Ecosystem Loss and Fragmentation

8/3/2019 Ecosystem Loss and Fragmentation

1/25

SYNTHESIS

Lessons in conservation

http://n ep.amnh.org/lin

Ecosystem Loss andFragmentationMelina F. Laverty * and James P. Gibbs

*American Museum o Natural History, New York, NY, U.S.A., email [email protected] SUNY-ESF, Syracuse, NY, U.S.A., email jpgibbs@es .edu

2

Ecosystem Loss and Fragmentation

K .

F r e y


2/25

SYNTHESIS



3

Table o Contents

Introduction...........................................................................................................74Habitat Loss by Biome..............................................................................................74

Terrestrial...........................................................................................................75Forests - Tropical and Temperate...................................................................... .75Grasslands - Tropical, Temperate, and Tundra.............. .......................................76

Aquatic..............................................................................................................77Wetlands........................................................................................................78Riverine Systems.............................................................................................78

Causes o Fragmentation.........................................................................................79Fragmentation Due to Natural Causes..................................................................79Fragmentation Due to Human Activity..................................................................79Natural Versus Human Fragmentation...................................................................80

E ects o Fragmentation.........................................................................................80Decreasing Patch Size.......................................................................... .................80Increased Edge E ects......................................................................... .................81Edge E ects........................................................................................................81

Edge E ects - Physical......................................................................................81Edge E ects - Biological...................................................... ............................82Invasion by Generalist Species........................................................................82Alteration o Plant Communities..................................................................82Alteration o Insect Communities and Nutrient Cylcing..............................82

Isolation Barriers to Dispersal................................ ...........................................83Species Response to Isolation........................................................................83E ect o Time on Isolation..........................................................................83

E ects o Di erent Types o Fragmentation.....................................................84E ects on Species Abundance, Richness, and Density......................................84Interactions Among Species and Ecological Processes.......................................85Box 1. Corridors and Connectivity....................................................................86Box 2. The Futi Corr idor Linking Tembe Elephant Park, South A rica to MaputoElephant Reserve, Mozambique.........................................................................87Box 3. Identi ying Species Vulnerable to Fragmentation.....................................88

Management o Fragmented Landscapes................................................................88Recommendations.................................................................................................89Terms o Use..........................................................................................................89Literature Cited......................................................................................................90Glossary..............................................................................................................95



3/25

SYNTHESIS



4


Ecosystem Loss and FragmentationMelina F. Laverty and James P. Gibbs

Introduction

Ecosystem loss and ragmentation has been termed the great-est worldwide threat to biodiversity and the primary causeo species extinction (Wilcox and Murphy, 1985; Rosenbergand Raphael, 1986; Simberlo , 1986). Today, as Laurence andBierregaard (1997) have stated, the ragmented landscape isbecoming one o the most ubiquitous eatures o the tropicalworld and indeed, o the entire planet. Moreover, eco-system ragmentation is as much an issue or biodiversity inaquatic, including marine, environments as it is or terrestrialones (Bostrom et al., 2006).

Ecosystem loss and ragmentation are related processes andtypically occur simultaneously. Indeed, some texts (e.g., Me -

e and Carroll, 1997) de ne ragmentation as the loss andisolation o natural habitats. However, the two processes aredistinct (Fahrig, 2003). Ecosystem loss re ers to the disap-pearance o an ecosystem, or an assemblage o organisms and

the physical environment in which they exchange energy andmatter. Many studies, however, examine loss with respect toa speci c organisms habitat . Habitat loss is the modi cationo an organisms environment to the extent that the qualitieso the environment no longer support its survival. Habitatloss usually begins as habitat degradation, the process where thequality o a species habitat declines. Once the habitats qual-ity has become so low that it no longer supports that speciesthen it is termed habitat loss. Fragmentation is usually a prod-uct o ecosystem loss and is best thought o as the subdivision

o a ormerly contiguous landscape into smaller units. Ulti-mately, ragmentation reduces continuity and inter eres withspecies dispersal and migration, thereby isolating populationsand disrupting the fow o individual plants and animals (andtheir genetic material) across a landscape. Generally speaking,habitat loss is o ar greater consequence to biological diver-sity than habitat ragmentation (Fahrig, 2003).

This process is well illustrated in southeastern Bolivia, wherea landscape that was once continuously orested has beentrans ormed into patches o orest surrounded by a matrix o agricultural land. A patc h is usually de ned by its area, perim-eter, shape, and composition (e.g., a land cover type - such aswater, orest, or grassland - a soil type, or other variable). Thematri x is simply the most common cover type in any givenlandscape.

Loss and ragmentation are tightly coupled processes as thepattern o loss a ects the degree o ragmentation. For exam-ple, in a 200-hectare orest, a single 100-hectare block couldbe cleared at one site or a arming operation. Alternatively,

orest could be cleared into many small plots across the land-scape, leaving 100 orest ragments o one hectare each. Inboth cases the landscape has lost 100 hectares o orest, butin the second scenario the landscape has a much higher levelo ragmentation. The potential consequences or plants andanimals are quite di erent in these two scenarios.

Habitat Loss by Biome

Loss and ragmentation impact most o the earths major bi-ome s rom tropical and temperate orests to grasslands and

rom wetlands to rivers. Quanti ying the extent o this lossand ragmentation is di cult one major problem is de-termining what vegetation existed historically to establish abenchmark or comparisons. Another issue is determining theextent that change is caused by humans versus natural orces

(Clark and Matthews, 1990; Fukami and Wardle, 2005). Manytextbooks show maps o the hypothetical distribution o theworlds biomes with todays climate, i there were no humans.These maps re er to the present potential vegetation thatis the potential vegetation i there were no humans to removeit. Additional maps illustrate earlier times when climates weredi erent and human impact was minimal: 5,000, 10,000 or


4/25

SYNTHESIS




more years ago. What is the basis o these maps and how ac-curate are they?

Maps or 5,000 or more years ago are largely determined bypast climate, as human infuence was still limited. Evidence o past climate patterns are compiled rom plant and zoological

ossils, as well as soil and sedimentological analyses. Maps o present potential vegetation combine existing vegetation andclimate patterns with remnant vegetation patches. With thesemaps there are obviously higher levels o uncertainty in areasthat are heavily infuenced by human activity versus thosethat have limited human impact. In other words, areas that

comparisons over time;limited groundtruthing o satellite data; andpoor or erratic government reporting.

These actors must all be kept in mind when examining dataon the extent and rate o ecosystem loss and ragmentation.Despite these challenges, these data are critical to conserva-tion e orts and monitoring. Because o its importance, in re-cent years e orts have been made by several organizations,such as the World Resources Institute (WRI), Wetlands Inter-national, and Tropical Ecosystem Environment Observationby Satellite (TREES), to streamline habitat classi cation and

have been heavily a ected by hu-man activity or thousands o years,such as Europe, are more di cult torecreate, while areas like the Arctictundra or Canadas boreal orest areeasier to establish. For a detailed dis-cussion o the challenges in recon-structing and understanding globalvegetation patterns, see Adams andFaure (1997).

E orts have been made to quan-ti y the extent and rate o loss o the worlds major biomes at vari-ous scales and or di erent timeperiods (Turner and Clark, 1990;Skole and Tucker, 1993; Adams andFaure, 1997; Davidson et al., 1999;Steininger et al., 2001; Achard et al.,2002; Etter et al., 2006). This process is complex and estimatesvary widely due to:

di erences in classi cation methods ( or example, wetlandinventories in the United States, Canada, and Mexico areall based on slightly di erent de nitions or wetlands);limited data or some regions ( or example, typically thereis less data or A rica than North America);lack o comparable land cover data rom di erent timeperiods (particularly historical data) that would allow

produce better comparisons on broad scales (Davidson et al.,1999; Matthews et al., 2000; White et al., 2000; Achard et al.,2003).

Terrestrial

Forests - Tropical and Temperate Today orest cover has shrunk to approximately hal o its

potential extent (Adams and Faure, 1997; Roper and Roberts,1999), replaced by agriculture, grazing, and settlement.

Deforestation in Madagascar (Source: L. Langham)


5/25

SYNTHESIS




Primary orest blocks o a signi cant size exist in only a ewcountries, such as the boreal orests o Northern Canada andRussia, and the Amazon basin o Brazil (Bryant et al., 1997).

The worlds orests began declining thousands o years ago,with the expansion o arming and herding in the MiddleEast and Europe. More recently, rapid population growth, in-dustrialization, and globalization are contributing to rapid de-

orestation in many tropical regions, with orest loss in Braziland Indonesia exceeding 3.5 million hectares in 1995 alone(Roper and Roberts, 1999, based on FAO gures). Whilethere is no question that orest loss and ragmentation is sub-stantial, determining the exact rate o these losses globallyis complex (Roper and Roberts, 1999). While determiningrates at smaller, local scales is o ten easier (Skole and Tucker,1993; Steininger et al., 2001), they too can be controversial.

