reconstructionfiretrap

NewlydiscoveredlandscapetrapsproduceregimeshiftsinwetforestsDavidB.
Lindenmayera,1,RichardJ.
Hobbsb,GeneE.
Likensa,c,1,CharlesJ.
Krebsd,andSamuelC.
BanksaaFennerSchoolofEnvironmentandSociety,AustralianNationalUniversity,CanberraACT0200,Australia;bSchoolofPlantBiology,UniversityofWesternAustralia,Crawley,WesternAustralia6009,Australia;cCaryInstituteofEcosystemStudies,Millbrook,NY,12545;anddDepartmentofZoology,UniversityofBritishColumbia,Vancouver,BC,CanadaV6T1Z4ContributedbyGeneE.
Likens,August4,2011(sentforreviewMay28,2011)Wedescribethe"landscapetrap"concept,wherebyentireland-scapesareshiftedinto,andthenmaintained(trapped)in,ahighlycompromisedstructuralandfunctionalstateastheresultofmul-tipletemporalandspatialfeedbacksbetweenhumanandnaturaldisturbanceregimes.
Thelandscapetrapconceptbuildsonideaslikestablealternativestatesandotherrelevantconcepts,butitsubstantivelyexpandstheconceptualthinkinginanumberofuniqueways.
Inthispaper,we(i)reviewtheliteraturetodeveloptheconceptoflandscapetraps,includingtheirgeneralfeatures;(ii)provideacasestudyasanexampleofalandscapetrapfromthemountainash(Eucalyptusregnans)forestsofsoutheasternAustralia;(iii)suggesthowlandscapetrapscanbedetectedbeforetheyareirrevocablyestablished;and(iv)presentevidenceofthegeneralityoflandscapetrapsindifferentecosystemsworldwide.
alteredecosystemprocesses|oldgrowthInmanyenvironmentsworldwide,keydriversofecosystemchangeinteractandreinforceoneanothertotriggercascadesofecosystemmodicationthataredifcultorimpossibletoreverse(1–3).
Thesecascadesareoftenreferredtoasregimeshifts(4–6).
Examplesofsignicantregimeshiftsincludeovershingandtrophiccascadesinmarinepredator–preysystems(7)andhumandisturbance-drivenlossesofdetritivorepopulationsandsubse-quentchangesinthedecompositionoforganicmatter(8).
Regimeshiftsarealmostalwaysidentiedinretrospect,makingitdifculttoknowhowtoavoidtheminadvanceandproblematictoreversetheireffects.
Therefore,understandingofthemechanisticpro-cessesbywhichregimeshiftsoccurmayprovideopportunitiestochangeresourcemanagementandavoidirreversibleandun-desirableecologicalchanges.
Inthispaper,wedescribethe"landscapetrap"concept,ofwhichtheoutcomeisaregimeshifttriggeredbyaseriesoffeed-backprocessesresultingfrominteractingnaturalandanthropo-genicdisturbances.
Wedenealandscapetrapasthatwhereinentirelandscapesareshiftedintoastateinwhichmajorfunctionalandecologicalattributesarecompromised.
Theseshiftsinalandscapeleadtofeedbackprocessesthateithermaintainanecosysteminacompromisedstateorpushitintoafurtherregimeshiftinwhichanentirelynewtypeofvegetationcoverdevelops.
Landscapetrapsarelarge-scaleecologicalphenomenathatarisethroughacombinationofalteredspatialcharacteristicsofalandscapecoupledwithsynergisticinteractionsamongmultiplehumanandnaturaldisturbances.
Thus,changesinthefrequencyandspatialcontagionoflarge-scaledisturbancesarethekeyinteractingfactorsdrivingentirelandscapesintoanundesirableandpotentiallyirreversiblestate(i.
e.
,landscapetrap).
Wedem-onstratetheconceptwithexamplesinvolvingspatialandtemporalfeedbackbetweenloggingandreinforestecosystemsandalsoprovideexamplesoflandscapetrapsinotherenvironments.
Likeotherkindsofecologicaltraps,thelandscapetrapconceptsharescharacteristicslikeshiftsbetweenalternativestablestatesandmultiplefeedbackprocesses(9).
However,afocusataland-scapescaleandontemporalandspatialchangesindisturbancessetsthelandscapetrapconceptapartfromotherkindsofecolog-icaltrapsandregimeshifts,suchaspopulationtrapsandextinc-tionvorticesinsmallpopulationsofanimals(10)andelevatedratesofanimalspecieslossbelowthresholdlevelsofnativeveg-etationcover(11).
Tothebestofourcollectiveknowledge,thelandscapetrapconcepthasnotbeenpreviouslyreported,yetwearguethatlandscapetrapsmaybemoreprevalentinecosystemsaroundtheworldthancurrentlyrecognized.
Commoningredientscon-tributingtolandscapetrapsare(i)overharvestingofnaturalresourcesinalandscape;(ii)climatechangeeffectsonspecies'lifehistoriesand/orthefrequencyandseverityofecologicaldisturbances;(iii)majorchangesinthespatialcharacteristicsoflandscapes;(iv)feedbackloopsbetweenthechangedenviron-mentalconditionsandothermajorstressors;and(v)severelyimpairedecologicalfunctionsofalandscapeinanalteredstate,suchas,forexample,reducedpopulationsofspeciesandhabitatsuitability,reducedcarbonstorage,andreducedwaterandtim-berproduction.
