theory. Theyshow .that for alarge dass ofpopulation dynamics arid arange ofbiological optirillzation ciiteria the optimal strategies aIways irivolve a threshold population size above whieh an excess individuals are haivesied and below whieh there is no harvesting. Insofar as possible, threshold harvesting strategies simultaneously increase expeeted yields from harvestirig arid reduee risk ofresource depletion or extinction. Tbe position of the threshold in this management approach detennines the level ofrisk. ,Tbe major drawback of threshold strategies is that they eniail ahigh vananee in . annual yield beeause, offrequerit years ofno harvest when the population is below the t:hieshold. Lande et aI. (1997) elucidate generaIized thTeshold strategieS that mitigate this drawback by , reducing the variance in anilUa! yield th.rOugh harVesting stnitegies that aCcOurit for limited haiVesting 'capacity ärid ~nceItaintY in populatiori estiniates. 4 Biological resources often are exploited using a constant effort or constant harvest rate strategy combined With a t.h.feshold or "escapement" level beIow which harvesting ceases. However. thresholds frequently are set far too low to sustain the resources. with harvesiing stopped omy arter severe depletion, as evidenced bythe large fuicition oföverexploited resoirrces on which harvesting coritinues, and numerous others that havecollapsed (Ludwig et al. 1993, Rosenberg et al. 1993, Hutchirigs arid Myers 1994). Lande et al. (1997) argue that the thecireticaljusiificarlon for conscini harvest raie strategies is extremely weaIc in comparison to thai for optiIrial threshold strategies. Opilirial threshold strategies have the importarii propenythat. insofar as possible. they , simultaneouslyreduce the nskofdepletion or