population density of the endangered florida key deer - BioOne

POPULATION DENSITY OF THE ENDANGERED FLORIDA KEY DEERROEL R. LOPEZ, 1 Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USANOVA J. SILVY, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USABRIAN L. PIERCE, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USAPHILIP A. FRANK, U.S. Fish and Wildlife Service, National Key Deer Refuge, Big Pine Key, FL 33043, USAMATTHEW T. WILSON, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USAKYLE M. BURKE, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843, USAAbstract: With decreased illegal hunting and better habitat conservation, the Florida Key deer (Odocoileus virginianusclavium) population grew from an estimated 25–50 animals in the late 1940s to approximately 200 animals onBig Pine and No Name keys, Florida, USA, by 1971, the last official survey. U.S. Fish and Wildlife Service (USFWS)trend data indicate that the deer population continued to increase after 1971; however, current deer density estimatesare necessary for the proposed reclassification of the Key deer from endangered to threatened. Our studyobjectives were to (1) obtain current population estimates of Florida Key deer and compare these to historical estimates,(2) evaluate survey methods (USFWS mortality and deer counts) in detecting changes in population trends,and (3) outline a protocol for future monitoring. Road counts (n = 889) were conducted from January 1971 toDecember 1971 and January 1976 to December 2001 on Big Pine and No Name keys. From mark–recapture data,we estimated that the Key deer population on these 2 islands increased by 240% between 1971 and 2001 (2001 estimate:453–517 deer). Trend data indicated that annual deer mortality was a function of deer density or populationsize (r s = 0.743). We compared the annual finite rate of increase (R) from USFWS annual deer counts and mortalitydata (R = 1.053–1.065) to mark–recapture studies (R = 1.038) and found them to be similar (P = 0.66–0.67).This similarity suggests that all 3 methods (USFWS deer counts and mortality data, and mark–recapture data) canbe used to monitor changes in Key deer density.JOURNAL OF WILDLIFE MANAGEMENT 68(3):570–575Key words: downlisting, endangered status, Florida Key deer, Odocoileus virginianus clavium, population density,recovery.The endangered Florida Key deer, the smallestsubspecies of white-tailed deer in the UnitedStates, is endemic to the Florida Keys on thesouthern end of peninsular Florida (Hardin et al.1984). Key deer occupy 20–25 islands within theboundaries of the National Key Deer Refuge(NKDR), with approximately 75% of the overallKey deer population on Big Pine and No Namekeys (Lopez 2001). Intense hunting pressure andhabitat loss resulted in a decline in the Key deerpopulation to an estimated 25–50 animals by thelate 1940s (Dickson 1955, USFWS 1999). With areduction in illegal hunting and the conservationof Key deer habitat through the establishment ofthe NKDR in 1957, the Key deer population grewto an estimated 200 animals on Big Pine and NoName keys by 1971, the last official survey (Silvy1975). In 1998, Texas A&M University (TAMU)and USFWS began studies to estimate the populationdensity of this federally protected subspecies.Trend data collected by USFWS biologistsindicated that the deer population had increased;however, current deer density estimates,necessary in the proposed reclassification of the1 E-mail: roel@tamu.edu570Key deer from endangered to threatened, arelacking (USFWS 1999).Key deer were marked as part of 2 separate fieldstudies (1968–1972, hereafter, historical study;and 1998–2001, hereafter, current study). In addition,USFWS biologists have collected deer mortalityand trend survey data since 1968. Collectively,we used these data to (1) obtain currentpopulation estimates and compare these to historicalestimates, (2) evaluate survey methods indetecting changes in population trends, and (3)outline a protocol for future monitoring.STUDY AREAThe Florida Keys are a chain of small islandsapproximately 200 km in length extending southwestfrom peninsular Florida, USA. Big Pine Key(2,548 ha) and No Name Key (461 ha) are withinthe boundaries of the NKDR and Monroe County,and they support approximately 75% of theKey deer population (Lopez 2001). Soils varyfrom marl deposits to bare rock of the ooliticlimestone formation (Dickson 1955). Typically,island areas near sea level (maritime zones) arecomprised of red mangrove (Rhizophora mangle),black mangrove (Avicennia germinans), whitemangrove (Laguncularia racemosa), and button-