TheinteractionofthesefactorsisshowninaconceptualmodelinFig.
1.
Wesuggestthatlandscapetrapsexistinmanyecosystems.
Forexample,loggedtropicalrainforestsinpartsofAsiahavebecomemorere-prone(12).
Postresalvagelogginginsomeoftheserainforests,inturn,changesthevegetationcompositiontowardmorere-pronegrasslandtaxa.
Additionalrefurtherdegradesre-sensitiveremnantrainforest,eventuallyleadingtoaregimeshifttoexoticre-promotinggrasslands,limitingopportunitiesforthevegetationtoreverttotropicalrainforest(13).
Suchkindsofinterrelationshipsbetweenloggingandalteredreregimesarewidespreadintropicalrainforestsinmanyotherpartsoftheworld,includingSouthAmericaandAfrica(14),asarerela-tionshipsbetweenloggingandexoticre-pronegrasses(15).
Temperateforestsarenotimmunetosuchtraps.
InmoisttemperateforestsofwesternNorthAmerica,logging-relatedalterationsinstandstructureincreasetheriskforbothoccur-renceandseverityofsubsequentwildresthroughchangesinfueltypesandconditions(16,17).
High-severitywildreskillyoungtreesplantedfollowingpreviousloggingoperations.
Thisnecessitatesreforestationefforts,buttheseyoungstandsaresusceptibletobeingkilledinsubsequentrecurringhigh-severityres(16).
Similarkindsofrelationshipsbetweenloggingregimesandalteredreregimeshavebeenreportedinarangeofforesttypeselsewherearoundtheworld(reviewedin18).
Authorcontributions:D.
B.
L.
,R.
J.
H.
,G.
E.
L.
,C.
J.
K.
,andS.
C.
B.
designedresearch;D.
B.
L.
,R.
J.
H.
,G.
E.
L.
,C.
J.
K.
,andS.
C.
B.
performedresearch;D.
B.
L.
,R.
J.
H.
,G.
E.
L.
,C.
J.
K.
,andS.
C.
B.
analyzeddata;andD.
B.
L.
,R.
J.
H.
,G.
E.
L.
,C.
J.
K.
,andS.
C.
B.
wrotethepaper.
Theauthorsdeclarenoconictofinterest.
FreelyavailableonlinethroughthePNASopenaccessoption.
1Towhomcorrespondencemaybeaddressed.
E-mail:david.
lindenmayer@anu.
edu.
auorlikensg@ecostudies.
org.
Thisarticlecontainssupportinginformationonlineatwww.
pnas.
org/lookup/suppl/doi:10.
1073/pnas.
1110245108/-/DCSupplemental.
www.
pnas.
org/cgi/doi/10.
1073/pnas.
1110245108PNAS|September20,2011|vol.
108|no.
38|15887e15891ECOLOGYDownloadedbyguestonMarch14,2021ResultsandDiscussionSpecicExampleofaLandscapeTrap:MountainAshForestsofVictoria,SoutheasternAustralia.
Thespecicexampleofaland-scapetrapthatwepresentcomesfromthemountainash(Euca-lyptusregnans)forestsofsoutheasternAustraliainthecentralhighlandsofVictoria.
Thelikelyregimeshiftisfromlandscapesdominatedbyold-growthforeststhatare200–450yofagetothosedominatedbyyoungre-proneforeststhatdonotsurvivetobe-comeoldgrowth.
Evidencecomesfromnewspatialinformationfollowingmassivewildresin2009,perhapsthemosteconomicallydestructiveinAustralianhistory(19),coupledwithunderstandingthathasemergedfrom28yofextensiveeldinformationandassociateddataanalysesinmountainashforests(20).
ThecentralhighlandsofVictoriasupportw121,000haofmountainashforest.
Thesearespectacularforestswithold-growthtreesreaching90mormoreinheight(14).
Mountainashforestspersistonlywithinaparticularreregime(sensu21).
BeforeEuropeansettlementover150yago,thereregimewasinfrequentseverewildrethatoccurredinlatesummer(22).
Youngseedlingsgerminatefromseedreleasedfromthecrownsofburnedmaturetreestoproduceaneweven-agedstand(20).
Wildresmaybestand-replacing,becausetheyoungtreesregeneratingafterrebelongtoasingleagecohort(23).
Whentheintervalbetweenstand-replacingdisturbancesislessthan20–30y(whichistheperiodrequiredfortreestoreachsexualma-turityandbeginproducingseed)(24),standsofmountainashforestwillbereplacedbyotherspecies,particularlywattle(Acaciaspp.
)(20).
Inthepastcentury,anewdisturbanceregime(logging)hasbeenaddedtothepreviousnaturalreregime.
Largeareasofmountainashhavebeensubjecttotimberandpulpwoodhar-vesting(Fig.
2).
Inthepast40y,thetraditionalmethodoflog-ginghasbeenclear-cutting,inwhichallmerchantabletreeswithina15-to40-haareaarecutinasingleoperation(25).
Followingclear-cutting,loggingdebrisisburnedtocreateabedofashesinwhichtheregenerationofaneweucalyptstandtakesplace,oftenbyarticialreseeding.
