-
J Appl Ecol. 2019;56:1839–1849.
wileyonlinelibrary.com/journal/jpe | 1839© 2019 The Authors.
Journal of Applied Ecology © 2019 British Ecological Society
Received:31January2019 | Accepted:10April2019DOI:
10.1111/1365-2664.13419
R E S E A R C H A R T I C L E
Remnant forest in Costa Rican working landscapes fosters bird
communities that are indistinguishable from protected areas
Daniel S. Karp1 | Alejandra Echeverri2 | Jim Zook3 | Pedro
Juárez4 | Alison Ke1 | Jaya Krishnan2 | Kai M.A. Chan2 | Luke O.
Frishkoff5
1DepartmentofWildlife,Fish,andConservationBiology,UniversityofCalifornia,Davis,California2InstituteforResources,Environment,andSustainability,UniversityofBritishColombia,Vancouver,BC,Canada3UnióndeOrnitólogosdeCostaRica,NaranjodeAlajuela,CostaRica4HerbarioNacionaldeCostaRica,MuseoNacionaldeCostaRica,SanJosé,CostaRica5DepartmentofBiology,UniversityofTexasatArlington,Arlington,Texas
CorrespondenceDaniel S. KarpEmail:[email protected]
Funding
informationNationalGeographicSociety,Grant/AwardNumber:9977-16;BelmontForum,Grant/AwardNumber:G8PJ-437336-2012;CanadianNetworkforResearchandInnovationinMachiningTechnology,NaturalSciencesandEngineeringResearchCouncilofCanada,Grant/AwardNumber:06-5566;GraduateLeadershipFellowship;KillamDoctoralFellowship;UniversityofToronto;NSFGRF;CanadaResearchChairs,Grant/AwardNumber:A15-0109;AnimalCareCommittee;CostaRicangovernment,Grant/AwardNumber:SINAC-SE-CUS-PI-R-036-2016andSINAC-SE-CUS-PI-R-030-2017
HandlingEditor:RicardoSolar
Abstract1. Theoutcomeof theongoingbiodiversitycrisisdependson
thecapacityof theEarth’s wildlife to persist in working landscapes.
Yet, the species that
occupyworkinglandscapesareoftendistinctfromthoseinprotectedareas,withalargegroupof“sensitivespecies”thoughttorarelyventureintohuman-dominatedland-scapes.Asgovernmentshavecommittedtorestoringdegradedlandsworld-wide,determiningwhetherandhowworkinglandscapescanberestoredtobenefitsen-sitivespeciesremainsamajorchallenge.
2. We surveyed Neotropical birds across Northwestern Costa Rica
in
protectedareas,farmsandforestsembeddedwithinworkinglandscapes.Weanalysedcom-munitycompositiontounderstandhowgradientsofforestcover,fragmentationandregionalprecipitationdeterminehowconserving(orrestoring)tropicalforestsinworking
landscapes could safeguard entire communities, especially
sensitivespecieswithlimitedranges.
3. We foundagricultural sitesmaintained
relativelyhighbirddiversitybuthostedvery distinct communities from
those found in protected areas. The averagerange size of species
found in agricultural communitieswas double the size
ofspeciesinprotectedareas.However,highforestcoversitesinworkinglandscapeshousedbirdcommunitieswithsmallrangesizesthatwereequivalenttothoseinnearbyprotectedareas,despitebeingtwiceasfragmentedandsignificantlymoredisturbed.
4.
Theeffectoflocalforestcoveronbirdcompositionwascontingentonbothland-scapecontextandregionalclimate.Whenlocalforestcoverincreasedinwetterregionsandmoreforested
landscapes,birdcommunities inworking landscapesexhibited a stronger
shift towards the assemblages found in protected
areas.Specifically,wefoundthatreforestingthewettestsiteswouldincreasesimilaritytoprotectedareasfourfoldcomparedtoonlyatwofoldincreaseinthedriestsites.
5. Synthesis and applications. Despite experiencingmuchmore
fragmentation anddegradation than protected areas, forests in Costa
Rican working
landscapescanmaintainbirdcommunitiesthatstronglyresemblethosefoundinprotectedareas.Thissuggeststhatconservingorrestoringforests
inworkinglandscapes,
www.wileyonlinelibrary.com/journal/jpemailto:https://orcid.org/0000-0002-3832-4428https://orcid.org/0000-0001-5738-2140mailto:[email protected]
-
1840 | Journal of Applied Ecology KARP et Al.
1 | INTRODUC TION
Conservationbiologistsandpractitionersareincreasinglyrecogniz-ing
the value ofworking landscapes for safeguarding
biodiversity(Chazdonetal.,2009;Kremen&Merenlender,2018).Indeed,“work-inglandscapes,”orhuman-dominatedlandscomposedofpastures,multiplecropspeciesandpatchesof
forests,grasslandsandothernatural habitats, have been repeatedly
shown to sustain diversecommunities (Melo, Arroyo-Rodriguez,
Fahrig,Martinez-Ramos,
&Tabarelli,2013).Yet,thespeciesthatoccupyworkinglandscapesareoftendistinctfromthoseinprotectedareas,withthemostvulnera-blespeciesfailingtopersist(Karpetal.,2015;Newboldetal.,2016;Pfeiferetal.,2017).Evenminordisturbancesinotherwiseintactfor-estssometimesexactdeclinesinvulnerableforestspecies,necessi-tatingthecreationofprotectedareas(Barlowetal.,2016;Bettsetal.,2017).Thus,despiteencouraging
findings related to
themain-tenanceoflocaldiversity,humanmodificationofintactlandscapesisstillrestructuringbiologicalcommunities(Newboldetal.,2016).
Nonetheless,ongoingandprojectedtrends in land-useare
im-pedingeffortstosufficientlyexpandtheglobalreservenetworktoslowtheongoingbiodiversitycrisis(Pouzolsetal.,2014).Therefore,while
reservecreationmust remainacornerstoneofconservationpolicy, the
fateof Earth'swildlifewill at least partially dependonthe
hospitability of working landscapes (Chazdon et al.,
2009).Ecologistsandpractitionersareincreasinglycallingforconservationinitiativesthattargetworkinglandscapes(Kremen&Merenlender,2018),includinglandscape-scalerestorationprojects.Forexample,theworld'sgovernmentshavebeenchallengedtorestore150Mhaofdegradedlandworld-wide—anareathesizeofMongolia
(Menz,Dixon,&Hobbs,2013).