J. Wildl. Manage. 68(3):2004 DENSITY OF KEY DEER • Lopez et al. 571wood (Conocarpus erecta) forests. With increasingelevation, maritime zones transition into hardwood(e.g., gumbo limbo [Bursera simaruba],Jamaican dogwood [Piscidia piscipula]) andpineland (e.g., slash pine [Pinus elliottii], saw palmetto[Serenoa repens]) upland forests with vegetationintolerant of salt water (Dickson 1955, Folk1991). Approximately 24% of native areas havebeen developed in the last 50 years (Lopez 2001).METHODSTrapping and MarkingKey deer were captured as part of 2 separate researchprojects conducted December 1968–June1972 (Southern Illinois University-Carbondale[SIU]; Silvy 1975) and January 1998–December2001 (TAMU; Lopez 2001) on Big Pine and NoName keys. Deer were captured with eitherportable drive nets (Silvy et al. 1975), drop nets(Lopez et al. 1998), and/or by hand (Silvy 1975).Deer were physically restrained after capture withan average holding time of 10–15 min (no drugswere used). Sex, age, capture location, bodyweight, radio frequency (if applicable), and bodycondition were recorded for each deer prior torelease. In addition, each animal captured wastattooed in an ear (Silvy 1975).Deer were marked in a variety of ways dependingon sex and age (Silvy 1975, Lopez 2001). Weused plastic neck collars (8-cm wide) primarilyfor females of all age classes, leather antler collars(0.25-cm wide) for yearling and adult malesonly, and elastic expandable neck collars (3-cmwide) primarily for male fawns/yearlings. A battery-powered,mortality-sensitive radiotransmitter(

572 DENSITY OF KEY DEER • Lopez et al.J. Wildl. Manage. 68(3):2004prior to necropsy or necropsied immediately. Carcassquality or ability to determine cause of deathranged from good to marginal (Nettles 1981,Nettles et al. 2002). Age, sex, body weight, andcause of death were recorded for each animalusing procedures described by Nettles (1981).Data AnalysisWe estimated density for Big Pine and NoName keys from SIU and TAMU survey datausing Lincoln-Petersen (Seber’s modification)and Schnabel estimators (White et al. 1982, Krebs1999). These survey routes met the requirementsfor both estimators because each survey adequatelycovered what was considered a closedpopulation (both areas are islands, with limiteddispersal between islands and small populationgrowth), and a segment of the population wasradiomarked when surveys were conducted.We compared population estimates betweenislands and time of day (sunrise, sunset) usingrepeated-measures analysis of variance (ANOVA;Tzilkowski and Storm 1993, SPSS 2001, von Ende2001). Repeated-measures ANOVA accounts forlack of independence when repeated observationsare obtained from the same experimentalunits (Zar 1996). We log transformed density estimatesand selected random samples within yearsto obtain a balanced design. Critical assumptions,such as normality and circularity within the variance–covariancematrix, were assessed prior tohypothesis testing. We tested for equality of errorvariances using Levene’s test, and we examinedtrends among years (within-subject factor) usingpolynomial transformations (orthagonal contrast;von Ende 2001). Following Lomax (2001),we reported conservative Greenhouse-Geisserprobabilities along with the standard F-statistic(Greenhouse and Geisser 1959). Last, we calculatedan overall Schnabel estimate (Krebs 1999)between islands and study (historical, current),and we compared these estimates to averagedLincoln-Petersen estimates.To evaluate Key deer population trends fromUSFWS survey and mortality data (1976–2000),we compared the average number of deer seenannually to annual deer mortality using a Spearman’srank correlation coefficient (Ott 1993). Wecalculated the