Thevastmajorityofmountainashlandscapeshavebecomedominatedbylargeareasofregrowthforestwithsmallareasofoldforestembeddedwithinthem.
Old-growthmountainashforest(sensu20)typicallycoverslessthan3%ofthemajorityofthe3,000-to6,000-hawoodproductionforestblocksinthecentralhighlands;however,insomecases,itislessthan1%(20).
Indeed,followingmorethanacenturyofloggingandwildresin1926,1932,1939,1983,and,mostrecently,2009,w1.
1%oftheentiremountainashforestestateisnowinanold-growthstage.
Thislandscapeisinstarkcontrasttomountainashlandscapes100–150yago,whichhis-toricalaccounts(e.
g.
,26),coupledwithstandreconstructionworkrelatingtotreeageandstemdiametersoflargedead(snag)treesremainingwithinyoungstands(27),suggestweredomi-natedbylargeareasofoldgrowth,possiblyashighas60–80%totalcoverinthecentralhighlandsofVictoria(20)(Fig.
2).
DevelopmentofaLandscapeFire-TrapinMountainAshForests.
Theinteractingeffectsofwildre,logging,andthecombinationofwildreandlogging(i.
e.
,salvagelogging)(sensu28)arecreatingapreviouslyunrecognizedlandscapetrapinwhichthedistur-bancedynamicsof"trapped"mountainashforestlandscapesaremarkedlydifferentfromthosebeforeEuropeansettlement(Figs.
S1andS2).
Thecoreprocessunderlyingthislandscapetrapisapositivefeedbackloopbetweenrefrequency/severityandareductioninforestageatthestandandlandscapelevels,leadingtoanincreasedriskfordenseyoungregeneratingstandsrepeatedlyreburningbeforetheyreachamorematurestate(Fig.
3).
Thelandscapetrapwillpotentiallycreateirreversiblechangesindisturbancedynamics,forestcover,landscapepattern,andvegetationstructure,andtherebyleadtoamajorregimeshiftoralternativestate.
Weexplainbelowtheevidenceforthepositivefeedbackprocessthatunderpinsthislandscapetrap(Fig.
S2)anddiscusswhyitishistoricallyunprecedentedandwhyitisbegin-ningtodominatethecontemporarylandscape.
Positivefeedbackloopbetweenreducedforeststandageandre.
Youngstandsofmountainashforestarecreatedbynaturalre-generationfollowingwildre.
Detailedon-sitemeasurementsfollowingthe2009wildreshaverevealedthatyoungforestburnsathigherseveritythanmatureforest.
Wesuggestthisisforfourkeyreasons:Fig.
1.
Conceptualmodelofalandscapetrap.
Thetrapresultsfromthereinforcingfeedbackloopshowninred.
Fig.
2.
Photomontageshowinghistoricallogginginextensivestandsofold-growthforest(A–C)andextensiveclear-cutareasofforestcutinthepast10y(DandE)inthemountainashforestinthecentralhighlandsofVictoria.
(PhotoscourtesyofNationalArchivesofAustralia,StateLibraryofVictoriaandD.
B.
L.
)FireorloggingClear-cutloggingReproductive,even-aged=LandscapetrapYoung,even-agedReseeded,singlespeciescanopyAcacia"OLDGROWTH""Mixedage"(relictoldtrees+regeneration)FireFig.
3.
DevelopmentofalandscapetrapinthemountainashforestsofthecentralhighlandsofVictoria.
15888|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1110245108Lindenmayeretal.
DownloadedbyguestonMarch14,2021i)Youngregeneratingstandsofmountainashtreesarechar-acterizedbydenselyspacedregrowthsaplings.
Therecanbeseveralmillioneucalyptseedlingsperhectaresoonafterareorlogging.
Throughprocessesofrapidnaturalself-thinning,thisdeclinestow400stemsperhectareat40yand40–80stemsperhectareinmatureforestafter150–200y(29).
Themarkedreductioninthenumberofstemsperunitareaovertimeisprimarilyattributabletocompetition-de-riveddeathandcollapseofsmallsuppressedpoleandsap-lingtrees,whichaddgreatlytothedensityofthevegetationinyoungregrowingforestsbutdonotgenerallyoccurinmatureandold-growthmountainashforests(30).
Denselyspacedstandsofregrowthsaplings,coupledwiththesub-sequentnaturalprocessesofrapidself-thinningthatcharac-terizetheearlystagesofstandregenerationinmountainashforests,createsignicantlymoreneandmediumfuelsthaninoldforests(31).
ii)Thecloselyspacedcrownsindenselystockedyoungstandsarereadilysusceptibletocarryingacrownre.
Thisisincontrasttoold-growthstands,whicharecharacterizedbylargerelativelywell-spacedtreeswithopencrownsandsmalllateralsubcrowns(24).
iii)Treesinyoungstandsareshorterthanthoseinold-growthstands.
Theameheightneededtoscorchorconsumethecanopyinyoungstandsisthereforesignicantlylowerthaninold-growthstands(22).
iv)Youngforestssupportsignicantlysmallerdiameterlogsonthegroundthanold-growthstands(32).
Suchsmallerdiameterlogssupportsignicantlylessdenseandluxuriantmossmatsthanlargerdiameterfallentrees.