Furthermore, depending on how orest is de ned, what or-est cover data is presented, or how it is analyzed, the picturewe obtain may end up being quite di erent; or example, bychanging the time periods used in an analysis, de orestation

rates may di er dramatically. According to estimates romthe Tropical Ecosystem Environment Observation by Satel-lite (TREES), a research program that uses satellite imagery

to estimate the extent o the worlds tropical humid orests,between 1990 and 1997, 5.8 (+/- 1.4) million hectares o humid orest were lost each year, which corresponds to a rateo 0.52% per year. A urther 2.3 (+/-0.71) million hectareswere obviously degraded, a rate o 0.20% a year (Achard et al.,2002; Eva et al., 2003). However, other scientists consideredthis result to be an underestimate o tropical orest loss, as itonly included humid tropical orest, while dry tropical orestsare disappearing more rapidly as those areas are o ten moreconducive to agricultural activities (Fearnside and Laurance,2003). For conservation planning, it is also critical to keep inmind the variation in de orestation rates at regional and localscales as di erent strategies might be needed. For example,the average de orestation rate across all o Latin America is0.38%, yet there is a very di erent picture o de orestation i

you look at the provincial level. Rates o de orestation in Bra-zils Acre province are 4.4 percent, substantially higher (Table1). Knowledge o this variation is essential or conservationplanning.

Grasslands - Tropical, Temperate, and TundraEstimates o the extent o the worlds grasslands range rom 40to 56 million km 2 or 30 to 40 percent o the earths land area(Table 2) (Whittaker and Likens, 1975; Atjay et al., 1979; Ol-son et al., 1983; Davidson et al., 2002). These estimates incor-porate temperate and tropical grasslands as well as shrublandand tundra (tundra occurs around the Arctic circle above thelatitude where trees can survive, and is dominated by shrubs,sedges, grasses, lichens, and mosses). Temperate grasslands de-velop in climates that typically have cold winters and summer droughts, and are ound in North America (prairies), Europe

and Asia (steppe), South America (pampas), and South A -rica (veldt) (Roxburgh and Noble, 2001). Tropical grasslandsusually develop in areas with distinct seasons o drought andrain, and include savanna, as well as tropical woodland andsavanna (this designation re ers to grassland associated withshrubs and trees). Herbivory and re are important elementso temperate and tropical grassland systems.

Table 1: De orestation rates

Hotspot areas by continent Annual de orestation rate or sample sites within hotspot area

(range)Latin America 0.38%

Central America 0.8-1.5%

Brazilian Amazon belt

Acre 4.4%

Rondonia 3.2%

Para 1.4-2.7%

Columbia-Ecuador border 1.5%

Peruvian Andes 0.5-1.0%

A rica 0.43% Madagascar 1.4-4.7%

Southeast Asia 0.91%

Southern Vietnam 1.2-3.2%

Source: Modi ed rom Achard et al., 2002


6/25

SYNTHESIS




Table 2: Extent o the worlds grasslands

Whittaker and Likens(1975) a

Atlay et al.(1979) a

Olsen et al.(1983)

Davidson et al.(2002)

Grassland Type Millionkm 2 % o TotalLand Area b Millionkm 2 % o TotalLand Area b Millionkm 2 % o TotalLand Area b Millionkm 2 % o TotalLandAreab

Savanna 15.0 11.6 12.0 9.3 - - 17.9 13.8Tropical Woodlandand Savanna

- - - - 7.3 5.6 - -

Dry Savanna andwoodland

8.5c 6.6 3.5 2.7 13.2 d 10.2 - -

Shrublands e - - 7.0 5.4 - - 16.5 12.7Non-woody grass-land and shrubland

- - - - 21.4 16.5 10.7 g 5.7

Temperate Grassland 9.0 7.0 12.5 9.7 - - - -Tundra 8.0 6.2 9.5 7.3 13.6 10.5 7.4 5.7Total Grassland 40.5 31.3 44.5 34.4 55.5 42.8 52.5 40.5a Desert and semi-desert scrub not includedb Total land area used or the world is 129,476,000 km 2 (excludes Greenland and Antarctica)c Includes woodland and shrublandd Includes dry orest and woodlande Includes hot, warm, or cool shrublands

Davidson et al. (WRI/PAGE) calculations based on GLCCD, 1998, Olsen, 1994 a and b, PAGE land area is based onland cover classi cations or savanna, woody savanna, closed and open shrubland, and non-woody grassland, plus Olsens

category or tundrag Includes non-woody grassland onlyNotes: - means data is not available or has been combined in another category

km 2 to 62,115 km 2)

Additional declines are occurring in grasslands in other partso the world as well. The rate and extent o these declinesis less well documented than in the U.S. and so is harder toquanti y accurately.

Aquatic

Although we o ten think o loss and ragmentation only ina terrestrial context, as these areas are easier to observe, lossand ragmentation is also a concern or aquatic ecosystems.Wetlands, mangroves, seagrasses, rivers, coral ree s, kelp orests,and rocky shorelines are ragmented by natural orces such as

Some o the highest rates o habitat loss and ragmentation inthe world have been in grassland areas, in large part becauseo their suitability or growing crops like wheat and corn,and or grazing (Parkinson, 1997). Conversion o grasslands to

armlands in Western Canada and the U.S. has le t only rem-nants o the original prairie grassland. The U.S. GeologicalSurvey (USGS) estimates that since 1830 over 1 million km 2

o the grasslands o the western US have disappeared.

The tall-grass prairie grassland has decreased by 97 per-cent ( rom 677,300 km 2 to 21,548 km 2)Mixed-grass prairie has declined 64 percent ( rom 628,000km 2 to 225,803 km 2)Short-grass prairie has declined 66 percent ( rom 181,790


7/25

SYNTHESIS




Table 3: Wetland extent (in hectares) in the United States and Canada based on the results o nationalwetland inventory in ormation

Country Wetland Extent(1780s)

Wetland Extent(1980s)

Wetland Extent(1985)

Wetland Extent(1988)

Wetland Extent(>1988)

United States (continental only) 89,488,127 a 42,238,851 a 41,356,092 b - 40,9000,000 b

United States (includes Alaskaand Territories)

158,389,525 a 111,056,479 a - - -

Canada - - - 127,199,000 c 150,000,000 d

a published Dah, 1990;b USFWS, 1998;c published NWWG, 1988;d approximate number based on data indicating total wetland extent in Canada may be as much as 150,000,000 ha based on in orma-tion indicating increase in peatland area (Polestar Geomatics, unpublished)

Source: Modi ed rom Davidson et al., 1999

mented or have had their fow modi ed by human interven-tion, primarily through the creation o dams (Dynesius andNilsson, 1994; Pringle, 1997). According to the World Reg-ister o Dams, between 1950 and 1986, the number o largedams in the world increased seven- old. Most dams are built

or irr igation or or hydroelectric needs; they ragment riversand surrounding environments and change natural water fow

patterns, trans orming loti c into lenti c systems. O the worldsmajor rivers (those greater than 125 miles or 201 km long),only two percent are ree fowing; the remaining 98 percenthave been ragmented or diverted (Benke, 1990).

Fragmentation o rivers has impacted many species. In thePaci c Northwest o the United States, dams have serious-ly a ected salmon populations by preventing salmon romreturning to their native streams to reproduce. Dams havealso contributed to declining reshwater mussel populations.

Ninety percent o the worlds reshwater mussels are oundin North America, and 73 percent o these ace extinction inthe United States. Many North American reshwater musselsmust spend a part o their li ecycle in sh gills to reproducesuccess ully. As an example, dams have blocked the movemento anadromous fsh, which the dwar wedge mussel ( Alasmidonta heterodon) depends upon during its li e cycle. This, coupled

bottom topography, wave action, currents, tides, storm surge,as well as human activities such as draining, diversion, extrac-tion o groundwater, dams, dredging, sedimentation, shing(e.g., trawling, dynamite shing), aquaculture, sea jetties, andboating. Here we highlight loss and ragmentation in two o the many aquatic systems: wetlands and rivers.

WetlandsWetlands have been drastically reduced in area and number inmany regions o the world as they are drained and lled or human use. A recent global review o wetlands identi ed sig-ni cant gaps in our knowledge o their extent and rate o loss(Davidson et al., 1999). Di erences in classi cation schemes aswell as gaps in data (data is especially limited or areas outsideNorth America and Europe) mean that current estimates o global wetland coverage vary widely, rom 560 to 1,279 mil-lion hectares. In the continental United States, where study o

wetlands has been more extensive, wetlands have declined bymore than hal , rom 89 to 42 million hectares between 1780and 1980. The rate o loss is speeding up; by 1985 more thanan additional one million hectares disappeared (see Table 3).

Riverine SystemsMany o the worlds major riverine systems are highly rag-


8/25

SYNTHESIS




with siltation and chemical runo , has led to substantial de-clines in their population.

Freshwater systems are also ragmented by groundwater re-moval, which o ten modi es the temperature structure o streams. For example, in the Southeastern United Statesextraction o groundwater has reduced the amount o coldwater that eeds many streams. Important game species, likestriped bass, use spring- ed areas o rivers as re uges duringhot summer months, as they have high oxygen needs andhigher oxygen levels are ound in colder water (Pr ingle, 1997).As these colder areas disappear, it a ects species that dependupon these conditions.

Causes o Fragmentation

Fragmentation is caused by both natural orces and humanactivities, each acting over various time rames and spatialscales.

Fragmentation Due to Natural Causes

Over long time rames (thousands or millions o years),landscapes are ragmented by geological orces (e.g.,continental dri t) and climate change (e.g., glaciations,changes in rain all, sea level rise).

2. Over short periods (decades or months), natural dis-turbances, such as orest res, volcanoes, foods,land slides, windstorms, tornadoes, hurricanes,and earthquakes, modi y and ragment landscapes.

In addition, landscapes are naturally ragmented by mountainridges, canyons, rivers, and lakes. Some ecosystems also com-monly occur in discrete patches and are thus naturally rag-

mented. Natural processes create the habitat heterogeneityand landscape diversity upon which many species depend.