Akeyquestionfacingsuchinitiativesiswheretotargeteffortstomaximizethebenefitsforbothpeopleandnature(Menzetal.,2013).Atregionalscales,therelativerelianceofdifferentspeciesonintactforestmayshiftacrossclimategradients,whichcouldinfluenceres-torationplacementstrategies(Karpetal.,2018).Atlandscapescales,avarietyof
factorsareknown tomediate
thespeedandcapacityforrestorationprojectstorecruitviablepopulationsofnativespe-cies
(Reid,Mendenhall, Rosales, Zahawi,&Holl, 2014). For
exam-ple,the“intermediatelandscape-complexityhypothesis”positsthatconservation
interventions should be targeted in
human-modifiedlandscapeswithintermediateamountsofremainingnaturalhabitat(Tscharntkeetal.,2012).Thethinkingisthatincompletelyclearedlandscapes,
sourcepopulationsmaynotexist to sendcolonists
torestoredsites,and,inveryintactlandscapes,somuchhabitatexists
thatcolonistsmay“spillover”intoclearedareas,withorwithoutanyconservationinterventions.
The intermediate landscape-complexity hypothesis,
however,wasprimarilyconceivedwiththegoalofbolsteringgeneralist,eco-system-serviceprovidersthatrelyonforestsbutreadilymoveintoagriculture.Consequently,
restoringorconservinghabitat in inter-mediate landscapes may fail
to conserve vulnerable species thatrarelyutilizeagriculture
(Tscharntkeet al.,
2012).Moreover,manyforest-restrictedbirdsrefusecrossinganydeforestedgapstocolo-nizenewfragments(Ibarra-Macias,Robinson,&Gaines,2011).Thus,restorationmaybemostsuccessfulatbolsteringforest-associatedspecieswhen
sites are located in landscapeswith large blocks
ofcontiguousforest(Reidetal.,2014).
We surveyed bird communities in Costa Rica to evaluate
thepotential forworking-landscape conservation to bolster
forest-re-stricted birds. Specifically, we censused birds at 150
sites over2 years in 5 reserves and 20working landscapes, arrayed
acrossan independent precipitation gradient encompassingwet and
dryforests(~1.5–2.8m,annualrainfall).Reservesvariedinsize(range:9.1–183
km2,mean: 59 km2) and time since establishment
(range:1974–1994,mean:1984).Agriculturalsitesencompassedpastures,rice,
sugarcane,andTaiwangrass (a foragecrop).Forests
inwork-inglandscapesvariedintheamountofforestcoverwithin50m(i.e.local
scale; range: 44%–100%,mean: 85%) andwithin 610m
(i.e.landscapescale;range:16%–96%,mean:63%;seemethodsforscaledefinitions).Ourworkwasorganizedaroundthreequestions.First,towhatextentcanforestsandfarmsinworkinglandscapesmaintainlocal
bird richness relative to protected areas? Second, given
thatforestsinworkinglandscapesareoftenfragmentedanddegraded,how
distinct are bird communities in working landscapes
versusprotectedareas?Inparticular,canspeciesofconservationconcernpersist?Third,whereshouldforestconservationandrestorationbetargetedtofacilitatereserve-likebirdcommunities?
2 | MATERIAL S AND METHODS
2.1 | Bird surveys
Weselected25focallandscapesinNorthwestCostaRica:20inwork-ing
landscapesand5 inprotectedareas
(ReservaBiológicaLomasBarbudal,ReservaNaturalMonteAlto,andParquesNacionalesPaloVerde,BarraHonda,andDiriá).Protectedareasencompassedmostoftheprecipitationgradient(1.6–2.4mvs.1.5–2.8matothersites).Otherprotectedareas
intheregionwouldnothaveservedasfair
particularlywithinwetterregionsandalreadyforestedlandscapes,maysafeguardbirdcommunitieswhencreatingprotectedareasisinfeasible.
K E Y W O R D S
avian,deforestation,fragmentation,habitatloss,landscapecontext,reserve,restoration,workinglandscapes
-
| 1841Journal of Applied EcologyKARP et Al.
comparisonstoourworking-landscapesites,astheywereeithertoofarawayorencompassedhabitattypesthatwedidnotsurvey.
Ineach landscape,we identified sixbird survey sites (N = 150
total).Inworkinglandscapes,threesiteswerelocatedinagricultureandthreeinadjacent,privatelyownedforests.Siteswithinthesamelandscapewere
separatedby500monaverageandwerechosensothatlocalforestcovervariedindependentlyfromlandscape-levelforestcoverandconfiguration.That
is,wesystemicallyvariedsitelocationswithineachlandscape,placingsitesinforestinteriorsandforestedges,inareassurroundedbysubstantialforestcoverandinsmallfragmentssurroundedbyagriculture(FigureS1andTableS1inMethodsS1).Inprotectedareas,foursiteswerelocatedinforestinteriorsandtwoatthereserves’edges.
Oneexpertobserver(J.Zook)conducted20min,50mfixedra-diuspointcountsateachsite.Becauseourfocuswasontheresidentavifauna,
survey effort was concentrated on the Boreal
summer(May–August).Dryseasonsurveyswouldhaveresultedinfewres-identdetections,asmanyspecies
leavedryforestsduringthedryseason and vocalize less frequently.
Sites in working landscapeswere surveyed in 2016 and2017; protected
areaswere surveyedonly in 2017. Approximately half of the siteswere
sampled
threetimeseachyear(forbinomialmixturemodelling,seebelow)andtheotherswereonly
surveyedonce.Zook surveyedone
farmorpro-tectedareaperday(sixsites).Surveysbeganatsunriseandcontin-uedfor~5hr.Allbirdsseenorheardduringcountswererecorded,inadditiontothetimeofday,thepresenceofloudnoise(e.g.cicadas,streamsandfarmmachinery),numberofpeoplenearby,windspeed(usingananemometer)anddistancetonearestriver.
2.2 | Environmental gradients
We quantified the local vegetation structure, surrounding
forestcoverand regionalprecipitationassociatedwitheachsite. In
four,5-m-radiussubplotsateachsurveysite,wequantifiedthefollowingvegetationstructurevariables:canopy
cover,proportion of trees with epiphytes or vines,proportion of
trees with lianas,understorey density,herbaceous ground cover,shrub
cover,tree species richness,number of tree stems,mean tree DBH and
mean vegetation height(seeMethodsS1). To quantify surrounding
forest cover, we hand-classified
alltrees,includingplantations,within1.5kmofeachsurveysiteusingGoogle
Earth imagery from2013 to 2017 (Karp et al.,
2018).Wegroundtruthedourfinalmapusingthe600vegetationplotsdetailedabove.Wethendefinedlocal
forest
coverasthefractionoftreecoverwithin50mofeachsurveysite.Landscape
forest coverwascalcu-lated atmultiple spatial scales andwas defined
as the
proportionoftreecoverwithin“doughnuts,”whichalwayshadaninnerradiusof50mbuttheouterradiusvariedfrom60mto1.5kmby10-mincrements.Tocalculatelandscapeconfigurationatmultiplescales,we
first deleted all isolated tree clusters
75%forestcoveratlocalandlandscapescales),low‐cover forest (N = 45;
forested sites inworking landscapeswith95%oftheposteriors.