annual finite rate of increase (R;Caughley 1977) for successive years for USFWSsurvey and mortality data, and we compared theaverage of these estimates to estimates of R fromSIU and TAMU survey data using a Student’st-test (Ott 1993, Minitab 1998).RESULTSTrapping and MarkingDuring December 1968–June 1972 and January1998–December 2001, we captured and marked434 Key deer (historical = 179, current = 255;fawn female = 41, yearling female = 43, adultfemale = 147, fawn male = 66, yearling male = 54,adult male = 83). The average number of markedanimals available on a weekly basis by island andstudy was 76 (range = 62–90) for Big Pine historical,48 (range = 34–64) for Big Pine current, and22 (range = 18–35) for No Name current.Approximately 67–75% of animals captured wereradiomarked during both studies (SIU =120/179, TAMU = 191/255).Density EstimatesWe conducted 889 roadside surveys (USFWS =266, SIU = 54, TAMU = 569; Table 1). From historicalweekly surveys (1971), we estimated 167 (95%CI: 149 to 185) and 34 (95% CI could not be calculated;Silvy 1975) Key deer on Big Pine and NoName keys, respectively (Table 1). From currentweekly surveys (1998–2001), we estimated 406(95% CI: 378 to 433) and 76 (95% CI: 72 to 80) Keydeer on Big Pine and No Name keys, respectively(Table 1). Conversely, Schnabel estimates indicated170 (95% CI: 154 to 191) Key deer on Big PineKey from historical surveys, and 390 (95% CI: 371to 411) and 79 (95% CI: 75 to 84) Key deer on BigPine and No Name keys, respectively, from currentsurveys (Table 1). Schnabel and weekly Lincoln-Petersen population estimates were similar (P >0.05; Table 1). Density estimates indicated theKey deer population on these 2 islands increasedby 240% (482/201) between 1971 and 2001.The repeated-measures ANOVA was conductedusing a balanced design that maximized thenumber of surveys within each island × year × surveyperiod combination. One data point (BPKsunset; estimate = 6.5 deer) was considered anoutlier and removed from consideration. Becauseno other data were available, the 1971 populationestimate of 34 (Silvy 1975) was used for No NameKey and resulted in a loss of homogeneity for thatyear (Levene’s test; P = 0.004). All other yearsdemonstrated equality of error variances by Levene’stest (P >0.132 for all remaining years). Therepeated-measures ANOVA results indicated noyear × survey (P = 0.743), year × island (P =0.294), or year × survey × island (P = 0.304) interactions;therefore, population estimates amongyears were not influenced by either time of survey

J. Wildl. Manage. 68(3):2004 DENSITY OF KEY DEER • Lopez et al. 573Table 1. Florida Key deer density estimates for Big Pine andNo Name keys, Florida, USA, 1971 and 1998–2001.Year Time n – x 95% CIPetersen estimatesBig Pine Key1971 sunrise 26 141 123 to 1581971 sunset 28 192 164 to 2191998 sunrise 36 447 364 to 5301998 sunset 36 366 322 to 4101999 sunrise 48 400 350 to 4491999 sunset 48 332 289 to 3762000 sunrise 32 498 415 to 5802000 sunset 33 408 331 to 4862001 sunrise 7 571 159 to 9832001 sunset 7 342 212 to 473historical sunrise 26 141 123 to 158sunset 28 192 164 to 219combined 54 167 149 to 185current sunrise 123 449 405 to 493sunset 124 363 332 to 393combined 247 406 378 to 433No Name Key1998 sunrise 36 105 94 to 1161998 sunset 36 81 73 to 891999 night 43 63 52 to 731999 sunrise 48 82 68 to 961999 sunset 48 74 60 to 882000 night 29 72 57 to 862000 sunrise 31 63 53 to 742000 sunset 34 69 59 to 782001 night 2 46 38 to 542001 sunrise 7 97 51 to 1432001 sunset 8 77 52 to 101historical a combined 34current sunrise 122 85 57 to 93sunset 126 75 68 to 81night 74 66 57 to 74combined 322 76 72 to 80Schnabel EstimatesBig Pine Keyhistorical combined 54 170 154 to 191current combined 247 390 371 to 411No Name Keyhistorical combined 34current combined 322 79 75 to 84a Raw data from historical surveys for No Name Key wereunavailable for analyses; however, Silvy (1975) reported anestimate of 34 deer from mark–recapture efforts for this islandin 1973, and we used this estimate in our comparison whenappropriate.or island. The main effect for year was significant(P