Fragmentation Due to Human Activity

Humans have modi ed landscapes or thousands o years.Early hunters infuenced the landscape by burning areas to

1.

avor certain game species, and today ranchers keep grasslandsopen in the same way (Schle, 1990). Many human activi-tiesagriculture, settlement (e.g., construction o buildings,

ences etc.), resource extraction (e.g., mining, timber), in-dustrial development (e.g. the construction o hydroelectricdams)alter and ragment landscapes. O these activities, ag-riculture is the leading cause o ecosystem loss and ragmen-tation throughout much o the world today (Vitousek et al.,1997; Tilman et al., 2001).

The process o human-caused ragmentation o ten proceedsin a airly predictable manner. First, an opening is ormed ina matrix o natural habitats: perhaps a road is built that crossesthe landscape. This opening becomes larger as settlement andde orestation occur along the road. Still, the landscape remainslargely orested and although there is habitat loss, ragmenta-tion is minimal. Second, smaller roads are constructed o themain road, increasing access to the orest. The newly accessedareas are subsequently cleared or crops. The landscape beginsto appear ragmented, even though the remaining patches o original orest are still large. This process o subdivision re-peats itsel at a ner and ner scale until the landscape shi ts toone predominated by cleared or degraded land, with patcheso isolated orest. Eventually, all o the landscape may be con-verted or human use, except those spots that are too wet, toodry, or too steep to be useable.

Humans also create distinctive patterns as they ragmentlandscapes, typically leaving patches that are non-randomin size and distribution. An analysis o de orestation in theTierras Bajas region o Southwest Bolivia revealed di erentland cover patterns created by our principal groups o people(Steininger et al., 2001). Colonization by peasant armers, insome cases planned and in others not, le t a complex mosaic o

cropland, secondary orest, and orest remnants. The plannedsettlements ormed pinwheel patterns o linear arms, radiat-ing rom a central town, while the unplanned settlements ap-peared as small, square or rectangular elds along roads. Men-nonite colonies, on the other hand, had settlements along theroad with large, rectangular arms extending behind them,leaving larger orest remnants than the peasant settlement pat-


9/25

SYNTHESIS



0


terns. Industrial soybean arms were distinguished rom theothers by their lack o settlements; these arms ormed linear strips with marked boundaries and windbreaks o trees 20

to 40 meters wide between the strips. Like the Mennonitearms, the industrial arms le t larger orest remnants.

There are several technical terms commonly used in theeld o landscape ecology to de ne di erent stages o theragmentation process or di erent orms o ragmentation o

a landscape. These include per oration (holes punched in alandscape), dissection (initial subdivision o a continuous land-scape), ragmentation (breaking into smaller parts), shrinkage(reduction in size o patches), and attrition (loss o patches).

Natural Versus Human Fragmentation

Several di erences exist between human-caused and naturallyragmented landscapes:

A naturally patchy landscape o ten has a complex structurewith many di erent types o patches. A human- ragment-ed landscape tends to have a simpli ed patch structurewith more distinct edges, o ten with a ew small patcheso natural habitats in a large area o developed land.Patch types in human-modi ed landscapes are o ten un-suitable to many species, while in a heterogeneous naturallandscape most patch types are suitable to a more diversegroup o species.The borders (or edges) o patches in naturally patchylandscapes tend to be less abrupt than in those created byhumans. (Edge e ects are discussed in detail later in thisdocument.)

Certain eatures o human- ragmented landscapes, such as

roads, are novel in the evolutionary history o most wild spe-cies and pose additional threats. Not only do they restrictmovement between populations, but heavily traveled roadsare a direct danger to wildli e (Forman and Alexander, 1998;Gibbs and Shriver, 2002). Furthermore, some animals avoidhabitats near roads due to noise pollution. Roads also havesecondary impacts on ecosystems and species. They are an ac-

1.

2.

3.

cess point, increasing a regions vulnerability to invasion byexotic species, and perhaps most importantly, making wildli ehabitats accessible to people or hunting or resource extrac-

tion (Findlay and Bourdages, 2000). In West A rica, or ex-ample, new roads or logging act as conduits or the bushmeattrade, which has contributed to the extirpation o many dui-ker species ( Cephalopus spp.) and the extinction o at least oneprimate species, Miss Waldrons red colobus monkey ( Procolobus badius waldroni ) (Newing, 2001; Whit eld, 2003).

E ects o Fragmentation

Fragmentation and loss o ecosystems are coupled process-es: ragmentation is a consequence o loss (Haila, 1999). Itis o ten di cult to distinguish between the e ects o thesetwo processes, since they o ten happen simultaneously. Losso habitat impacts species principally by reducing availableresources and microenvironments. Fragmentation has addi-tional consequences or species on top o those caused bylossmost importantly, a ecting movement and dispersal andmodi ying behavior.

As ragmentation progresses in a landscape, three major con-sequences are apparent:

decreasing patch size;increased edge e ects; andincreased patch isolation

Decreasing Patch Size

Once a landscape has been ragmented, the size o the re-maining patches is a critical actor in determining the num-ber and type o species that can survive within them. For all

specieslarge or smallthat cannot or will not cross a orestedge or leave a patch, all requirements to complete their li ecycle must be met within the patch, rom nding ood tomates. This is especially important or species with complexli e cycles, each with distinct habitat requirements. For ex-ample, many amphibian species have an aquatic larval stageand an upland adult phase. Also, some species require large

1.2.3.


10/25

SYNTHESIS



1


areas o continuous habitat and cannot survive in small patch-esthey are re erred to as area-sensitive species. Furthermore,large patches typically support larger populations o a given

species and thereby bu er them against extinction, inbreeding depression, and genetic dri t .

Increased Edge E ects

One o the most obvious changes to a ragmented landscapeis the increase in edge environment. Edge environment s or eco-tones mark the transition between two di erent habitats. In anaturally orested landscape, edge is usually limited to a smallarea, such as along streams or landslides (Laurance and Bier-regaard, 1997). Natural edges are usually less abrupt than hu-man- ormed edges and show a gradual transition rom onehabitat type to another. In Rondonia, Brazil, de orestationpatterns show a herringbone pattern that closely ollows theroad that was cut through the original orest. Along agr icul-tural rontiers, the orig inal landscape may be ragmented intolong narrow strips or shreds, interspersed with areas o agri-culture (Feinsinger, 1997). These strips may separate di erentcrops, thus serving as windbreaks, or the boundary betweentwo landowners. As a result this remaining ragment is entirelymade up o edge environment. Residual trees along riversprovide another example o narrow, edge-dominated envi-ronments.

The extent o edge environment in a ragment patch is deter-mined in part by its shape. The ratio o the perimeter to area(or the amount o edge environment to the amount o inte-rior) is one measure o patch shape. A circular patch has themaximum area per unit edge and will have less edge environ-ment and ewer edge e ects than a rectangular patch o thesame size. Because edge e ects may extend 200 meters (and

sometimes more), small patches may be entirely composed o edge environment. For example, a new reserve is being cre-ated with an area o one square km. The reserve can either berectangular: Reserve A (2 km by 0.5 m), or square: Reserve B(1 km by 1 km). As illustrated, both have the same total areabut Reserve A will be composed entirely o edge environ-ment and its core size will be 0 square km, whereas Reserve

B will have a core area o 0.25 square km.

Edge E ects

Many studies have examined the e ects o edges on the phys-ical environment and biological communities that remaina ter ragmentation (Lovejoy et al., 1986; Bierregaard et al.,1992; Malcolm, 1994; Camargo and Kapos, 1995; Murcia,1995; Didham, 1997; Laurance et al., 1998; Carvalho and Vas-concelos, 1999). The longest running and perhaps the mostdetailed study o ragmentation e ects ever conducted is theBiological Dynamics o Forest Fragments project, which be-gan in 1979. This pioneering project, located in the Amazonregion north o Manaus, Brazil, has generated some o the

ndings described here and in ormed much o our generalunderstanding o the e ects o orest ragmentation. Edge e

ects is a general term used to describe a number o di erentimpacts, and can be categorized into several types: physical(e.g., microclimatic changes), direct biological (changes inspecies composition, abundance, and distribution), and indi-rect biological (changes in species interactions such as pre-dation, competition, pollination, and seed dispersal) (Matlackand Litvaitis, 1999). Moreover, many o the e ects o rag-mentation are synergistic; or example, ragmentation can leadto increased re risk, increased vulnerability to invasive spe-cies, or increased hunting pressure (Hobbs, 2001; Lauranceand Williamson, 2001; Peres, 2001).

Edge E ects - Physical Some o the most signi cant edge e ects are the microcli-matic changes that take place along a ragments edge (Harper et al., 2005). Edge areas in orests are typically warmer, moreexposed to light and wind, and drier than interior areas. Gra-dients o these microclimatic conditions extend into the in-

terior approximately 15 to 75 meters (Kapos, 1989; Lauranceand Bierregaard, 1997). Microclimatic changes along edgeso ten have secondary e ects, such as altering vegetation struc-ture and, eventually, plant and animal communities (Matlack,1993).