To compare vegetation structure and landscape attributes
be-tweenlanduses,weimplementedLinearMixedModels(LMMs;Bates,Maechler,Bolker,&Walker,2015)thatincludedcategoricallanduseasthesolefixedeffectandarandomeffectoflandscapetoaccountforspatialautocorrelation.Wetransformedresponsevariableswhennecessary
to satisfymodel assumptions (FigureS8
inMethodsS1).Variablesignificancewasassessedusing
likelihoodratiotests,com-paringnestedmodelswithandwithoutthecategoricalland-usefixedeffect(Zuur,Ieno,Walker,Saveliev,&Smith,2009).
2.4 | Modelling species richness, range size, and similarity to
reserves
We used the binomial mixture model to estimate species
abun-dancesacrosssites.Specifically,weextracted
themodelledabun-dance of each species at each site in 2017
(Ni,j,2017)—the year in
-
1842 | Journal of Applied Ecology KARP et Al.
whichallsitesweresampled—alongeachiteration(N=2,000)oftheposterior
(Karpetal.,2018).Usingeachof these2,000
“posteriorcommunities,”wecalculated thespecies
richnessofeachsite.Wealsocalculatedtheaveragerangesizeacrossallspeciespresentateachsite,usingestimatesfromBirdLifeInternational(2019).
Next, we quantified the bird community similarity
betweeneachpairofsitesusingpresence-absence(Sørensonsimilarity)andabundance
(Bray–Curtissimilarity)metrics.Wedecomposedthesemetrics to analyse
their turnover components, using “betapart” inr
(Baselga,Orme,Villegar,Bortoli,&Leprieur,2018;RCoreTeam,2018).
We visualized differences between sites in their
speciescompositionsusingnon-metricmultidimensionalscaling,andtestedwhetherdifferentlanduseshostedsignificantlydifferentcommuni-tiesviapermutationalmultivariateanalysisofvariance.Wealsousedthecommunitysimilaritymetricstocalculatetheoverlapbetweenthebirdcommunityfoundateachsiteandcommunityfoundinthenearestprotectedarea.
Specifically, for eachmetric
andposteriorcommunity,wecalculatedthemultivariatebirdcommunitydistancefromeachsitetothecentroidoftheclosestprotectedarea(Karpetal.,2018).Aswewereinterestedinthemost“intact”reservecom-munity,weexcludedthetwoedgesitesineachprotectedareawhencalculatingsimilaritytoreservecommunities.
Wemodelledspeciesrichness,averagerangesize,andreservesimilarityusingLMMswithlandscapeasarandomeffectandthefol-lowing
fixedeffects: local forest cover (linearandquadratic
terms),landscape forest cover and edge, precipitation, vegetation
structure, and interactions between local forest cover and
precipitation,landscape forest cover,andforest
edge.Forthespeciesrichnessandrangesizeanalyses, we included reserve
sites and added reserve status
(i.e.,whetherthesitewasinareserveornot)asanotherpredictor.Forthereservesimilarityanalysis,weomittedreservesitesbutincludeddistance
to nearest reserve as another fixed effect to account
forcommunitysimilaritydecayingwithdistance(Karpetal.,2018).Weomitteddistance
to
reservesinmodelsthatdidnotfocusoncommu-nitysimilarity.Inallmodels,vegetation
structure
wasmeasuredasthesecondandthirdprincipalcomponentsofaPCAonallvegetationstructure
variables. The first principal componentwas highly
cor-relatedwithlocal forest cover(Pearson'sr=0.88,df=148).
All fixed effects were standardized prior to analysis.
Modelswere weighted by the posterior variance of the species
richness,reserve size, and reserve similarity estimates (Karp et
al.,
2018).AllmodelsconformedtoLMMassumptions(i.e.normality,hetero-scedasticity)andnoneoftheincludedparametersdisplayedsevereevidenceofcollinearity(varianceinflationfactors
-
| 1843Journal of Applied EcologyKARP et Al.
While habitat fragmentation per se (measured as landscape-scale
forestedge
length)didnotaffectspeciesrichnessalone,wedidobserveaninteractionbetweenforest
edge and local forest cover
amount,suchthatthestrongestrichnessincreaseswithlocal forest
coverwereobservedintheleastfragmentedlandscapes(Figure1b).Noneofourresultschangedwhenimplementingforwardorback-wardsmodel
selection. Resultswere also largely
consistentwhenexaminingrawdetections(ratherthanmodelledabundances).Rawdetectionmodelsdid,however,suggestaninteractionbetweenpre-cipitationand
local forestcover, such that
richnessdisproportion-atelyincreasedinwetter,moreforestedsites.Restrictingourfocusto
forested sites and excluding agriculture, all models
supportedtrends of increasing richness at sites with more landscape
forestcoverandinwetterregions(TableS3andFigureS4inMethodsS1).Wefoundlessconsistentsupport,however,forrichnessincreasinginstandswithfewer,largertreesandatsiteswithmoreunderstoreydensityandherbaceousgroundcover.
3.2 | Are bird communities in working landscapes distinct from
those in reserves?
We found strong evidence that community composition
shiftedacross the environmental gradients (Table S4 in Methods
S1;Figure2andFigureS5inMethodsS1).However,communitieswere
muchmoreresponsivetolocalforestcoverand,toalesserextent,regionalprecipitationthanothervariables.Wealsofoundthatcom-munitycompositionsignificantlydifferedbetweenprotectedareas,high-coverforest(i.e.sitesinworkinglandscapeswith>75%forestcoveratlocalandlandscapescales),
low-coverforestsandagricul-ture.Onekeyexceptionwasprotectedareasandhigh-coverforests,whichcouldnotbedifferentiated
(p>0.05).
Indeed,birdcommu-nitiesrapidlyshiftedatsiteswith>75%localand
landscapeforestcover, with “reserve-affiliated species” increasing
in proportionalabundance(FigureS6inMethodsS1).
Average range sizes reflected these community shifts.