Mossmatsholdlargeamountsofwater(1,100%ofdryweight)(33);theyplayasignicantroleinmoistureretentionwithinlogs,andtherebymayreducetheriskforburning.
WhyhasthispositivefeedbackloopnotoccurredhistoricallyBeforeEuropeansettlement,frequent,widespread,high-severitywild-resinmountainashforestswouldhavebeensuppressedbyacombinationofextendedperiodsofwetclimaticconditionsandtheabsenceoftheintensivehumandisturbancesresultingfromclear-cutlogging.
Thisfavoredanegativefeedbackloopbetweenforestageandre,enablingyoungforesttomatureintoalessre-pronestatethatwasnotconducivetowidespreadhigh-severitywildre(Fig.
S1).
WhyisthispositivefeedbackloopnowbeginningtodevelopTwomajorchangeshaveoccurredrelativelyrecentlytofavorthepositivefeedbackloop:reducedforestageinmountainashforestsandincreasedrefrequency(Fig.
3andFigs.
S1andS2).
First,therehasbeena25%reductioninrainfallinsoutheasternAustraliaoverthepastfewdecades(34).
Second,logginghasconvertedmorethan90%offormerlyoldforesttoyoungregeneratingstands.
Youngforestresultingfromclear-cutlog-ginghastwoaddedelementsofreproneness:(i)nefuelscreatedbyloggingoperationsareaddedtothosefromthecol-lapseofsmall-diameterstemsandsheddingofbranchesduringnaturalself-thinningandself-pruningprocessesindenselystockedregeneratingstands,and(ii)thespatialpatternofstandageclassesinmountainashlandscapeshasbeenaltered,withanincreasedprevalenceofyoungdenselystockedforestandasig-nicantlyreducedareaof(mesic)old-growthforest.
This,inturn,hasincreasedtherecontagioninthelandscape.
Codesofloggingpracticeandthepracticallogisticsofhar-vestingoperationsmeanthatclear-cuttingisappliedtoatterandmoreaccessiblepartsofmountainashlandscapes.
However,theseplacesarealsowhereold-growthstandswereformerlymostlikelytooccur.
Evidenceforthiscomesfromworkinclosed-watercatchmentsofthecentralhighlandsofVictoria,wheretherewerenoconfoundingeffectsofpastandpresenthumandisturbancesthatwouldhaveotherwiseobscuredkeyspatialpatternsofforestageclasses(22).
Beforethe2009wildres,oldgrowthmountainashoccupiedasubsetoftheoverallenvironmentaldomainofmountainashperse,typicallywithinanarrowbandofmesicsitesratherthanridgesorsteepslopes.
Thisenvironmentaldomainwasnotonlyfavorablefortreegrowthbutinteractedwithspatialdifferencesinnaturaldisturbanceregimes(35).
Mesicsitessupporttallertrees.
Theyarealsoplaceswhereboththerefrequencyandtheintensityofpastwildreswereattenuated(22).
Formerareasofold-growthforestonatterrainhavenowbeenconvertedtoyoungregen-eratingstandsandarespatiallyconnectedtoyoungburnedorloggedforestonmidslopesandridges.
Importantly,themorewidespreadthatyoungloggedandregeneratedforestbecomes,thegreateristheriskforincreasingspatialcontagioninthespreadofwildrethroughlandscapes(31),becausemoistrem-nantareasthatwouldhaveslowedorhaltedthespreadofre(andformerlysupportedoldforest)havebeenconvertedtoyoungforest.
Spatialcontagioninrecurrenthigh-severityremaythereforereinforceapatternofincreasinghomogeneityinthecoverofyoungforestinalandscape(Fig.
S2).
Thispatternoccursbecausesomeareasofrerefugia(e.
g.
,atplateau,deepsouth-facingvalleyoors)thatweretraditionallycharacterizedbyalongabsenceofre(particularlyhigh-severityre)andsupportedstandsofmultiagedforestorold-growthforest(35)becomemoresusceptibletobeingburnedbyhigh-severitycon-agrationsthatspreadfromadjacentmoreammableloggedandyoungregeneratingareas(Figs.
S1andS2).
Notably,al-thoughnaturaldisturbanceregimesoftenincreaseheterogeneityinmanylandscapes(36),theoppositefrequentlyoccursinareassubjecttolandscapetrapphenomena,inwhichthecombinationofhumanandnaturaldisturbanceregimescanleadtoincreasedlandscapehomogeneity.
Researchinmoistforestsaroundtheworldsuggeststhatotherfactorsassociatedwithloggingmayincreasesusceptibilityofyoungregeneratingforeststobeingburnedorreburningathighseverity.
Forexample,thelargequantitiesofloggingslashcre-atedbyharvestingoperationscansustainresforlongerthanfuelsinunloggedforest(12).
Similarly,lightningstrikeignitionismorelikelytooccurinharvestedstandsbecauseofincreasednefuelsresultingfromloggingslash,andthiseffectmayremainfor10–30yfollowinglogging(37).
Finally,theremovaloftreesbyloggingcreatesmicroclimaticconditionsthatleadtoincreaseddryingofunderstoryvegetationandtheforestoor,andacor-respondinglyelevatedrerisk(38).