574 DENSITY OF KEY DEER • Lopez et al.J. Wildl. Manage. 68(3):2004The similarity in results between the mark–recaptureestimates (R = 1.038), USFWS road-survey data(R = 1.053), and mortality data (R = 1.065) suggestthat all 3 methods can be used to monitor changesin Key deer trends. In particular, the USFWS roadsurveyand mortality data are useful in monitoringpopulation changes of this federally protected subspeciesbecause data are collected each year byUSFWS biologists. We recommend the continuationof both of these 25-plus year data sources.Fig. 2. Average number of deer seen on U.S. Fish and WildlifeService monthly road counts and annual deer mortality on BigPine Key, Florida, USA, 1976–2000.illegal hunting as the primary reasons for theincrease (Lopez 2001). For example, in the last30 years, NKDR and other natural resource agencieshave purchased and managed nearly 64% ofthese 2 islands for Key deer conservation (Lopez2001). Furthermore, documented Key deerpoaching has decreased nearly 10-fold since the1970s (USFWS, unpublished data). We proposethat urban development also might be responsiblefor the increase in Key deer numbers. Forexample, since the establishment of NKDR in 1957,approximately 717 ha (24%) of native habitats onBig Pine and No Name keys were developed(Lopez 2001) in concert with the Key deer populationincrease. The results of early development(prior to mid-1980s) probably benefited Key deerby the conversion of tidal areas such as mangroveand buttonwood forests into “uplands” (Gallagher1991). Initially, the conversion and filling of thesehabitats provided Key deer with both native andornamental vegetation (Lopez 2001). For thesereasons, we propose that urban development in theform of land clearing increased the overall carryingcapacity for Key deer on these 2 islands. Continueddevelopment in the form of houses and/orbusinesses, however, might not provide the samebenefits as increases in secondary impacts (i.e.,road mortality, habitat loss, fence entanglement)also would be expected (Lopez et al. 2003).Population TrendsIn comparing annual R estimates for the Keydeer population, we found USFWS trend datawere similar to and validated SIU/TAMU estimates.Since 1971, we estimate that the Key deerpopulation has grown approximately 5% annually.MANAGEMENT IMPLICATIONSThe decision to consider downlisting the Keydeer is based on current deer density estimatesand recovery criteria that state an R ≥1.0 for a 14-year period would warrant reclassification(USFWS 1999). As a result of our study, the reclassificationof the Key deer from endangered tothreatened has been proposed and currently isbeing considered by USFWS (Jay Slack, EcologicalServices Field Supervisor, USFWS, personalcommunication). Other recovery criteria thatshould be addressed in the next several years includethe establishment of other local populationsof Key deer on outer islands. Overall, effortsby USFWS, state agencies, and conservation organizationshave been instrumental in the first steptoward Key deer recovery.ACKNOWLEDGMENTSThanks to TAMU student interns who assistedin the collection of field data, and J. Morgart, C.Faulhaber, N. Wilkins, W. Grant, T. Peterson, and2 anonymous reviewers for constructive criticismin the preparation of this manuscript. Funding wasprovided by TAMU System, Rob and Bessie WelderWildlife Foundation, Florida Fish and WildlifeConservation Commission, and USFWS (SpecialUse Permit No. 97-14). Special thanks to the staffof the National Key Deer Refuge, Monroe County,Florida. This manuscript is supported by theWelder Wildlife Foundation, Contribution 611.LITERATURE CITEDCAUGHLEY, G. 1977. Analysis of vertebrate populations.John Wiley & Sons, New York, New York, USA.DICKSON, J. D., III. 1955. An ecological study of the Keydeer. Florida Game and Fresh Water Fish Commission,Technical Bulletin 3, Tallahassee, Florida, USA.FOLK, M. L. 1991. Habitat of the Key deer. Dissertation,Southern Illinois University, Carbondale, Illinois, USA.GALLAGHER, D. 1991. Impact of the built environmenton the natural environment. Pages 226–229 in J. Gato,editor. Monroe County environmental story. MonroeCounty Environmental Education Task Force, BigPine Key, Florida, USA.