Wide-rangingbirds replacednarrow-rangedbirds at siteswith less
localforest cover, especially in less forested landscapes and in
wetterregions (Figure 1). As a result, average range sizes in
agriculturalcommunitiesweremorethantwiceaslargeascommunitiesinpro-tectedareas(FigureS3inMethodsS1).Rangesizetendedtodeclineinfragmentedlandscapes(Figure1),andinforestedsiteswithfewerlianasandmoreherbaceousvegetation
(FigureS7
inMethodsS1).Critically,averagerangesizedidnotdifferbetweenprotectedareasandhigh-coverforests(FigureS3inMethodsS1).
Indeed,wefoundthatveryfewbirdspeciesuniformlydeclinedin
abundance outside protected areas, after controlling for
differ-encesinforestcover.Specifically,only3ofthe150surveyedspeciesweresignificantlymoreabundant
inprotectedareas thanworking
F I G U R E 1
Speciesrichnessandaveragerangesizechangesalonggradients.Richnessexhibitedaslight,nonlinearincreasewithlocalforestcover,peakingat~80%(dottedline;a).Increasesinrichnesswithlocalforestcoverweremostpronouncedatsiteswithlessforestedgeinthelandscape(b).Richnessalsoincreasedwithprecipitation(c)andwiththesecondprincipalcomponentofavegetationstructurePCAthatdifferentiatedsiteswithmanythintreesfromsiteswithfewer,widertrees(d).Unlikerichness,averagerangesizeacrossbirdcommunitiesdeclinedsharplywithlocalforestcover(e).Declinesweremorerapidinsiteswithmorelandscapeforestcover(f)andinwetterregions(g).Rangesizealsodeclinedinmorefragmentedlandscapes(h).Linesindicatepredictedtrends;greyregionsdelineate95%confidenceintervals.Plussignsaresitesinprotectedareas(PA);greycirclesaresitesinworkinglandscapes
-
1844 | Journal of Applied Ecology KARP et Al.
landscapes(FigureS8inMethodsS1).Similarly,whenourbinomialmixture
model was modified to compare land-use categories,
wefoundthat7,32and50speciesweresignificantlymoreabundantinprotectedareasthanhigh-coverforests,low-coverforestsandagri-culturerespectively(AppendixS2andFigure3).
This near equivalency in bird community composition
betweenprotectedareasandhigh-coverforestsexisteddespitestructuraldif-ferences(FigureS9inMethodsS1).High-coverforestswereembeddedinlandscapesthatweremorethantwiceasfragmentedasprotectedareas.
Logging and other disturbances also contributed to distinct
vegetationstructures.Specifically,comparedtoprotectedareas,high-cover
forests exhibited 13% lower canopy cover, 30% shorter
treeheights,20%lowertreerichnessand35%lessdenseunderstories.
3.3 | How could working landscapes be managed to promote
reserve‐like bird communities?
Communitiesinworkinglandscapeswithmoreforestcoveratlocalandlandscapescalesexhibitedahighdegreeofoverlapwithcom-munitiesinnearbyreserves(Figure4;TableS7inMethodsS1).The
F I G U R E 2
Non-metricmultidimensionalscalingplotsdepictingshiftsinbirdcommunitycompositionalongenvironmentalgradients.Thedistancebetweensites(points)correspondstodifferencesincommunitycomposition(Bray–Curtissimilarity).Sitesthatsharedsimilarlevelsoflocalforestcover(a;greentoorangegradient)orprecipitation(b;purpletoredgradient)alsohostedsimilarbirdcommunities.(c)Birdcommunitiesinreserves(darkgreen;plussigns),high-coverforests(sitesinworkinglandscapeswith>75%forestcoveratlocalandlandscapescales;olive),low-coverforests(lightgreen)andagriculture(orange)weredistinct,exceptthatreserveandhigh-coverforestcommunitiesoverlapped.OvalsareordinationellipsesbasedontheSDofpointscores
F I G U R E 3
Differencesinbirdabundancebetweenreservesandotherlanduses.(a)Barplotsdepictcomparisonsofspeciesabundancesbetweenreservesandhigh-coverforest(sitesinworkinglandscapeswith>75%forestcoveratlocalandlandscapescales;leftbar),low-coverforests(
-
| 1845Journal of Applied EcologyKARP et Al.
effect of local forest cover was nonlinear, with sites
acceleratinginreservesimilarityasforestcoverincreased.Thiswasmostlytheresult
of accelerating abundance increases of “reserve-affiliatedspecies”
at high levels of forest cover; “agriculture-affiliated
spe-cies”exhibitedamore lineardeclinewith forestcover
(FigureS10inMethodsS1).Thus,themostforestedsitesinworkinglandscapesexhibited
roughly the same degree of reserve similarity as actualreserve
sites didwith sites in different reserves (Figure
4c,f).Wealsofoundthatincreasinglocalforestcoverinwetterregionsand/or
in more forested landscapes increased reserve similarity moreso
than anequivalent amountof forest increase indry
regionsordeforestedlandscapes.Fragmentationhadnodetectableeffectoncommunitysimilaritywithreserves.Theseresultswerequalitativelysimilarwhenanalysingrawdetectionsandtheturnovercomponentofdissimilarity(TablesS7andS8inMethodsS1).Ourfindingswerealsogenerallyconsistentwhenreservesimilaritywascalculatedwithmetrics
thatonlyconsideredspeciespresencesand
ignoredabun-dances(FigureS11andTableS7inMethodsS1).
Withinforestsites,findingswerelessconsistentacrosssimilaritymetrics,modelselectionproceduresandanalysistargets(i.e.mod-elledcommunitiesvs.rawdetections;TablesS9andS10andFigureS12
inMethodsS1). In every case,we found that reserve similar-ity
increasedatsiteswithmore landscape forestcover.Wefoundless
consistent support for increased similarity with greater
treeheights, canopy cover, understorey density, proportion of
lianas,
tree richness and shrub cover. Finally,we foundvery
inconsistentsupport that reserve similarity declined with
precipitation, forestedge,numberoftreestemsandDBH.
Ourmodelofabundance-basedreservesimilarityadequatelyfittheobserveddata(conditionalR2=0.90)andcouldthusbeusedtopredictspatialvariationintheprojectedoverlapbetweenbirdcom-munities
found inworking landscapes andnearby reserves acrossNorthwest Costa
Rica. Both across Northwest Costa Rica andwithin
theCorredorBiológicoHojancha-Nandayure (Figure5),
re-sultingmapshighlightedthevalueofmaintainingorrestoringforestsinwetterregionsandinregionswithmorelandscapeforestcover.Specifically,modelspredictedthatrestoringagriculturalpixelsinthewettestareasresultedintwicethegainsinreservesimilarityasre-storing
in thedriest areas. Similarly, restoring agriculture in100%forested
landscapeswas2.65timesmoreeffective in termsof
in-creasingreservesimilaritythanindeforestedlandscapes.