Onceamountainashforestlandscapeisdominatedbywide-spreadareasofyoungre-proneforest,theelevatedriskforhigh-severityspatiallycontagiousredecreasestheprobabilitythatthelandscapecanreturntoitsformermaturestate,particularlyunderthedrierandwarmerconditionsassociatedwithclimatechange.
Hence,thedynamicsoftrappedmountainashforestlandscapesaredifferentfromthoseinthepast(>100yago)(Fig.
3andFigs.
S1andS2).
Thecurrentsetofinteractingdisturbanceregimesofre,logging,andpostre(salvage)loggingdidnotexistbeforeEuropeansettlement.
Importantly,thereisamajorasymmetryintheperiodduringwhichmountainashforestecosystemshavecoevolvedwithnaturaldisturbances(>20milliony)comparedwiththe20–100yduringwhichtheinteractinghumanandnaturaldisturbanceregimeshaveproducedalandscapetrap.
Endpoint:RegimeshiftThepositivefeedbackcycleofwidespreadyoungregeneratingstandsandfrequenthigh-severitywildremeansthateitherextensiveareasoftrappedyoungmountainashforestwillbemaintainedorafurtherregimeshiftwilloccurinwhichanewtypeofvegetationcoverdevelops,particularlywattle(Acaciaspp.
)(Fig.
3andFigs.
S1andS2).
Oncemountainashhasbeeneliminatedfromanextensivearea,itrecolonizesslowlybecausetheseedreleasedfromthecrownsofburnedmaturetreesdispersesw1.
5–2.
0crownheightsfromasourceLindenmayeretal.
PNAS|September20,2011|vol.
108|no.
38|15889ECOLOGYDownloadedbyguestonMarch14,2021treeandsuccessfulregeneration(re)eventsmayoccurevery30–400y.
Therefore,theregenerationniche,whichisakeypartofthelifecycleofmountainash(39),ismaladaptedtothealteredlandscapeconditionsandalteredreregimecreatedbyrecurrentloggingandwildre.
Recurrenthigh-frequencywildremayresultinrepeatedlyburnedareasthatwereformerlydom-inatedbymountainashbeingcolonizedbyothereucalyptspeciesthatdonotdependonseedlingregenerationbut,instead,recoverafterwildreviastrategieslikeepicormicresprouting[e.
g.
,shininggum(Eucalyptusnitens),messmate(E.
obliqua)].
Irrespectiveofwhethermountainashforestlandscapesremaintrappedaswidespread,young,re-pronestandsorundergoaregimeshifttoextensiveareasdominatedbyAcaciaspp.
andotherspecies,suchchangeswillresultinsignicantimpairmentofecologicalfunctionslikecarbonstorage,waterproduction(40,41),andbiodiversityconservation.
Forexample,neitheryoungsmall-diametermountainashtreesnorAcaciaspp.
supportthecavitiesthatarecrucialnestinganddenningsitesformanyspeciesofanimals.
Theyalsolackcriticalstructuralfeatures,suchasextensivebarkstreamers,thatarekeyforagingmicro-habitatsforwildlife(42).
Thesechangesinvegetationstructurearelikelytoleadtoirreversiblelossesinhabitatsuitabilityforw40speciesofvertebratesinmountainashforeststhatarede-pendentonlarge120-to150+-y-oldtreeswithhollows.
AvoidingaLandscapeTrapinMountainAshForestsofVictoria.
ThreeimportantstrategiesareneededtoreducetheproblemscreatedbythelandscapetrapinthemountainashforestsofVictoria.
First,large(>1,000ha)areasofcurrentlyunburnedforestneedtoberetained,whereinthenumberofanthropogenicstressorsisreduced.
Theareaofgreenforestwasreduceddramaticallybythe2009wildres;hence,relativebiodiversity,carbonstorage,andwaterproductionvaluesofremainingunburnedforesthaveincreased.
However,suchuncommonareasofunloggedforestareincreasinglysoughtafterfortimberandpulpwoodharvestingbecause(i)theyareamongthedecliningnumberofplacessuitableforcuttingasaconsequenceofpastresandpast(prere)loggingoperations,(ii)therearelegislatedguaranteestoprovideloggingcontractorswithforesttocutfortimberandpulpwood(43),and(iii)cuttingburntforest(i.
e.
,salvagelog-ging)hasmajornegativeenvironmentalimpactsandlong-termeffectsonforestrecoveryandforestbiodiversity(28).
Targetinglimitedremainingareasofunburnedforestforloggingdepletestheoverallamountoftheseforests,withlong-termeconomicimplicationsforharvestcontractors.
Increasedloggingpressureongreenareashasotherecologicalimplications:Remainingareasofgreenforestareimportantrefugiaforbiodiversityfol-lowingwildresandarecriticalforunderpinningpostreeco-logicalrecovery(32).
Legislativeandotherimpedimentstoreducingharvestlevelshighlighttheexistenceofmanagementandsocioeconomictrapswithinlandscapetraps,andtheseneedseriousandtimelyreview.
Asecondstrategytoavoidthedevelopmentofalandscapetrapinthenowhighlyre-pronemountainashlandscapesofVictoriaistorecalculatethesustainedyieldtoaccommodatefuturelossesoftimberresultingfromtheinevitableburningofsomepartsofforestlandscapes.