Increased wind along the edge o ragments physically dam-


11/25

SYNTHESIS



2


ages trees, causing stunted growth or tree alls (Essen, 1994;Laurance, 1994). This is especially obvious when a ragment

rst orms, since interior plant species are o ten not structur-

ally adapted to handle high wind stress. Furthermore, windtends to dry out the soil, decrease air humidity, and increasewater loss (evapotranspiration rates) rom lea sur aces, creat-ing a drier microclimate. This drier environment has a higher

re risk. Several studies have examined the increased risk o res in ragmented environments, particularly those that bor-

der grazing lands (Uhl and Bushbacher, 1985; Cochrane and

Invasion by Generalist SpeciesEdges are more susceptible to invasion by generalist or weedyspecies that are better adapted to handle disturbance and the

new microclimate. These species might be plants (such as lia-nas, vines, creepers, and exotic weeds), animals, or diseases.Simultaneously, long-lived interior canopy species, epiphytes,and other mature orest taxa decline in abundance (King andChapman, 1983). Wind along edges also increases the trans-

er o seeds rom outlying areas, thereby aiding invasion ooreign, generalist, or weedy species. Introduction o animals,

Bolivian road (Source: E. Sterling and K. Frey)

Schulze, 1999; Nep-stad et al., 1999; Co-chrane, 2001; Hobbs,2001).

Edge E ects - Biological The creation o edge

ollowing ragmenta-tion causes a number o biological changes(Harper et al., 2005).These changes areo ten similar or cou-pled to the biologi-cal changes that result

rom the creation o the ragment itsel .These include changes in species composition, abundance, anddistribution, as well as changes in species interactions such aspredation, competition, pollination, and seed dispersal. Alongthe edge o a ragment, biological changes may extend ar-ther than the physical ones. In one study, invasion by a distur-bance-adapted butterfy species extended nearly 250 metersinto the orest (Laurance et al., 2000). Here we examine three

biological changes particularly associated with the ormationo edge: invasion by generalist species, alteration o plant com-munities, and alteration o insect communities and nutrientcycling. Additional biological changes as a consequence o

ragmentation are detailed in subsequent sections: E ects onSpecies Abundance, Richness, and Density and InteractionsAmong Species and Ecological Processes.

loving species at the expense o slower-growing shade-lov-ing ones (Harper et al., 2005). Studies o orest ragments inthe Amazon noted a dramatic loss o plant biomass overall;although secondary vegetation (especially vines and lianas)proli erated, this new biomass did not compensate or the losso interior tree species (Laurance et al., 1997). Since manytree species have long li e spans, it is important to examine the

changes in plant communities over extended periods. It maytake hundreds o years or the ull consequences o ragmen-tation to be revealed.

Alteration o Insect Communities and Nutrient Cycling Only a ew studies have been conducted to date on the e ecto ragmentation on insect communities (Aizen and Feins-

adapted to disturbedenvironments and hu-man presence, such asdomestic cats, rats, andmice, is o ten a prob-lem along edges, as isdisease transmissionbetween wildli e anddomestic animals.

Alteration o Plant CommunitiesThe increased lightalong edges a ecboth the rate and typeo plant growth, avor-ing ast-growing light-


12/25

SYNTHESIS



3


inger, 1994; Gibbs and Stanton, 2001). Fragmentation, how-ever, appears to alter both the abundance and composition o insect communities, thus a ecting lea litter decomposition

and hence nutrient cycling (Didham, 1998).

Beetles (o the amilies, Carabidae, Staphylinidae, Scarabaei-dae) that are common to continuous interior orest disappear

rom orest ragments, a surprising result given their small sizeand generalist habitat requirements (Klein, 1989). Their disap-pearance may be the result o the drier microclimate or losso species they depend on (i.e., less mammal dung and allen

ruit on which to reproduce). Another possible reason or their disappearance is that these insects actually travel tremendousdistances in search o decaying material or their reproduc-tion and may not be able to cross the matrix between patches.Whatever the cause, there are a number o implications or ecosystem unction, including a decreased rate o nutrient cy-cling. Also, the incidence o disease may be elevated, as dung isle t on the ground longer, allowing fies to breed there.

Isolation-Barriers to Dispersal

The degree o isolation o a patch helps determine what bio-logical communities it can sustain. While patches may appear isolated, their actual biological connectivit y depends on thehabitat that separates them. In ragmented landscapes, patcheso high-quality habitat are typically interspersed with areaso poor habitat. In a very isolated patch, species that cannotdisperse may be unable to nd adequate resources or mates.They may become separated rom other populations and thusprone to genetic inbreeding and possibly local extinction.

Species Response to IsolationA species response to ragmentation depends on its disper-

sal ability as well as its perception o the environment. For example, species that fy (e.g., birds, bats, fying insects) aretypically less a ected by patch isolation than less mobile spe-cies (e.g. rogs and beetles). For some species, crossing an open

eld or two kilometers is not a problem. However, speciesthat spend most o their time in treetops (e.g., some specieso primates and marsupials) or in dark, interior orest may

never cross such a large opening. A species that disperses over long distances, such as an A rican elephant ( Loxodonta sp.)will perceive a particular landscape as more connected than a

species with short-range dispersal, such as a shrew (species o the amily Soricidae).

Species without the bene t o an aerial view o a landscapemake decisions primar ily based on the habitat directly in ronto them (Gibbs et al., 1998). A study in the Amazon conductedby Malcolm (1998) revealed distinct responses o similar ani-mals to ragmentation. Two species o opposumthe wooly(Didelphys lanigera) and the mouse ( Didelphys murina)weretracked using radio transmitters to determine i they wouldtravel a gap o 135 to 275 meters to reach the ragment on theother side. Mouse opposums were able and willing to crossthe gap, while the more strictly arboreal species, the woolyopposum, was not.

In the marine environment, responses to ragmentation aremore complex because the environment is three-dimensional,and many marine species are mobile or have a mobile larvalstage, and breed ar rom where they complete their adult li ecycle. These traits mean that marine species are less likely toexperience the kind o isolation that occurs in a ragmentedterrestrial system. The circumstances depend largely on theparticular marine system or species (e.g., ragmentation o mangroves mimics terrestrial ragmentation more closely thanthat o other marine systems) and the degree o ragmentation(small or large scale). Studies o larval dispersal that examinethe link between physical oceanography (e.g., currents) andreproductive li e cycles o marine species are shedding newlight on the level o connectivity o marine systems (Roberts,1997; Cowen et al., 2000; Taylor and Hellberg, 2003).

E ect o Time on IsolationFragmentation is a dynamic process, o ten with delayed e -

ects; knowing the amount o time a patch has been isolated iscritical to understanding the consequences o ragmentation.In long-lived species, such as trees, it may take a hundred yearsto observe the impact o ragmentation. Individual trees con-tinue to survive immediately ollowing ragmentation; how-


13/25

SYNTHESIS



4


Table 4: A comparison o matrix habitats rom a wildli e conservation perpective

Benefts

Allows Provides Provides Provides Protection FromMatrix Habitat Gene

Flow EcosystemServices

Wildli eHabitat

Climatic Extremes

ExoticSpecies

Fire TotalScore

Fully protected orest 4 4 4 4 4 4 24Low intensity selective logging 4 4 3 4 3 3 21Traditional orest management 3 4 3 4 3 3 20Medium-high intensity logging 3 3 3 3 3 1 16Low-diversity agro orestry 2 2 2 2 2 2 12Plantation Forests 1.5 3 1.5 3 2 3 14

Row crops 1 0 1 0 0 1 3Cattle pastures 1 1 1 0 0 0 3Note: Each habitat was scored by a panel o 15 researchers. The most avorable habitats received the highest number andthe least avorable received the lowest.Source: Modi ed rom Laurance et al., 1997

ever, they may no longer reproduce perhaps they are toospread apart to exchange pollen by wind, or their pollinatorsor seed dispersers have disappeared. In this case, it is only amatter o time be ore the population becomes locally extinct.

Janzen (1986) coined the term living dead to describe theates o species in such situations.

E ects o Di erent Types o Fragmentation

The e ects o ragmentation also vary depending on thecause o ragmentation ( or example, ragmentation or agr i-culture versus or logging). It is di cult to make generaliza-tions about the e ects o a speci c type o ragmentation on aparticular landscape, since the consequences may be very di -

erent in a temperate versus a tropical region or in a grasslandversus a orest, largely because the plants and animals presenthave di erent sensitivities to ragmentation.

Keeping these issues in mind, we can estimate the potentiale ects o a particular type o ragmentation based on how thenew environment is perceived by the original species present

and whether the change to the landscape is permanent or temporary. For example, selective logging is typically less dis-ruptive than clear-cutting orested areas. This is because a ter selective logging the orest is still relatively intact. While di -

ering rom the original orest, selectively-logged orest doesnot orm a large, unusable gap in habitat, as o ten occurs whena orested area is replaced with agricultural land (Table 4.).

The matrix that surrounds ragments has a large e ect onwhat species remain within the ragments and their dispersalability between ragments. Table 4 illustrates some o the ben-e ts provided by di erent matrix types as subjectively rankedby a panel o 15 researchers rom the Biological Dynamicso Forest Fragments project. The table displays a hierarchy

o matrix habitats rom most avorable to least avorable omany species.

E ects on Species Abundance, Richness, and Density

Fragmentations impact on species abundance, richness, anddensity is complex, and there is no clear rule what these e -


14/25

SYNTHESIS




ects may be. Studies o the e ects o ragmentation on speciesabundance, richness, or density relative to ragment size havehad inconsistent results (Debinski and Holt, 2000), some indi-

cating an increase in species, in others, a decline. However, it isimportant to keep in mind that simply counting the number o species does not measure impacts o ragmentation on be-havior, dispersal ability, or genetic diversity.