4 | DISCUSSION
Our study suggests that there is great scope for
conservingNeotropicalbirdsinCostaRicanworkinglandscapes.Speciesrich-ness
was no higher in reserves than in working landscapes
withsubstantialforestcoveratlocalandlandscapescales,andfewspe-cies(
-
1846 | Journal of Applied Ecology KARP et Al.
forests.Consequently,high-coverforestsandreserveswerelargelyindistinguishableinspeciescompositionandbothhostedcommuni-tiesofbirdswithsmallrangesizes.Thiswassurprisingashigh-coverfragmentswereembedded
inworking landscapes
thatweretwiceasfragmentedasprotectedareas.Moreover,regular
logging,fires,huntingandotherdisturbances
likelyalteredthevegetationstruc-tureofforestsinworkinglandscapescomparedtoprotectedareas.Indeed,high-coverforestsinworkinglandscapeshadlowercanopycover,shorter
treeheights, fewertreespecies,and
lessdenseun-derstoriesthanprotectedareas.Landownerslikelymaintainforestformultiplereasonsincludinglawsthatmandatereforestationnearrivers(Ley7575,276),CostaRica'sflagshipPaymentforEcosystemServicesprogramthatincentivizesforestconservation,andbecausesomeareasmaybemarginalareasforcultivation(e.g.steeperslopes;FigureS13inMethodsS1).Regardless,thesefindingssuggestthatmaintainingforestinCostaRicanworkinglandscapescanpromoteavian
communities that resemble those in protected areas,
evenwhensubjecttodisturbances.
Critically,we found thatmany species alsopersisted in
totallydeforested sites. Specifically, we detected a nonlinear,
saturatingrelationship between species richness and local tree
cover,
suchthatrichnessonlydeclinedby20%(sevenspecies)atthemostde-forestedsites.This
finding isnotunique toour region:agriculture
hasbeenshowntomaintainspeciosecommunitiesinotherareasofCostaRica(Karpetal.,2015)andabroad(e.g.Ranganathan,Daniels,Chandran,
Ehrlich,&Daily, 2008;Waltert et al., 2005).
Especiallybecausestudiesoftropicalwildlifeoftenfocusonforest-restrictedspecies(e.g.Barlowetal.,2016),thisdiversecommunityofagricul-turalspeciesbearsfurtherstudy.
Still, it is importanttoconsiderthe
identityofthespeciesthatpersisted in farmland. Only two surveyed
species are not listedas “least concern” by the IUCN—Great Curassow
(Crax rubra) andYellow-napedParrot(Amazona
auropalliata)—bothwhichweremostregularlydetected in forests
inworking landscapes (butnot inag-riculture).More
importantly,wefoundthat
theaveragerangesizeofspeciesinagriculturalsiteswasmorethandoublethatofspeciesinprotectedareas,indicatingthateitheragriculturefacilitatesnatu-rallywide-rangingoverendemicspeciesorthatagriculturefavoursspecies
thatwere initially range-restrictedbut
laterbecamewide-rangingasagricultureexpanded.
Adivergenceinaveragerangesizereflectsourbroaderobserva-tion that
aviancommunity composition stronglydifferedbetweenagriculture and
protected areas. Other studies across the tropicshave documented
strong shifts in community composition withhabitat conversion
(Gibson et al., 2011; Newbold et al., 2016). InCameroon,
similarbird richnessbetween forests and farmsbelied
F I G U R E 5
Mapsofstudysitesandsimilarityofbirdcommunitiesinworkinglandscapestoreservecommunities.(a)NorthwestCostaRicaencompassesastrongprecipitationgradient.(b)Mapdepictsprotectedareas(darkgreen),forestsinworkinglandscapes(blue)andagriculture(orange)acrosstheregion.In(a)and(b),whitedotsindicatereservesitesandreddotsaresitesinworkinglandscapes.(c)Wettersitesandsiteswithmoreforestcoveratlocalandlandscapescaleswerepredictedtohostmorereserve-likebirdcommunities.Greenpolygonsdenoteterrestrialreserves(UNEP-WCMC&IUCN,2018).YellowpolygondenotestheCorredorBiológicoHojancha-Nandayure,aregionbeingprioritizedforreforestation.(d)Simulatedreforestationofthecorridor'sagriculturalsitesfrom0%to100%localforestcoversuggestsgreaterincreasesinbirdcommunitysimilaritytoreservesforsiteslocatedinwetterregions(i.e.theSouthwest)andsurroundedbymoreforest(forestinworkinglandscapes=lightgreen;reserves=darkgreen;asopposedtootherland=grey)
-
| 1847Journal of Applied EcologyKARP et Al.
markeddifferencesincommunitycomposition(Waltertetal.,2005).Similarly,
inBrazil'sAtlanticforest,vertebratecommunitiesexperi-encedanabruptshiftfromforestspecialiststodisturbance-adaptedspecies
when landscape-level forest cover declined below 30%(Banks-Leite et
al., 2014). Here, community similarity to
reservesalsorespondedstronglytoforestcover,withreserve-affiliatedspe-ciesrapidlyincreasinginabundanceatsiteswithmorethan75%for-estcoveratlocalandlandscapesites.
Insomesenses,ourfindingthattheeffectoflocalforestcoveramplified
in forested landscapes contradicts the
intermediatelandscape-complexity hypothesis, which predicts
stronger effectsof local management in landscapes of intermediate
forest cover(Tscharntke et al., 22012). However, this hypothesis
originally
fo-cusedongeneralist,ecosystem-serviceproviders,ratherthanmoresensitiveforestspecies.Ourresultssuggestthatmaintainingorre-storinglargerswathsoftropicalforestmaybeessentialifthegoalistopreservemorereserve-likebiologicalcommunities,completewithrange-restrictedspeciesorothersofconservationconcern(Bettsetal.,2017;Pfeiferetal.,2017;Reidetal.,2014).
Thatsaid,forestedlandscapesmaymaintainseeddispersersthatcould
facilitate restoration without active management
(Hooper,Legendre,&Condit, 2005),meaning intermediately forested
land-scapes should still be targeted for active restoration
(Tambosi,Martensen,Ribeiro,&Metzger,2014).Moreover,recentapproachesareabletoprioritizesitesforrestorationbasedoncostsandmulti-plebenefits(i.e.notjustreservesimilarity;Strassburgetal.,2019).Forexample,ifanothergoalistoenhanceecosystemservices,thenrestoringforestindeforestedlandscapeswouldincreasefarmland–forestinterfacesandallowhabitatgeneraliststomovefromforeststofarms(Karpetal.,2015;Ricketts&Lonsdorf,2013).Similarly,agri-culturemaybecompatiblewithwaterbirdconservation:inourstudysystem,80%ofdetectedwaterbirdspecies(e.g.herons,ducksetc.)weremoreabundantinagriculturethanprotectedareas.