Thisstrategyhastheadvantageofnotovercommittingremainingunloggedgreenforestintheeventofwildres,therebyresultinginmoreconservativeman-agementofnaturalresourcesandmoreexplicitrecognitionoftheuncertaintycreatedbymajornaturaldisturbances.
GiventheextentofrecentlyburnedforestinVictoria,athirdimportantstrategytoreducetherisksfordevelopmentofaland-scapetrapistotrytolimittheamountoffuturere.
Althoughmountainashtreesaredependentonretopromoteregeneration,reshavebeenextensiveinthepast25–100y;anotherreinthecoming20ywithincurrentlyyoungregeneratingstandsislikelytoleadtoamajorregimeshift(Fig.
3).
Reducingtheamountofreinmountainashforestsisasignicantchallenge.
Broad-areapre-scribedburningisnotaviablemanagementoptionbecausehighlevelsofmoistureinthevegetationandlargequantitiesofbiomassmakeplannedresextremelydifculttocontrol(20).
However,prescribedburningaspartofaregimeofrecanbeanappropriatemanagementoptionindrierforesttypesthatareadjacenttomountainashforests.
Carefullyappliedstrategicburninginsuchdrierenvironmentsmayhelptoreducetheextentofspatialcon-tagioninwildrethatoccursintheseareasand,inturn,reducetheriskforadjacentstandsofmountainashforestbeingburned(44).
ExamplesofLandscapeTrapsinEcosystemsOtherThanForests.
Wecontendthatlandscapetrapsmaybeprevalentinmanyecosys-tems.
Forexample,climatechangeandovershinghavefacili-tatedtheconversionofsubtidalkelp(Macrocystispyrifera)forestsinTasmaniancoastalwatersto"barrens"habitatresultingfromovergrazingbytheseaurchinCentrostephanusrodgersii.
Oceanwarmingandalteredcirculationpatternshaveenabledthepolewardspreadofthisseaurchin(45),andovershingofpredators,suchasthesouthernrocklobster(Jasusedwardsii),hasenabledC.
rodgersiitoestablishhigh-populationdensitybarrensthatresultinthelossofbiodiversityandareductionintheproductivityofsheriesandcontributetothedeclineofsuchpredatorsasJ.
edwardsii(46).
Aquaticenvironmentswherewaterqualitycanberadicallyalteredbynutrientinputsfromhumanactivities(e.
g.
,47)alsoaresusceptibletothedevelopmentoflandscapetraps.
GrazingonpubliclandsinthewesternUnitedStateshasbeenblamedforreducingbiodiversityand,togetherwithexoticweeds,mayhaveledthesegrasslandecosystemsintoalandscapetrapthatproducesaplantcommunityfromwhichthereisnogoingback(48).
LivestockgrazinginwesternUnitedStatesmayhavereducedtheabundanceofpreferredplantspecieswhilesub-jectingthesoiltoweedinvasion,suchthatlargeareasarenowdegradedrangelandsinthesamemannerillustratedineasternAustraliabythe"woodyweed"probleminsemiaridwoodlands(49).
Introducedgrasses,suchascheatgrass(Bromustectorum),cansimilarlymovegrasslandcommunitiesintheintermountainwesternUnitedStatesintoaregimechangethatisnearlyim-possibletoreverse(50,51).
Alackofreversiblechangemaybebestillustratedbylandscapetrapsinregionsheavilyimpactedbydisturbanceslikemountaintopmining(52).
ConcludingCommentsWesuggestthatstrategiesandmanagementinterventionsareneededtoreducetheprobabilityoflandscapetrapsdeveloping(Fig.
4).
Oneapproachistorecognizethatlandscapetrapscanexistandidentifythesuiteofspatialandtemporalcharacteristicsthatcancombinetogiverisetothem,including(i)exploitationofthenaturalresourcesinalandscapethroughunsustainablelevelsofharvesting;(ii)alterationinthespatialcharacteristicsoflandscapes,includingmodicationstothefrequencyandseverityofecologicaldisturbances;(iii)feedbacksbetweenalteredenvi-ronmentalconditionsandothermajoranthropogenicstressors;Fig.
4.
Conceptualmodelhighlightingsignalsandinterventionsrequiredtoreversethedevelopmentofalandscapetrap.
15890|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1110245108Lindenmayeretal.
DownloadedbyguestonMarch14,2021and(iv)severelyimpairedlandscapeprocessesandfunctions.
Asecondapproachistolimitthenumberofanthropogenicstressorsinlandscapesandreducethepotentialfornegativeinteractionsamongmultiplestressors.
Thismayequatetoamoreconservativeapproachtotheharvestingofnaturalresourcesor,inothercases,applicationofmanagementstrategiesthatreducefeedbacks(e.
g.
,fuelreductionthroughprescribedburning).
Sustainedyieldsofnaturalresourcesalsomayneedtoberapidlyreassessedfollow-ingcatastrophiceventstoavoidovercommittingremainingintactareasandfurtherincreasingtheriskforcreatingalandscapetrap.
Wesuggestthattheneedforproactivemanagementtopre-ventthedevelopmentoflandscapetrapsiscritical,giventhat(i)landscapetrapsmightbeatincreasedriskfordevelopmentinresponsetosignicant"events"likemajornaturaldisturbances,whicharelikelytobecomemorefrequent,moresevere,orbothunderrapidclimatechangeinmanyregions(e.
g.