J. Wildl. Manage. 68(3):2004 DENSITY OF KEY DEER • Lopez et al. 575GREENHOUSE, S. W., AND S. GEISSER. 1959. On methods inthe analysis of profile data. Psychometrika 24:95–112.HARDIN, J. W., W. D. KLIMSTRA, AND N. J. SILVY. 1984.Florida Keys. Pages 381–390 in L. K. Halls, editor.White-tailed deer: ecology and management. StackpoleBooks, Harrisburg, Pennsylvania, USA.HUMPHREY, S. R., AND B. BELL. 1986. The Key deer populationis declining. Wildlife Society Bulletin 14:261–265.KREBS, C. J. 1999. Ecological methodology. Second edition.Addison-Welsey Publishers, Menlo Park, California,USA.LOMAX, R. G. 2001. Statistical concepts: a second coursefor education and the behavioral sciences. Secondedition. Lawrence Erlbaum Associates, Mahwah, NewJersey, USA.LOPEZ, R. R. 2001. Population ecology of Florida Keydeer. Dissertation, Texas A&M University, College Station,Texas, USA.———, N. J. SILVY, J. D. SEBESTA, S. D. HIGGS, AND M.SALAZAR. 1998. A portable drop net for capturingurban deer. Proceedings of the Southeastern Associationof Fish and Wildlife Agencies 52:206–209.———, M. E. VIERA, N. J. SILVY, P. A. FRANK, S. W.WHISENANT, AND D. A. JONES. 2003. Survival, mortality,and life expectancy of Florida Key deer. Journal ofWildlife Management 67:34–45.MINITAB. 1998. Minitab user’s guide. Version 12. Fourthedition. Minitab, Inc., State College, Pennsylvania,USA.NETTLES, V. F. 1981. Necropsy procedures. Pages 6–16 inW. R. Davidson, F. A. Hayes, V. F. Nettles, and F. E.Kellogg, editors. Diseases and parasites of white-taileddeer. Tall Timbers Research Station, Tallahassee,Florida, USA.———, C. F. QUIST, R. R. LOPEZ, T. J. WILMERS, P. FRANK,W. ROBERTS, S. CHITWOOD, AND W. R. DAVIDSON. 2002.Morbidity and mortality factors in Key deer, Odocoileusvirginianus clavium. Journal of Wildlife Diseases38:685–692.OTT, R. L. 1993. An introduction to statistical methodsand data analysis. Fourth edition. Duxbury Press, Belmont,California, USA.SILVY, N. J. 1975. Population density, movements, andhabitat utilization of Key deer, Odocoileus virginianusclavium. Dissertation, Southern Illinois University,Carbondale, Illinois, USA.———, J. W. HARDIN, AND W. D. KLIMSTRA. 1975. Use ofa portable net to capture free-ranging deer. WildlifeSociety Bulletin 3:27–29.SPSS. 2001. SPSS advanced models. Version 11. SPSS,Inc., Chicago, Illinois, USA.TZILKOWSKI, W. M., AND G. L. STORM. 1993. Detectingchange using repeated measures analysis: white-taileddeer abundance at Gettysburg National Military Park.Wildlife Society Bulletin 21:411–414.USFWS (U.S. FISH AND WILDLIFE SERVICE). 1999. SouthFlorida multi-species recovery plan: Key deer recoveryplan. U.S. Fish and Wildlife Service, Atlanta,Georgia, USA.VON ENDE, C. N. 2001. Repeated-measures analysis:growth and other time-dependent measures. Pages134–157 in S. M. Scheiner and J. Gurevitch, editors.Design and analysis of ecological experiments.Oxford University Press, New York, New York, USA.WHITE, G. C., D. R. ANDERSON, K. P. BURNHAM, AND D. L.OTIS. 1982. Capture-recapture and removal methodsfor sampling closed populations. Los Alamos NationalLaboratory, Los Alamos, New Mexico, USA.———, AND R. A. GARROTT. 1990. Analysis of wildliferadio-tracking data. Academic Press, San Diego, California,USA.ZAR, J. H. 1996. Biostatistical analysis. Third edition.Prentice Hall, Upper Saddle River, New Jersey, USA.Received 5 January 2003.Accepted 16 February 2004.Associate Editor: Morgart.