Itisalsoimportanttoacknowledgetheregionalcontextbeforeadvocatingthebenefitsofforestconservationinworkinglandscapeselsewhere.First,itispossiblethatoursurveysmissedespeciallyrareforest
specialists that are restricted to protected
areas.However,only21additional (non-migratory) speciesweredetected
in
a15-yearbirdcensuseffortat16sitesinourstudyregion(Frishkoffetal.,2014),allofwhicharelistedas“leastconcern.”Apossibilitythatwecannotexclude,however,isthatthemostsensitivespecieshaveal-readybeenextirpatedfromtheregion.Ourstudyregionexperiencedhigh
rates of deforestation until the 1980s, atwhich point
forestcoverbegan
increasing,from23%in1986to47%in2005(Calvo-Alvarado,McLennan,Sanchez-Azofeifa,&Garvin,2009).Thismeansthatsensitivespeciescouldhaveregionallyextirpatedpriortothe1980s.Second,formalprotectedareasareoftensubjecttodegrada-tion,callingintoquestiontheirutilityasabaseline.Indeed,one-thirdofprotectedareasfaceseverehumanpressureandCostaRicaisnoexception(Jonesetal.,2018).Third,ourpriorworksuggeststhatdryforest-associated
speciesmay be preadapted to thrive
inworkinglandscapes(Frishkoffetal.,2016;Karpetal.,2018).Highsimilaritybetweencommunitiesinreservesandforestedworkinglandscapes
may thus partially reflect the fact that someof our reserves
pro-tecteddryforests.Indeed,inwetterAmazoniansites,communitiesturnoverbetweenprimaryforestsandthesecondaryforeststhattypifyworkinglandscapes(Barlow,Mestre,Gardner,&Peres,2007).Finally,othertaxamayresponddifferentlythanbirds;forexample,arecentstudyfounditmaytakecenturiesforplantcommunitiesinsecondaryforeststoresemblethoseinprimaryhabitats(Rozendaaletal.,2019).Thus,prioritizingconservationoffragmented,disturbedforests
inworking landscapesmay be still be inadvisable in
areasthathostmoreendangeredspecies,are lessdegradedand/or
thatarelocatedinwetterregions(Barlowetal.,2016).
5 | CONCLUSIONS
Ourworkyielded several conservation-relevant insights.
First,wereportthatwhilebiodiversecommunitiespersistedinagriculture,as-semblageslackedtherange-restrictedspeciesofconservationcon-cernfoundinforests.Second,wefoundthatonly5%ofthespeciessurveyedweremore
abundant in reserves than in high-cover
for-ests,suggestingvulnerablespeciescouldgreatlybenefitfromcon-servingorrestoringforestinCostaRicanworkinglandscapes,eveniftheforestishighlyfragmentedanddisturbed.Third,ourmappingexerciseillustrateshowourmethodcanbeusedtoprovideconcreteguidanceforsitingconservationinitiatives(e.g.ongoingrestorationeffortsintheCorredorBiológicoHojancha-Nandayure;Figure5d).Indeed,NorthwestCostaRicaisexpectedtoexperiencefuturecli-matedrying
(Rauscher,Giorgi,Diffenbaugh,&Seth,2008)andwehavepreviouslyshownthatbirdsassociatedwithwetterandmoreforestedsitesarethemostvulnerabletothesechanges(Karpetal.,2018).Thus,ourworksuggeststhattargetingfuturerestorationandconservationinwetterregionsandmoreforestedlandscapescouldhelp
optimize biodiversity conservation, at least for
forest-associ-atedbirdsinCostaRica.
ACKNOWLEDG EMENTS
We thank X. Campos, E. Rodríguez, D. Steyn, L. Bogantes,
theFurturAgua team, SINAC staff and landowners for supporting
ourfieldwork and analyses. Our work was supported by a
GraduateLeadership Fellowship and a Killam Doctoral Fellowship to
A.E.,
aUniversityofTorontopostdoctoralfellowshiptoL.O.F.,anNSFGRFtoA.K.andtheCanadaResearchChairsProgram(K.M.A.C.).Additionalfunding
included grants from National Geographic (#9977-16),
theBelmontForum(#G8PJ-437336-2012)andNSERC(UBC#06-5566).This
research was conducted with approval from the Animal CareCommittee
(A15-0109) and Costa Rican government
(permits:SINAC-SE-CUS-PI-R-036-2016andSINAC-SE-CUS-PI-R-030-2017).
AUTHORS’ CONTRIBUTIONS
D.S.K.,A.E.,K.M.A.C.andL.O.F.designedresearch;D.S.K.,A.E.,J.Z.,P.J.,
andA.K. collecteddata;D.S.K., L.O.F.,A.E. and J.K. analysed
-
1848 | Journal of Applied Ecology KARP et Al.
data;D.S.K.wrotethemanuscript.Allauthorscontributedtodraftsandgavefinalapprovalforpublication.
DATA ACCE SSIBILIT Y
Data available via the Dryad Digital Repository
https://doi.org/10.5061/dryad.fg8kq66(Frishkoff&Karp,2019).
ORCID
Daniel S. Karp https://orcid.org/0000-0002-3832-4428
Luke O. Frishkoff https://orcid.org/0000-0001-5738-2140
R E FE R E N C E S
Banks-Leite,C.,Pardini,R.,Tambosi,L.R.,Pearse,W.D.,Bueno,A.A.,Bruscagin,R.T.,…Metzger,J.P.(2014).Usingecologicalthresholdsto
evaluate the costs and benefits of set-asides in a
biodiversityhotspot. Science, 345, 1041–1045.
https://doi.org/10.1126/science.1255768
Barlow,J.,Lennox,G.D.,Ferreira,J.,Berenguer,E.,Lees,A.C.,Nally,R.M.,…Gardner,T.A.
(2016).Anthropogenicdisturbance in
tropicalforestscandoublebiodiversitylossfromdeforestation.Nature,535,144–147.https://doi.org/10.1038/nature18326
Barlow,J.,Mestre,L.A.M.,Gardner,T.A.,&Peres,C.A.(2007).Thevalueof
primary, secondary and plantation forests forAmazonian
birds.Biological Conservation, 136, 212–231.
https://doi.org/10.1016/j.biocon.2006.11.021
Baselga,A.,Orme,D.L.,Villegar,S.,DeBortoli,J.,&Leprieur,F.(2018).betapart:
Partitioning beta diversity into turnover and
nestednesscomponents.
Bates, D.,Maechler,M., Bolker, B., &Walker, S. (2015).