Some species respond positively to ragmentation (Brown andHutchings, 1997; Laurance and Bierregaard, 1997; Lynam,1997; Malcolm, 1997). Fragmentation may increase speciesrichness by allowing generalist species to invade. In a study o the impact o ragmentation on rogs in a lowland Amazonian

orest, species richness was strongly and positively related toragment area (Tocher et al., 1997). A ter ragmentation, spe-

cies richness increased largely as a result o invasion by rogspecies rom the surrounding matrix into the remaining orest

ragments. It is unclear i this increase will be sustained over time. For example, i this same spot were re-surveyed in 50

years, total species richness might decline as interior orestspecies disappear.

Immediately ollowing ragmentation, the density o individ-

uals may increase as animals crowd into the remaining orest(Schmiegelow et al., 1997; Collinge and Forman, 1998). Thisinfation o density will ultimately prove short-lived because

patches are rarely adequate to support the same populationdensity as more extensive habitats. This phenomenon under-scores the need to monitor ragmentation e ects over longtime scales.

Interactions Among Species and Ecological Processes

Fragmentation causes the loss o animal populations by aprocess termed aunal relaxation, the selective disappearanceo species and replacement by more common species (Dia-mond, 2001). Large-bodied vertebrates, especially those athigh trophic levels, are particularly susceptible to habitat lossand ragmentation, and are among the rst species to dis-appear. Thus, predators are o ten lost be ore their prey, andthose species that do survive on small ragments (usually her-bivores) tend to become ar more abundant than populationso the same species on larger species-rich ragments. There aretwo principal explanations or this increased abundance. The

rst is ecological release rom competition: when competing spe-cies are removed, the resources they utilized become available

to the persisting species. The secondis that prey escape predators thatnormally limit their abundance onlarger ragments. Lack o predatorsin small ragments can also lead toan overabundance o herbivores thattend to weed out palatable plant spe-cies and convert the landscape intoa orest o herbivore-proo plants.Furthermore, as large predators dis-appear, smaller predators o ten in-

crease; this is known as mesopredato release (Soul et al., 1988; Terborgh etal., 1997). For example, in Cali ornia,as coyotes disappear rom ragments,there is an overabundance o smaller predators, such as skunks, raccoons,grey ox, and cats (Saether, 1999).Cassava eld burning in Vietnam (Source: K. Frey)


15/25

SYNTHESIS




These smaller predators then prey on scrub-breeding birds.Fragmentation thus triggers distortions in ecological inter-actions that drive a process o species loss, the end point o

which is a greatly simpli ed ecological system lacking mucho the initial diversity (Terborgh et al., 1997; Terborgh et al.,2001).

While predator-prey relationships are o ten altered in rag-mented landscapes, it is not always possible to predict whatthe change will be. A number o review papers have exam-ined nest predation in ragmented landscapes; however, theresults have been inconsistent (Andren, 1994; Paton, 1994;Major and Kendal, 1996; Hartley and Hunter, 1998; Chal ounet al., 2002). Studies in Central Canada, or example, oundthat nests in orest patches adjacent to agricultural land hadincreased predation, while those next to logged areas did not(Bayne and Hobson, 1997, 1998). It appeared that the preda-tor community did not change in the logged areas, while or-est patches next to agricultural land had increased densitieso red squirrels that preyed on the nests. Other studies haveshown that songbirds are subject to increased predation alongedges, particularly in de orested areas. In other words, the typeo ragmentation and the habitat adjoining the ragment in-

fuences predator-prey relations: nest predation is less a ectedby a single road bisecting an area, but is greatly a ected alongedges o areas that have been de orested (Hartley and Hunter,

1998).

Overall a combination o landscape type and structure, preda-tor community, and level o parasitism are important in an-ticipating the outcomes o ragmentation. For example, unlikestudies in the Midwest and Northeast o the United States, astudy in the American West, where the landscape has histori-cally been patchy, ound that predation rates actually decreasedas human-caused ragmentation increased (Tewksbury et al.,1998). This study indicated that the type o predators in anarea, as well as the habitat structure, were key inputs to an-ticipate the impact o ragmentation on bird nest predationrates.

In addition, not all groups o species experience an increasein predation due to ragmentation. A recent analysis o the lit-erature ound that avian predators were more likely to bene t

rom ragmentation than mammalian predators (Chal oun etal., 2002). Another study surprisingly ound that turtle nestslocated along roads had lower predation rates than those lo-

Box 1. Corr idors and Connectivity

When existing protected areas are small, connecting them to other protected areas may increase their ability tosustain their auna and fora. Connectivity between protected areas is critical as ew o them are large enough tosustain species on their own (Hunter and Gibbs, 2006). Four basic species movements are important to consider to ensure landscape connectivity: daily, small-scale home range movements; annual seasonal migrations; dispersalo young rom their parents; and geographic range shi ts (Hunter, 1997). These di erent species movements as wellas the types o species ound in a particular landscape are all important actors when increasing connectivity or designing protected area networks. One way to increase connectivity is by creating wildli e corridors. Corridor s are

linear strips o land that allow species to move among di erent habitat types or breeding, birthing, eeding, roost-ing, annual migrations, dispersal o young animals away rom their parents, and as an escape path rom predatorsor disturbance. Riparian zones are good examples o corr idors that link orest patches. The value o corr idors hasbeen the center o considerable debate (Noss, 1987; Simberlo and Cox, 1987; Soul and Gilpin, 1991; Simberlo et al., 1992; Tewksbury et al., 2002). Part o this debate is due to the theoretical nature o the corridor concept.There are ew studies that show that animals actually use corridors, or that can separate between the e ect o thecorridor itsel rom that o the additional habitat provided by its creation.


16/25

SYNTHESIS




cated in edges or in orests (Hamilton et al., 2002).

An increase in invasive plants ollowing ragmentation may

indirectly enhance predator success on bird nests. Schmidtand Whelan (1999) ound that invasive plants o the genera

Loniceraand Rhamnus were not only pre erred nesting sites or American robins ( Turdus migratorius) and Wood thrushes ( Hy-locichla mustelina), but also acilitated predator access to nests.

The invasive plants lea out earlier, and so are requently cho-sen as nesting sites; the lack o thorns and lower nest height

Box 2. The Futi Corr idor Linking Tembe Elephant Park, South A rica to Maputo Elephant Reserve,Mozambique

Landscapes have naturally occurring borders that are not determined by political boundaries. Many political bor-ders are reely crossed by animals to access the resources they need or survival, while others, such as many interna-tional borders, not only appear on maps, but are bounded by ences or other obstacles that ragment landscapes andecosystems. These boundary markers may present an impenetrable barrier to species that can limit a populations

access to needed resources or prevent migration and movement through a landscape. In these situations removalo border obstacles and creation o designated corridors to acilitate animal movement has sometimes proven tobe a worthwhile solution.

The border between Mozambique and South A rica is an example o such a solution. A ence constructed alongthe border divided a population o elephants, the only indigenous population remaining on the coastal plain o Southern Mozambique and Kwa-Zulu Natal province in South A rica. These elephant traveled along the FutiCorridor (a seasonal river and marshland) that links Tembe Elephant Park in South A rica to Maputo ElephantReserve in Mozambique.

With the end o political unrest an opportunity arose to assess the need or the ence and the potential or reuni-ying the elephant population. On June 22, 2000, the governments o Mozambique, Swaziland, and South A rica

signed the Lubombo Trans rontier Trilateral Protocol, an agreement whose goal is to remove borders to supportconservation. Scientists at the Conservation Ecology Research Unit (CERU) at the University o Pretoria spentthree years tracking the movements o elephants along a section o the Mozambique/South A rica border (VanAarde and Fairall, 2002). Using satellite radio tracking, they ound that the populations still traveled the traditionalroutes they had used prior to installation o the ence. Examining the elephant populations movement patterns,and their impact on the landscape and interaction with humans, a series o recommendations to acilitate move-ment across the boundary while minimizing disruptions to the landscape and the human population were pro-posed. The recommendations included removing the border ences entirely, ormal designation o the corridoras a protected area in Mozambique, and speci c boundary parameters or the corridor. Plans are currently under-way to implement the recommendations and establish a conservation area that will cross the political boundaries(Peace Parks, 2003).

Cross border conservation solutions have been used more and more requently to acilitate conservation coopera-tion between countries around the world. Typically solutions like this are called Trans rontier Conservation Areas(TFCAs) or Transboundary Natural Resource Management solutions. These cross border e orts are instrumentalin reuni ying arti cially-divided landscapes and can acilitate development o coordinated conservation practices.Other bene ts include improved political relationships between countries, increased tourism opportunities, andthe involvement o local communities in cra ting conservation solutions that will provide direct local bene ts.[For more details see the module on Transboundary Protected Areas].


17/25

SYNTHESIS




o these shrubs in turn seems to aid predators in reaching the

nests.

Fragmentation can also take an indirect toll on plants whosepollinators or seed dispersers are orced to navigate an in-creasingly ragmented landscape in search o their host plants(Aizen and Feinsinger, 1994). In western Australia, only small,isolated populations o the cone-bearing shrub, Goods bank-sia (Banksia goodi ), remain, and many o these no longer repro-duce because their pollinators have disappeared (Buchmannand Nabhan, 1997).

Fragmentation o ten alters animal behavior, due to changesin the environment or predator activity. For example, Hobsonand Villard (1998) ound that one bird, the American Redstart(Setophaga ruticilla) acted more aggressively when con rontedwith a model o a nest parasitethe Brown-headed Cowbird

(Molothrus ater )in ragmented landscapes than in un rag-

mented ones. This appears to be because Cowbirds are morecommon in ragmented areas, and are thus a greater threat tothe Redstarts breeding success.