,53,54),and(ii)markedasymmetryexistsbetweentherapiditywithwhichlandscapetrapsmaydevelopandtheprolongedtimescales(hundredstothousandsofyears)thatcharacterizenaturaleco-logicalprocessesandnaturaldisturbanceregimes.
ACKNOWLEDGMENTS.
Prof.
R.
Mitchell,Dr.
D.
DellaSala,Prof.
D.
Bowman,andDr.
A.
Gillmadeastutecommentsthatimprovedearlierversionsofthismanuscript.
1.
BeisnerBE,HaydonDT,CuddingtonK(2003)Alternativestablestatesinecology.
FrontEcolEnviron1:376e382.
2.
CarpenterSR,etal.
(2011)Earlywarningsofregimeshifts:Awhole-ecosystemex-periment.
Science332:1079e1082.
3.
PaineRT,TegnerMJ,JohnsonEA(1998)Compoundedperturbationsyieldecologicalsurprises.
Ecosystems(NewYork,N.
Y.
)1:535e545.
4.
BiggsR,CarpenterSR,BrockWA(2009)Turningbackfromthebrink:Detectinganimpendingregimeshiftintimetoavertit.
ProcNatlAcadSciUSA106:826e831.
5.
FolkeC,etal.
(2004)Regimeshifts,resilience,andbiodiversityinecosystemman-agement.
AnnuRevEcolSyst35:557e581.
6.
WarmanL,MolesAT(2009)AlternativestablestatesinAustralia'swettropics:Atheoreticalframeworkfortheelddataandaeld-caseforthetheory.
LandscapeEcol24:1e13.
7.
CasiniM,etal.
(2009)Trophiccascadespromotethreshold-likeshiftsinpelagicmarineecosystems.
ProcNatlAcadSciUSA106:197e202.
8.
KleinBC(1989)Effectsofforestfragmentationondungandcarrionbeetlecommu-nitiesincentralAmazonia.
Ecology70:1715e1725.
9.
WalkerBH,SaltD(2006)ResilienceThinking(IslandPress,Washington,DC).
10.
GilpinME,SouléME(1986)ConservationBiology.
TheScienceofScarcityandDi-versity,edSouléME(Sinauer,Sunderland,MA),pp19e134.
11.
AndrenH(1994)Effectsofhabitatfragmentationonbirdsandmammalsinland-scapeswithdifferentproportionsofsuitablehabitat—Areview.
Oikos71:355e366.
12.
CochraneMA,SchulzeMD(1999)FireasarecurrenteventintropicalforestsoftheeasternAmazon:Effectsofforeststructure,biomass,andspeciescomposition.
Biotropica31:2e16.
13.
vanNieuwstadtMG,ShielD,KartawinataD(2001)TheecologicalconsequencesofloggingintheburnedforestsofeastKalimantan,Indonesia.
ConservBiol15:1183e1186.
14.
MalhiY,etal.
(2009)Exploringthelikelihoodandmechanismofaclimate-change-induceddiebackoftheAmazonrainforest.
ProcNatlAcadSciUSA106:20610e20615.
15.
VeldmanJW,MostacedoB,Pena-ClarosM,PutzFE(2009)SelectiveloggingandreasdriversofaliengrassinvasioninBoliviandryforest.
ForEcolManage258:1643e1649.
16.
ThompsonJR,SpiesTA,GanioLM(2007)Reburnseverityinmanagedandun-managedvegetationinalargewildre.
ProcNatlAcadSciUSA104:10743e10748.
17.
OdionDC,etal.
(2004)PatternsofreseverityandforestconditionsinthewesternKlamathMountains,California.
ConservBiol18:927e936.
18.
LindenmayerDB,HunterML,BurtonPJ,GibbonsP(2009)Effectsofloggingonreregimesinmoistforests.
ConservLett2:271e277.
19.
2009VictorianBushresRoyalCommission(2010)FinalReport(ParliamentofVicto-ria,Melbourne).
20.
LindenmayerDB(2009)ForestPatternandEcologicalProcess:ASynthesisof25YearsofResearch(CSIROPublishing,Melbourne).
21.
GillAM(1975)FireandtheAustralianora:Areview.
AustFor38:4e25.
22.
MackeyB,LindenmayerDB,GillAM,McCarthyMA,LindesayJA(2002)Wildlife,FireandFutureClimate:AForestEcosystemAnalysis(CSIROPublishing,Melbourne).
23.
AshtonDH(1981)FireandtheAustralianBiota,edsGillAM,GrovesRH,NobleIR(AustralianAcademyofScience,Canberra,Australia),pp339e366.
24.
AshtonDH(1975)TherootandshootdevelopmentofEucalyptusregnansF.
Muell.
AustJBot23:867e887.
25.
LutzeMT,CampbellRG,FaggPC(1999)DevelopmentofsilvicultureinthenativeStateforestsofVictoria.
AustFor62:236e244.
26.
HoughtonN(1986)TimberMountain(LightRailwayResearchSocietyofAustralia,Melbourne),p106.
27.
LindenmayerDB,McCarthyMA(2002)Congruencebetweennaturalandhumanforestdisturbance:AcasestudyfromAustralianmontaneashforests.