Fitting linearmixed-effectsmodelsusing lme4.Journal of Statistical
Software,67,1–48.
Betts,M.G.,Wolf,C., Ripple,W. J.,
Phalan,B.,Millers,K.A.,Duarte,A., … Levi, T. (2017). Global forest
loss disproportionately
erodesbiodiversityinintactlandscapes.Nature,547,441–444.https://doi.org/10.1038/nature23285
BirdLife International (2019). IUCN Red List for birds.
Retrieved fromhttp://www.birdlife.org
Calvo-Alvarado, J., McLennan, B., Sanchez-Azofeifa, A., &
Garvin,T. (2009). Deforestation and forest restoration in
Guanacaste,Costa Rica: Putting conservation policies in context.
Forest Ecology and Management,258, 931–940.
https://doi.org/10.1016/j.foreco.2008.10.035
Chazdon, R. L.,Harvey,C.A., Komar,O.,Griffith,D.M., Ferguson,
B.G., Martínez-Ramos, M., … Philpott, S. M. (2009). Beyond
re-serves: A research agenda for conserving biodiversity in
human-modified tropical landscapes. Biotropica, 41, 142–153.
https://doi.org/10.1111/j.1744-7429.2008.00471.x
Frishkoff, L. O., & Karp, D. S. (2019). Data from:
Species-specific
re-sponsestohabitatconversionacrossscalessynergisticallyrestruc-tureNeotropicalbirdcommunities.Dryad
Digital Repository,https://doi.org/10.5061/dryad.fg8kq66
Frishkoff,L.O.,Karp,D.S.,Flanders,J.R.,Zook,J.,Hadly,E.A.,Daily,G.C.,&M’Gonigle,L.K.(2016).Climatechangeandhabitatconversionfavourthesamespecies.Ecology
Letters,19,1081–1090.https://doi.org/10.1111/ele.12645
Frishkoff,L.O.,Karp,D.S.,M'Gonigle,L.K.,Mendenhall,C.D.,Zook,J.,Kremen,C.,…Daily,G.C.(2014).Lossofavianphylogeneticdiversity
inneotropicalagriculturalsystems.Science,345,1343–1346.https://doi.org/10.1126/science.1254610
Frishkoff,L.O.,Mahler,D.L.,&Fortin,M.-J.(2017).Integratingoverun-certaintyinspatialscaleofresponsewithinmultispeciesoccupancymodels
yieldsmore accurate assessments of community
composi-tion.BioRxiv.https://doi.org/10.1101/143669
Gibson,L.,Lee,T.M.,Koh,L.P.,Brook,B.W.,Gardner,T.A.,Barlow,J.,…Sodhi,N.S.(2011).Primaryforestsareirreplaceableforsustainingtropicalbiodiversity.Nature,478,378–381.https://doi.org/10.1038/nature10425
Hooper, E., Legendre, P., & Condit, R. (2005). Barriers to
for-est regeneration of deforested and abandoned land inPanama.
Journal of Applied Ecology, 42, 1165–1174.
https://doi.org/10.1111/j.1365-2664.2005.01106.x
Ibarra-Macias,A.,Robinson,W.D.,&Gaines,M.S.(2011).ExperimentalevaluationofbirdmovementsinafragmentedNeotropicallandscape.Biological
Conservation, 144, 703–712.
https://doi.org/10.1016/j.biocon.2010.08.006
Jones,K.R.,Venter,O.,Fuller,R.A.,Allan,J.R.,Maxwell,S.L.,Negret,P.J.,&Watson,J.E.M.(2018).One-thirdofglobalprotectedlandisunder
intense human pressure.Science,360, 788–791.
https://doi.org/10.1126/science.aap9565
Karp,D. S., Frishkoff, L.O., Echeverri,A., Zook, J.,
Juárez,P.,&Chan,K. M. A. (2018). Agriculture erases
climate-driven β-diversity inNeotropicalbirdcommunities.Global
Change Biology,24,338–349.https://doi.org/10.1111/gcb.13821
Karp,D.S.,Mendenhall,C.D.,Callaway,E.,Frishkoff,L.O.,Kareiva,P.M.,
Ehrlich, P. R., &Daily,G. C. (2015). Confronting and
resolvingcompeting values behind conservation objectives.
Proceedings of the National Academy of Sciences of the United
States of
America,112,11132–11137.https://doi.org/10.1073/pnas.1504788112
Kremen, C., & Merenlender, A. M. (2018). Landscapes that
work
forbiodiversityandpeople.Science,362,1–9.https://doi.org/10.1126/science.aau6020
Melo, F. P. L.,Arroyo-Rodriguez,V., Fahrig,
L.,Martinez-Ramos,M.,&Tabarelli,M. (2013).On the hope for
biodiversity-friendly tropicallandscapes.Trends in Ecology &
Evolution,28,462–468.https://doi.org/10.1016/j.tree.2013.01.001
Menz,M.H.M.,Dixon,K.W.,&Hobbs,R.J.(2013).Hurdlesandoppor-tunitiesforlandscape-scalerestoration.Science,339,526–527.https://doi.org/10.1126/science.1228334
Newbold,T.,Hudson,L.N.,Hill,S.L.L.,Contu,S.,Gray,C.L.,Scharlemann,J.P.W.,…Purvis,A.(2016).Globalpatternsofterrestrialassemblageturnover
within and among land uses. Ecography, 39,
1151–1163.https://doi.org/10.1111/ecog.01932
Pfeifer, M., Lefebvre, V., Peres, C. A., Banks-Leite, C., Wearn,
O. R.,Marsh,C.J.,…Ewers,R.M. (2017).Creationofforestedgeshasaglobal
impacton
forestvertebrates.Nature,551,187–191.https://doi.org/10.1038/nature24457
Pouzols, M. F., Toivonen, T., DiMinin, E., Kukkala, A. S.,
Kullberg,
P.,Kuusterä,J.,…Moilanen,A.(2014).Globalprotectedareaexpansioniscompromisedbyprojectedland-useandparochialism.Nature,516,383–386.https://doi.org/10.1038/nature14032
RCore Team (2018).R: A language and environment for statistical
com‐puting. Vienna, Austria: R Foundation for Statistical
Computing.Retrievedfromhttps://www.r-project.org/
Ranganathan, J., Daniels, R., Chandran, S., Ehrlich, P. R.,
& Daily, G.C. (2008). Sustaining biodiversity in ancient
tropical country-side. Proceedings of the National Academy of
Sciences of the United States of America, 105(46), 17852–17854.
https://doi.org/10.1073/pnas.0808874105
Rauscher,S.A.,Giorgi,Æ.F.,Diffenbaugh,N.S.,&Seth,Æ.A.