Management o Fragmented Landscapes

Increasingly, conservation pro essionals are aced with man-aging ragmented landscapes. This challenge is complicatedby the diverse responses o species to ragmentation and thecomplex decisions surrounding conservation o land. As with

any conservation management plan, when examining a rag-mented landscape, it is essential to identi y clear goals. For example, or a wide-ranging species, such as the black bear (Ursus americanus), habitat connectivity is critical, so it is im-portant to maintain a large un ragmented area; however, toconserve a rare species with speci c habitat needs, it may be

Box 3. Identi ying Species Vulnerable to Fragmentation

Knowing which species are most vulnerable is critical to understanding the impact o ragmentation. Behavioralpatterns, resource needs, reproductive biology, and natural history can be used to identi y species that are mostvulnerable to ragmentation. Below is a list o characteristics that are typical o species more vulnerable to rag-mentation (modi ed rom Laurance and Bierregaard, 1997):

Rare species with restricted distr ibutions (Andersen et al., 1997)Rare species with small populations (Andersen et al., 1997)Species with large home ranges (Soul et al., 1979; Newmark, 1987)Species that require heterogeneous landscapesSpecies that avoid matrix habitats (Warburton, 1997)Species with very specialized habitat requirementsSpecies with limited dispersal abilities (Laurance, 1990, 1991)Species with low ecundity (Sieving and Karr, 1997)Species with variable population sizes using patchy resourcesGround nesters vulnerable to medium-sized predators at edges (Bayne and Hobson 1997, 1998)Species vulnerable to hunting (Red ord and Robinson, 1987)Species that are arboreal (canopy dwellers)Co-evolved species (e.g., plants with speci c pollinators) (Gilbert, 1980)


18/25

SYNTHESIS




more important to preserve a speci c place than a large area,so that smaller ragments are more valuable than larger rag-ments (Dale et al., 1994; Laurance and Bierregaard, 1997).

While there are many di erent conservation strategies andoptions (Laurance and Gascon, 1997), here we will exploresome o the strategies specially aimed at ragmented land-scapes [For a detailed discussion o conservation managementstrategies see modules on Conservation Planning in and out-side Protected Areas].

Recommendations

The ollowing are important considerations to manage rag-mented landscapes (Laurance and Gascon, 1997; Me e andCarroll, 1997):

Conduct a landscape analysis to determine where the bigblocks o land suitable or protection exist and where po-tential connections among them lie.Evaluate the landscape and each patch in a regional context. I all surrounding landscapes are heavily ragmented and

your ocal landscape is not, its role in biodiversity con-servation is important at a regional level. Protection andconservation action should be elevated accordingly. Incontrast, i surrounding areas are largely un ragmented,

ragmentation issues in your ocal region may be less im-portant.Increase connectivity. Examine di erent planning options toavoid or reduce ragmentation. Can roads be re-routed, al-ternative land uses be ound, or protected areas be placedstrategically? [See Box 1. Corridors and Connectivity]Minimize edge e ects. Land managers o ten have some con-trol over which land uses will be adjacent to one another.Land management policies can be established to ensure

that a ragments size and shape maximizes the e ectivearea o protected land and reduces edge e ects. Adequatebu er zones (where land use is compatible with speciesneeds) around protected land also minimize edge e ects.Remember small ragments. They may not sustain jaguars or tapirs, but they still retain huge diversities o invertebrates,small vertebrates, plants, and perhaps rare or unique eco-

systems and species.Identi y species most vulnerable to ragmentation.It is impor-tant to identi y those species most likely to be impacted

by ragmentation and to consider them when designingmanagement and monitoring plans. [See Box 3. Identi y-ing Species Vulnerable to Fragmentation]

Terms o Use

Reproduction o this material is authorized by the recipientinstitution or non-pro t/non-commercial educational useand distribution to students enrolled in course work at theinstitution. Distribution may be made by photocopying or viathe institutions intranet restricted to enrolled students. Re-cipient agrees not to make commercial use, such as, withoutlimitation, in publications distributed by a commercial pub-lisher, without the prior express written consent o AMNH.

All reproduction or distribution must provide both ull cita-tion o the original work, and a copyright notice as ollows:

Laverty, M.F. and J.P. Gibbs. 2007. Ecosystem Loss and Frag-mentation. Synthesis. American Museum o Natural History,Lessons in Conservation. Available at http://ncep.amnh.org/linc.

Copyright 2006, by the authors o the material, with licenseor use granted to the Center or Biodiversity and Conserva-

tion o the American Museum o Natural History. All rightsreserved.

This material is based on work supported by the Nation-al Science Foundation under the Course, Curriculum andLaboratory Improvement program (NSF 0127506), and the

United States Fish and Wildli e Service (Grant AgreementNo. 98210-1-G017).

Any opinions, ndings and conclusions, or recommendationsexpressed in this material are those o the authors and donot necessarily refect the views o the American Museumo Natural History, the National Science Foundation, or the


19/25

SYNTHESIS



0


United States Fish and Wildli e Service.

Literature Cited

Achard, F., H.D. Eva, H-J. Stibig, P. Mayaux, J. Gallego, T.Richards, and J-P. Malingreau, 2002. Determination o de-

orestation rates o the worlds humid tropical orests. Sci-ence 297:999-1002.

Adams, J.M. and H. Faure. 1997. Preliminary vegetation mapso the world since the last glacial maximum: an aid to ar-chaeological understanding. Journal o Archaeologcal Sci-ence 24:623-647. See also Quaternary Environments Net-work (QEN) or online maps. Available rom: http://www.esd.ornl.gov/projects/qen/adams1.html (accessed August28, 2003).

Aizen, M.A. and P. Feinsinger. 1994. Forest ragmentation,pollination, and plant reproduction in Chaco dry orest,Argentina. Ecology 75:33051.

Andersen, M., A. Thornhill, and H. Koopowitz. 1997. Chap-ter 18: Tropical orest disruption and stochastic biodiversityloss. Pages138155 in W.F. Laurance, and R.O. Bierregaard

Jr., editors. Tropical Forest Fragments. University o Chi-cago Press, Chicago, Illinois, U.S.A.

Andren, H. 1994. E ects o habitat ragmentation on birdsand mammals in landscapes with di erent proportions o suitable habitats: a review. Oikos 71:355-366.

Atlay, G.L., P. Ketner, and P. Duvigneaud. 1979. Terrestrialprimary production and phytomass. Pages 120-181 in B.Bolin, E.T. Degens, S. Kempe, and P. Ketner, editors. TheGlobal Carbon Cycle. John Wiley and Sons, Chichester,U.K.

Bayne, E. M. and K.A. Hobson. 1997. Comparing the e -ects o landscape ragmentation by orestry and agricul-

ture on predation o arti cial nests. Conservation Biology

11(6):1418-1429.Bayne, E.M. and K.A. Hobson. 1998. The e ects o habitat

ragmentation by orestry and agriculture on the abun-dance o small mammals in the southern boreal mixedwood orest. Canadian Journal o Zoology 76:62-69.

Benke, A. C. 1990. A perspective on Americas vanishingstreams. Journal o the North American Benthological So-

ciety 9:77-88.Bierregaard, R.O., Jr., T.E. Lovejoy, V. Kapos, A.A. dos San-

tos, and R.W. Hutchings. 1992. The biological dynamics o

tropical rain orest ragments. BioScience 42:859-66.Bostrom, C., E. L. Jackson, and C. A. Simenstad. 2006. Sea-grass landscapes and their e ects on associated auna: A re-view. Estuarine Coastal and Shel Science 68: 383-403.

Brown, Jr., K.S. and R. W. Hutchings. 1997. Disturbance, rag-mentation, and the dynamics o diversity in Amazonian

orest butterfies. Pages 91-110 in W. F. Laurance and R. O.Bierregaard, editors. Tropical Forest Remnants. Universityo Chicago Press, Chicago, Illinois, U.S.A.

Bryant, D., D. Nielsen, and L. Tangley. 1997. Last rontier or-ests: Ecosystems and economies on the edge. World Re-sources Institute, Washington, D.C., U.S.A. Available rom:http://www.wri.org/ /l -eng/index.html (accessed Au-gust 28, 2003)

Buchmann, S.L. and G.P. Nabhan. 1997. The Forgotten Pol-linators. Island Press, Washington, D.C., U.S.A.

Camargo, J.L.C. and V. Kapos. 1995. Complex edge e ects onsoil moisture and microclimate in central Amazonian or-est. Journal o Tropical Ecology 11:205-11.

Carvalho, J.L.C. and V. Vasconcelos. 1999. Forest ragmenta-tion in central Amazonia and its e ects on litterdwellingants. Biological Conservation 91(23):151157.

Chal oun, A.D., F.R. Thompson III, and M.J. Ratnaswamy.2002. Nest predators and ragmentation: a review andmeta-analysis. Conservation Biology 16(2):306-318.

Clark, W.C. and J.T. Matthews. 1990. Page 140 in B.L. Turn-er, and W.C. Clark, editors. The Earth as Trans ormed byHuman Action. Cambridge University Press, Cambridge,U.K.

Cochrane, M. A. and M. D. Schulze. 1999. Fire as a recurrentevent in tropical orests o the eastern Amazon: e ects on

orest structure, biomass and species composition. Biotro-pica 15:2-16.