ForEcolManage155:319e335.
28.
LindenmayerDB,BurtonPJ,FranklinJF(2008)SalvageLoggingandItsEcologicalConsequences(IslandPress,Washington,DC).
29.
AshtonDA,AttiwillP(1994)AustralianVegetation,edGrovesRH(CambridgeUnivPress,Melbourne),pp157e196.
30.
AshtonDH(1976)Thedevelopmentofeven-agedstandsofEucalyptusregnansF.
Muell.
incentralVictoria.
AustJBot24:397e414.
31.
WhelanRJ(1995)TheEcologyofFire(CambridgeUnivPress,Cambridge,UK).
32.
BanksSC,DujardinM,McBurneyL,BlairD,LindenmayerDB(2011)Startingpointsforsmallmammalpopulationrecoveryafterwildre:Recolonization,refugiaorresidualpopulationsOikos120:26e37.
33.
AshtonDH(1986)EcologyofbryophyticcommunitiesinmatureEucalyptusregnansF.
Muell.
forestatWallabyCreek,Victoria.
AustJBot34:107e129.
34.
CaiW,CowanT(2008)DynamicsoflateautumnrainfallreductionoversoutheasternAustralia.
GeophysResLett35:L09708.
35.
LindenmayerDB,etal.
(1999)Factorsaffectingstandstructureinforests—ArethereclimaticandtopographicdeterminantsForEcolManage123:55e63.
36.
LindenmayerDB,FranklinJF(2002)ConservingForestBiodiversity:AComprehensiveMultiscaledApproach(IslandPress,Washington,DC).
37.
KrawchukMA,CummingSG(2009)Disturbancehistoryaffectslightningreinitiationinthemixedwoodborealforest:Observationsandsimulations.
ForEcolManage257:1613e1622.
38.
UhlC,KauffmanJB(1990)Deforestation,resusceptibility,andpotentialtreere-sponsestoreintheEasternAmazon.
Ecology71:437e449.
39.
NitschkeC,HickeyG(2007)AssessingthevulnerabilityofVictoria'sCentralHighlandforeststoclimatechange.
UniversityofMelbourneTechnicalReport1/2007(De-partmentofSustainabilityandEnvironment,Melbourne).
40.
KeithH,MackeyBG,LindenmayerDB(2009)Re-evaluationofforestbiomasscarbonstocksandlessonsfromtheworld'smostcarbon-denseforests.
ProcNatlAcadSciUSA106:11635e11640.
41.
VertesseyRA,WatsonFG(2001)FactorsdeterminingrelationsbetweenstandageandcatchmentwaterbalanceinMountainAshforests.
ForEcolManage143:13e26.
42.
LindenmayerDB,CunninghamRB,DonnellyCF,FranklinJF(2000)StructuralfeaturesofoldgrowthAustralianmontaneashforests.
ForEcolManage134:189e204.
43.
VictorianGovernment(2009)2009Victoria'sTimberIndustryStrategy(DepartmentofPrimaryIndustries,Melbourne).
44.
KirkpatrickJB,DellaSalaDA(2011)TemperateandBorealRainforestsoftheWorld:EcologyandConservation,edDellaSalaDA(IslandPress,Washington,DC),pp195e212.
45.
LingSD,JohnsonCR,RidgwayK,HobdayAJ,HaddonM(2009)Climate-drivenrangeextensionofaseaurchin:Inferringfuturetrendsbyanalysisofrecentpopulationdynamics.
GlobChangeBiol15:719e731.
46.
LingSD,JohnsonCR,FrusherSD,RidgwayKR(2009)Overshingreducesresilienceofkelpbedstoclimate-drivencatastrophicphaseshift.
ProcNatlAcadSciUSA106:22341e22345.
47.
HaslerAD(1947)Eutrophicationoflakesbydomesticdrainage.
Ecology28:383e395.
48.
FreilichJE,EmlenJM,DudaJJ,FreemanDC,CafaroPJ(2003)Ecologicaleffectsofranching:Asix-pointcritique.
Bioscience53:759e763.
49.
NobleJC(1997)TheDelicateandNoxiousScrub:CSIROStudiesonNativeTreeandShrubProliferationintheSemi-AridWoodlandsofEasternAustralia(CSIROWildifeandEcology,Canberra,Australia).
50.
YoungJA,ClementsCD(2009)Cheatgrass:FireandForageontheRange(UniversityofNevadaPress,Reno,NV).
51.
D'AntonioCM,VitousekPM(1992)Biologicalinvasionsbyexoticgrasses,thegrass/recycle,andglobalchange.
AnnuRevEcolSyst23:63e87.
52.
PalmerMA,etal.
(2010)Scienceandregulation.
Mountaintopminingconsequences.
Science327:148e149.
53.
LenihanJM,DrapekR,BacheletD,NeilsonRP(2003)Climatechangeeffectonveg-etationdistribution,carbon,andreinCalifornia.
EcolAppl13:1667e1681.
54.
MarlonJR,etal.
(2009)WildreresponsestoabruptclimatechangeinNorthAmerica.
ProcNatlAcadSciUSA106:2519e2524.
Lindenmayeretal.
PNAS|September20,2011|vol.
108|no.
38|15891ECOLOGYDownloadedbyguestonMarch14,2021