(2008).Extension and intensification of the Meso-American
mid-summerdroughtinthetwenty-firstcentury.Climate
Dynamics,31,551–571.https://doi.org/10.1007/s00382-007-0359-1
https://doi.org/10.5061/dryad.fg8kq66https://doi.org/10.5061/dryad.fg8kq66https://orcid.org/0000-0002-3832-4428https://orcid.org/0000-0002-3832-4428https://orcid.org/0000-0001-5738-2140https://orcid.org/0000-0001-5738-2140https://doi.org/10.1126/science.1255768https://doi.org/10.1126/science.1255768https://doi.org/10.1038/nature18326https://doi.org/10.1016/j.biocon.2006.11.021https://doi.org/10.1016/j.biocon.2006.11.021https://doi.org/10.1038/nature23285https://doi.org/10.1038/nature23285http://www.birdlife.orghttps://doi.org/10.1016/j.foreco.2008.10.035https://doi.org/10.1016/j.foreco.2008.10.035https://doi.org/10.1111/j.1744-7429.2008.00471.xhttps://doi.org/10.1111/j.1744-7429.2008.00471.xhttps://doi.org/10.5061/dryad.fg8kq66https://doi.org/10.5061/dryad.fg8kq66https://doi.org/10.1111/ele.12645https://doi.org/10.1111/ele.12645https://doi.org/10.1126/science.1254610https://doi.org/10.1126/science.1254610https://doi.org/10.1101/143669https://doi.org/10.1038/nature10425https://doi.org/10.1038/nature10425https://doi.org/10.1111/j.1365-2664.2005.01106.xhttps://doi.org/10.1111/j.1365-2664.2005.01106.xhttps://doi.org/10.1016/j.biocon.2010.08.006https://doi.org/10.1016/j.biocon.2010.08.006https://doi.org/10.1126/science.aap9565https://doi.org/10.1126/science.aap9565https://doi.org/10.1111/gcb.13821https://doi.org/10.1073/pnas.1504788112https://doi.org/10.1126/science.aau6020https://doi.org/10.1126/science.aau6020https://doi.org/10.1016/j.tree.2013.01.001https://doi.org/10.1016/j.tree.2013.01.001https://doi.org/10.1126/science.1228334https://doi.org/10.1126/science.1228334https://doi.org/10.1111/ecog.01932https://doi.org/10.1038/nature24457https://doi.org/10.1038/nature24457https://doi.org/10.1038/nature14032https://www.r-project.org/https://doi.org/10.1073/pnas.0808874105https://doi.org/10.1073/pnas.0808874105https://doi.org/10.1007/s00382-007-0359-1
-
| 1849Journal of Applied EcologyKARP et Al.
Reid,J.L.,Mendenhall,C.D.,Rosales,J.A.,Zahawi,R.A.,&Holl,K.D.(2014).Landscapecontextmediatesavianhabitatchoiceintropicalforest
restoration. PLoS ONE, 9, e90573.
https://doi.org/10.1371/journal.pone.0090573
Ricketts,T.H.,&Lonsdorf,E. (2013).Mapping
themargin:Comparingmarginal values of tropical forest remnants for
pollination ser-vices. Ecological Applications, 23, 1113–1123.
https://doi.org/10.1890/12-1600.1
Royle,J.,&Dorazio,R.(2008).Hierarchical modeling and
inference in ecol‐ogy.London,UK:AcademicPress.
Rozendaal, D. M. A., Bongers, F., Aide, T. M., Alvarez-Dávila,
E.,Ascarrunz,N.,Balvanera,P.,…Poorter,L.(2019).Biodiversityrecov-eryofNeotropicalsecondaryforests.Science
Advances,5,eaau3114.https://doi.org/10.1126/sciadv.aau3114
Strassburg,B.B.N.,Beyer,H. L.,Crouzeilles,R.,
Iribarrem,A.,Barros,F.,deSiqueira,M.F.,…Uriarte,M.
(2019).Strategicapproachestorestoringecosystemscantripleconservationgainsandhalvecosts.Nature
Ecology and Evolution, 3, 62–70.
https://doi.org/10.1038/s41559-018-0743-8
Tambosi,L.R.,Martensen,A.C.,Ribeiro,M.C.,&Metzger,J.P.(2014).A
framework tooptimizebiodiversity
restorationeffortsbasedonhabitatamountandlandscapeconnectivity.Restoration
Ecology,22,169–177.https://doi.org/10.1111/rec.12049
Tscharntke, T., Tylianakis, J.M., Rand, T. A., Didham, R. K.,
Fahrig,
L.,Batáry,P.,…Westphal,C.(2012).Landscapemoderationofbiodiver-sitypatternsandprocesses-eighthypotheses.Biological
Reviews,87,661–685.https://doi.org/10.1111/j.1469-185X.2011.00216.x
UNEP-WCMC and IUCN. (2018). Protected planet: The world database
on protected areas
(WDPA).Cambridge,UK:Author:Retrievedfromwww.protectedplanet.net
Waltert,M.,Bobo,K.S.,Sainge,N.M.,Fermon,H.,Muhlenberg,M.,&Mulenberg,M.
(2005).Fromforesttofarmland:HabitateffectsonAfrotropical forest
birddiversity.Ecological Applications,15,
1351–1366.https://doi.org/10.1890/04-1002
Zuur,A.F.,Ieno,E.N.,Walker,N.J.,Saveliev,A.A.,&Smith,G.M.(2009).Mixed
effects models and extensions in ecology with
R.NewYork,NY:Springer.
SUPPORTING INFORMATION
Additional supporting information may be found online in
theSupportingInformationsectionattheendofthearticle.
How to cite this
article:KarpDS,EcheverriA,ZookJ,etal.RemnantforestinCostaRicanworkinglandscapesfostersbirdcommunitiesthatareindistinguishablefromprotectedareas.J
Appl Ecol. 2019;56:1839–1849.
https://doi.org/10.1111/1365-2664.13419
https://doi.org/10.1371/journal.pone.0090573https://doi.org/10.1371/journal.pone.0090573https://doi.org/10.1890/12-1600.1https://doi.org/10.1890/12-1600.1https://doi.org/10.1126/sciadv.aau3114https://doi.org/10.1038/s41559-018-0743-8https://doi.org/10.1038/s41559-018-0743-8https://doi.org/10.1111/rec.12049https://doi.org/10.1111/j.1469-185X.2011.00216.xhttp://www.protectedplanet.nethttps://doi.org/10.1890/04-1002https://doi.org/10.1111/1365-2664.13419https://doi.org/10.1111/1365-2664.13419