Cochrane, M.A. 2001. Synergistic interactions between habi-tat ragmentation and re in Evergreen Tropical Forests.Conservation Biology 15(6):1515-1521.

Collinge, S.K. and R.T.T. Forman. 1998. A conceptual mod-el o land conversion processes: predictions and evidence


20/25

SYNTHESIS



1


rom a microlandscape experiment with grassland insects.Oikos 82:6684.

Corsi, F., J. de Leeuw, and A. Skidmore. 2000. Pages 392-396

in Boitani, L. and T.K. Fuller, editors. Research Techniquesin Animal Ecology: Controversies and Consequences. Co-lumbia University Press, New York, New York, U.S.A.

Cowen, R.K., K.M.M. Lwiza, S. Sponaugle, C.B. Paris, andD.B. Olson. 2000. Connectivity o Marine Populations:Open or Closed? Science 287:857-859.

Dale, V.H., S.M. Pearson, H.L. O erman, and R.V. ONeill.1994. Relating patterns o land-use change to aunal bio-diversity in the Central Amazon. Conservation Biology4:1027-1036.

Davidson, I., R. Vanderkam, and M. Padilla. 1999. Reviewo wetland inventory in ormation in North America. InC.M. Finlayson and A.G. Spiers, editors. Global Reviewo Wetland Resources and Priorities or Wetland Inventory.CD-ROM, Supervising Scientist Report 144, Canberra,Australia. Available rom: http://www.wetlands.org/in-ventory&/GRoWI/welcome.html (accessed August 28,2003)

Debinski, D.M. and R.D. Holt. 2000. A survey and overviewo habitat ragmentation experiments. Conservation Biol-ogy 14(2): 342-356.

Diamond, J. 2001. ECOLOGY: Dammed Experiments! Sci-ence 294: 1847-1848.

Didham, R.K. 1998. Altered lea -litter decomposition rates intropical orest ragments. Oecologia 116:397-406.

Didham, R. K. 1997. An overview o invertebrate responsesto orest ragmentation. Pages 303-320 in A. D. Watt, N. E.Stork, and M. D. Hunter, editors. Forests and Insects. Chap-man & Hall, London, U.K.

Dynesius, M. and C. Nilsson. 1994. Fragmentation and fowregulation o river systems in the Northern third o the

world. Science 266:753-762.Essen, P.A. 1994. Tree mortality patterns a ter experimental

ragmentation o an old-growth coni er orest. BiologicalConservation 68:19-29.

Etter, A., C. McAlpine, S. Phinn, D. Pullar, and H. Possing-ham. 2006. Characterizing a tropical de orestation wave:a dynamic spatial analysis o a de orestation hotspot in the

Colombian Amazon. Global Change Biology 12: 1409-1420.

Eva, H.D., F. Achard, H-J. Stibig, and P. Mayaux. 2003. Reponse

to Comment on Determination o de orestation rates o the worlds humid tropical orests. Science 299:1015.Fahrig, L. 2003. E ects o habitat ragmentation on biodi-

versity. Annual Review o Ecology, Evolution, and Sys-tematics 34: 487-515.

Fearnside, P.M. and W.F. Laurance. 2003. Comment on De-termination o de orestation rates o the worlds humidtropical orests. Science 299:1015a.

Feinsinger, P. 1997. Habitat Shredding. Pages 270-272 in G.K.Me e and C.R. Carroll, editors. Principles o Conserva-tion Biology. Sinauer, Sunderland, Massachusetts, U.S.A.

Findlay, S.T, and J. Bourdages. 2000. Response time o wet-land biodiversity to road construction on adjacent lands.Conservation Biology 14(1): 8694.

Forman, R.T.T., and L.E. Alexander. 1998. Roads and their major ecological e ects. Annual Review o Ecology andSystematics 29: 207231.

Fukami, T., and D. A. Wardle. 2005. Long-term ecologicaldynamics: reciprocal insights rom natural and anthropo-genic gradients. Proceedings o the Royal Society B-Bio-logical Sciences 272 : 2105-2115.

Gibbs, J.P., M.L. Hunter, and E.J. Sterling. 1998. Problem-solving in Conservation Biology and Wildli e Manage-ment: Exercises or class, eld and laboratory. BlackwellScience, Malden, Massachusetts, U.S.A.

Gibbs, J.P. and E.J. Stanton. 2001. Habitat ragmentation andarthopod community change: carrion beetles, phoreticmites, and fies. Ecological Applications 11:79-85.

Gibbs, J.P. and W.G. Shriver. 2002. Estimating the E ects o Road Mortality on Turtle Populations. Conservation Biol-ogy 16(6):1647.

Gilbert, L.E. 1980. Food web organization and the conserva-tion o Neotropical diversity. Pages 11-33 in M.E. Souland B.A. Cox, editors. Conservation Biology: An Evolu-tionary Perspective. Sinauer Associates, Sunderland, Mas-sachusetts, U.S.A.

Haila, Y. 1999. Chapter 7: Islands and ragments. Hunter, M.,editor. Maintaining Biodiversity in Forest Ecosystems.


21/25


22/25

SYNTHESIS



3


mals in Amazonian orest ragments. Pages 207-221 in W.F.Laurance and R.O. Bierregaard, Jr., editors. Tropical For-est Remnants: Ecology, management, and conservation o

ragmented Communities. University o Chicago Press,Chicago, Illinois, U.S.A.Malcolm, J.R. 1998. Fragments o the orest: High roads to

oblivion. Natural History, American Museum o NaturalHistory, New York, July/August: 46-49.

Matthews, E., R. Payne, M. Rohweder, and S. Murray. 2000.Pilot analysis o global ecosystems(PAGE): Forest ecosys-tems. World Resources Institute (WRI), Washington, D.C.,U.S.A. Available rom: http:// orests.wri.org/pubs_de-scription.c m?PubID=3055 (accessed May 5, 2003).

Matlack, G.R. 1993. Microenvironment variation within andamong orest edge sites in the Eastern United States. Bio-logical Conservation 66:185-194.

Matlack, G.R. and J.A. Litvaitis. 1999. Chapter 6. Forest edges.Pages 210-233 in M. Hunter, editor. Maintaining Biodi-versity in Forest Ecosystems. Cambridge University Press,Cambridge, U.K.

Me e, G.K. and C.R. Carroll. 1997. Principles o Conserva-tion Biology, Second edition. Sinauer Associates, Sunder-land, Masschusetts, U.S.A.

Murcia, C. 1995. Edge e ects in ragmented orests:implica-tions or conservation. Trends in Ecology and Evolution19:58-62.

Nepstad, D.C., A. Verssimo, A. Alencar, C. Nobre, E. Lima, P.Le ebvre, P. Schlesinger, C. Potter, P. Moutinho, E. Men-doza, M.A. Cochrane, and V. Brooks. 1999. Large-scaleImpoverishment o Amazonian orests by logging and re.Nature 398:505-508.

Newing, H. 2001. Bushmeat hunting and management: im-plications o duiker ecology and interspeci c competition.Biodiversity and Conservation 10 (1):99-118.

Newmark, W.D. 1987. A land-br idge perspective on mamma-lian extinctions in western North American parks. Nature325: 430-32.

Noss, R.F. 1987. Corr idors in real landscapes: A reply to Sim-berlo and Cox. Conservation Biology 1:159-64.

Olson, J.S., J.A. Watts, and L.J. Allison. 1983. Carbon in LiveVegetation o Major World Ecosystems. Report ORNL-

5862. Oak Ridge National Laboratory, Tennessee, U.S.A.Note that the revised version rom 2001 is available rom:http://cdiac.esd.ornl.gov/epubs/ndp/ndp017/ndp017.

html (accessed August 27, 2003).Paton, W.C. 1994. The e ect o edge on avian nest success: howstrong is the evidence? Conservation Biology 8(1):17-26.

Parkinson, C.L. 1997. Earth rom Above: Using color-codedsatellite images to examine the global environment. Uni-versity Science Books, Sausalito, Cali ornia, U.S.A.

Peace Parks Foundation. 2003. Available rom: http://www.peaceparks.org/ (accessed September 30, 2003)

Peres, C.A. 2001. Synergistic e ects o subsistence huntingand habitat ragmentation on Amazonian orest vertebrates.Conservation Biology 15 (6): 1490-1505.

Pringle, C.M. 1997. Fragmentation in stream ecosystems.Pages 289-290 in G.K. Me e and C.R. Carroll, editors.Principles o Conservation Biology, second edition. Sinau-er, Sunderland, Massachusetts, U.S.A.

Red ord, K.H. and J.G. Robinson. 1987. The game o choice:patterns o Indian and colonist hunting in the Neotropics.American Anthropologist 89: 650-67.

Roberts, CM. 1997. Connectivity and management o Carib-bean coral ree s. Science 278:1454-1457.

Roper, J. and R.W. Roberts. 1999. De orestation: tropicalorest in decline. Canadian International Development

Agency (CIDA) Forestry Advisers Network. Availablerom: http://www.rc a-c an.org/english/issues.12.html

(accessed March 25, 2003).Rosenberg, K.V. and M.G. Raphael 1986. E ects o orest

ragmentation on vertebrates in Douglas r orests. Page32 in J. Verner, M.L. Morrison, and C.J. Ralph editors.Wildli e 2000: Modeling Habitat Relationships o Terres-trial Vertebrates, University o Wisconsin Press, Madison,Wisconsin, U.S.A.

Roxburgh, S. and I. Noble. 2001. Terrestrial Ecosystems. Page637 in S.A. Levin, editor. Encyclopedia o Biodiversity, Vol-ume 5. Academic Press, San Diego, Cali ornia, U.S.A.

Saether, B-E. 1999. Top dogs maintain diversity. Nature400:510 511.

Schmidt, K.A. and C.J. Whelan. 1999. E ects o exoticLonicera and Rhamnus on Songbird Nest predation. Con-


23/25

SYNTHESIS



4


servation Biology 13(6):1502-1506.Schmiegelow, F.K.A., C.S. Machtans, and S.J. Hannon. 1997.

Are boreal birds resilient to orest ragmentation? An ex-

perimental study o shortterm community responses.Ecology 78:19141932.Schle, W. 1990. Landscapes and climate on prehistory: In-

teractions o wildli e, man, and re. Pages 273-318 in J.G.Goldammer, editor. Fire in the tropical biota. Springer-Verlag, Berlin, Germany.

Sieving, K.E. and J.R. Karr. 1997. Avian extinction and per-sistence mechanisms in lowland Panama. Pages 156-170 inW.F. Laurance and R.O. Bierregaard, Jr., editors. TropicalForest Remnants: Ecology, management, and conservationo ragmented communities. University o Chicago Press,Chicago, Illinois, U.S.A.

Simberlo , D., 1986. The proximate causes o extinction. Pag-es 259-276 in D.M. Raup and D. Jablonski, editors. Pat-terns and Processes in the History o Li e. Springer-Verlag,Berlin, Germany.

Simberlo , D.S. and J. Cox. 1987. Consequences and costs o conservation corridors. Conservation Biology 1: 63-71.

Simberlo , D.S., J.A. Farr, J. Cox, and D.W. Mehlman. 1992.Movement corridors: Conservation bargains or poor in-vestments? Conservation Biology 6: 493-504.

Skole, D., and C. Tucker. 1993. Tropical de orestation and hab-itat ragmentation in the Amazon: Satellite data rom 1978to 1988. Science 260: 1905-09. Available rom: http://www.ciesin.org/docs/002-115/002-115.html (accessedSeptember 25, 2003).

Soul, M.E., B.A. Wilcox, and C. Holtby. 1979. Benign ne-glect: a model o aunal collapse in the game reserves o East A rica. Biological Conservation 15: 259-72.

Soul, M.E., D.T. Bolger, A.C. Alberts, J. Wright, M. Sorice,and S. Hill. 1988. Reconstructed dynamics o rapid extinc-

tions o chapparal-requiring birds in urban habitat islands.Conservation Biology 2(1):75-92.

Soul, M.E., and M.E. Gilpin. 1991. The theory o wildli ecorridor capability. Pages 3-8 in D.A. Saunders and R.J.Hobbs, editors. Nature Conservation 2: the role o cor-ridors, Surrey Beatty, Sydney, Australia.

Steininger, M.K., C.J. Tucker, P. Ersts, T.J. Killeen, Z. Ville-

gas, and S.B. Hecht. 2001. Clearance and ragmentation o tropical deciduous orest in the Tierras Bajas, Santa Cruz,Bolivia. Conservation Biology 15(4): 856-866.

Taylor, M.S. and M. E. Hellberg. 2003. Genetic evidence or local retention o pelagic larvae in a Caribbean ree sh.Science 299:107-109.

Terborgh J., L. Lopez, P. Nunez, M. Rao, G. Shahabuddin, G.Orihuela, M. Riveros, R. Ascanio, G.H. Adler, T.D. Lam-bert, and L. Balbas. 2001. Ecological Meltdown in Preda-tor-Free Forest Fragments. Science 294:1923-1926.

Terborgh, J., L. Lopez, J. Tello, D. Yu, and A.R. Bruni. 1997.Chapter 17: Transitory states in relaxing ecosystems o landbridge islands. Pages 256-279 in W.F. Laurance and R.O.Bierregaard, Jr., editors.1997. Tropical Forest Remnants.University o Chicago Press, Chicago, Illinois, U.S.A.

Tewksbury, J.J., S.J. Heil, and T.E. Martin. 1998. Breeding pro-ductivity does not decline with increasing ragmentationin a western landscape. Ecology 79:2890-2903.

Tewksbury, J.J., D.J. Levey, N.M. Haddad, S. Sargent, J.L. Or-rock, A. Weldon, B.J. Danielson, J. Brinkerho , E.I. Dam-schen, and P. Townsend. 2002. Corridors a ect plants,animals, and their interactions in ragmented landscapes.Proceedings o the National Academy o Science 99(20):12923-12926.

Tilman, D., J. Fargione, B. Wol , C. DAntonio, A. Dobson,R. Howarth, D. Schindler, W.H. Schlesinger, D. Simberlo ,and D. Swackhamer. 2001. Forecasting agriculturally driv-en global environmental change. Science 292:281-284.

Tocher, M.D., C. Gascon, and B.L. Zimmerman. 1997. Chap-ter 9: Fragmentation e ects on a Central Amazonian rogcommunity. Pages 124-137 in W.F. Laurance and R.O.Bierregaard, Jr., editors. Tropical Forest Remnants. Univer-sity o Chicago Press, Chicago, Illinois, U.S.A.

Turner, B.L. and W.C. Clark. 1990. The Earth as Trans ormed

by Human Action Cambridge University Press, Cam-bridge, U.K.

Uhl, C., and R. Bushbacher. 1985. A disturbing synergismbetween cattle-ranch burning practices and selective treeharvesting in the eastern Amazon. Biotropica 17:265-268.

Van Aarde, R.J. and N. Fairall. 2002. Restoration o theTembe-Futi-Maputo Coastal Plains elephant population.


24/25

SYNTHESIS




Annual report to US Fish & Wildli e Service 2001/2002.CERU Technical Report 12, University o Pretoria, SouthA rica.

Vitousek, O.M., H.A. Mooney, J. Lubchenco, and J.M. Mel-illo. 1997. Human domination o earths ecosystems. Sci-ence 277: 494-499.

Warburton, N.H. 1997. Chapter 13. Structure and conser-vation o orest avi auna in isolated remnants in tropicalAustralia. Pages 190-206 in W.F. Laurance and R.O. Bier-regaard, Jr., editors. Tropical Forest Remnants. Universityo Chicago Press, Chicago, Illinois, U.S.A.

Whit eld, J. 2003. Bushmeat: the law o the jungle. Nature421:8-9.

Whittaker, R.H. and E. Likens. 1975. The Biosphere and Man.In H. Lieth and R. H. Whittaker, editors. Primary Produc-tivity o the Biosphere. Ecological Studies 14. Springer-Verlag, Berlin, Germany.

Wilcox, B.A., and D.D. Murphy. 1985. Conservation strat-egy: The e ects o ragmentation on extinction. AmericanNaturalist 125:879-87.

White, R., S. Murray, and M. Rohweder. 2000. Pilot analy-sis o global ecosystems (PAGE): Grassland ecosystems.World Resources Institute, Washington, D.C., U.S.A.Available rom: http:// orests.wri.org/pubs_description.c m?PubID=3057 (accessed August 28, 2003).

Glossary

Anadromous sh : sh that return rom the sea to the riverswhere they were born to breed (e.g. salmon).

Area-sensitive species : species that require large areas o con-tinuous habitat and cannot survive in small patches.

Biome : a major biotic classi cation characterized by similar vegetation structure and climate, but not necessarily the samespecies.

Connectivity : the degree to which patches in a landscape arelinked.

Corridors : linear strips o protected land.

Ecological release rom competition : when competing species

are removed, the resources they utilized become available tothe persisting species.

Ecosystem: an assemblage o organisms and the physical envi-ronment in which it exchanges energy and matter.

Ecosystem loss : the disappearance o an assemblage o organ-isms and its physical environment such that it no longer unc-tions.

Edge environments or ecotones : the transition between twodi erent habitats.

Faunal relaxation : the selective disappearance o some speciesand replacement by more common species.

Fragmentation : the subdivision o a ormerly contiguouslandscape into smaller units.

Genetic dri t: a random change in allele requency in a smallbreeding population leading to a loss o genetic variation.

Habitat : there are two common usages o the term habitat.The rst de nes habitat as a species use o the environment,while the second de nes it as an attr ibute o the land and re-

ers more broadly to habitat or an assemblage o species. For adiscussion o di erent usages o habitat see Corsi et al., (2000)In this module we use habitat and habitat type to di erenti-ate between the two common usages o the term.Habitat degradation : the process where the quality o a spe-cies habitat declines.

Habitat loss : the modi cation o an organisms environmentto the extent that the qualities o the environment no longer support its survival.

Inbreeding depression : reduction in reproductive abilityand survival rates as a result o breeding among related


25/25

SYNTHESIS

individuals.

Lentic : relating to or living in still or slow-moving water.

Lotic: relating to or living in swi t-fowing water.

Matrix : the most common cover type in any given landscape.As it occupies the most area, it is the dominant eature o thelandscape and usually the most connected cover type.

Meso predator release : as large predators disappear, the popu-lation o smaller predators o ten increases.

Patch : usually de ned by its area, perimeter, shape, and com-position, such as a land cover type (such as water, orest, or grassland), a soil type, or other variable.

Potential extent : the extent o coverage o a particular biometype, assuming there were no humans and based on currentclimatic conditions.

Trophic level : stage in a ood chain or web leading rom pri-mary producers (lowest trophic level) through herbivores toprimary and secondary carnivores (highest trophic level).

Ecosystem Loss and Fragmentation

Documents

Transcript of Ecosystem Loss and Fragmentation