Freshwater Ecoregions of the World: A New Map of Biogeographic Units for Freshwater Biodiversity Conservation.

Articles

Growth of the human population, rising consumption,and rapid globalization have caused widespread de-

gradation and disruption of natural systems, especially inthe freshwater realm.Freshwater ecosystems have lost a greaterproportion of their species and habitat than ecosystems onland or in the oceans, and they face increasing threats fromdams, water withdrawals, pollution, invasive species, andoverharvesting (MEA 2005, Revenga et al. 2005). Freshwater

ecosystems and the diverse communities of species found inlakes, rivers, and wetlands may be the most endangered of all(MEA 2005).

These stressed systems support an extraordinarily highproportion of the world’s biodiversity. In terms of area, fresh-water ecosystems occupy only 0.8%of Earth’s surface, but theyare estimated to harbor at least 100,000 species, or nearly 6%of all described species (Dudgeon et al. 2006). Each year,

Robin Abell (e-mail: [email protected]), Michele L. Thieme, Rebecca Ng, Nikolai Sindorf, and EricWikramanayake are withWWF inWashington, DC. Carmen

Revenga, Mark Bryer (Bethesda), James Robertson, Eric Armijo (Bolivia), Jonathan V. Higgins (Chicago), Thomas J. Heibel, and Paulo Petry (Boston) are with the

Nature Conservancy, headquartered in Arlington, Virginia. Paulo Petry is also an associate in ichthyology at the Museum of Comparative Zoology at Harvard

University in Massachusetts. Maurice Kottelat is an independent consultant in Switzerland and an honorary research associate at the Raffles Museum of Biodiversity

Research at the National University of Singapore. Nina Bogutskaya and Alexander Naseka are senior researchers at the Zoological Institute of the Russian Academy

of Sciences in St. Petersburg. Brian Coad is a research scientist at the Canadian Museum of Nature in Ottawa. Nick Mandrak is a research scientist at the Great Lakes

Laboratory for Fisheries and Aquatic Sciences, Fisheries and Oceans Canada, Burlington, Canada. Salvador Contreras Balderas is professor emeritus of the

Universidad Autónoma de Nuevo Leon in Monterey, Mexico. William Bussing is professor emeritus at the Universidad de Costa Rica. Melanie L. J. Stiassny is the

Axelrod Research Curator of Ichthyology at the American Museum of Natural History and an adjunct professor at Columbia University in New York City. Paul

Skelton is managing director of the South African Institute for Aquatic Biodiversity and professor at Rhodes University in Grahamstown, South Africa. Gerald R.

Allen is a research associate at Western Australian Museum in Perth. Peter Unmack is a postdoctoral associate in the Department of Integrative Biology at Brigham

Young University in Utah. David Olson is director of science and stewardship at Irvine Ranch Conservancy in California. Hugo L. López is head of the vertebrate

zoology department at the Museo de La Plata, assistant professor in the Facultad de Ciencias Naturales y Museo, and researcher at CIC (Buenos Aires) in Argentina.

Roberto E. Reis is a professor at Católica do Rio Grande do Sul in Porto Alegre, Brazil. John G. Lundberg is chair and curator of ichthyology, and Mark H. Sabaj Pérez

is collection manager, at the Academy of Natural Sciences in Philadelphia. © 2008 American Institute of Biological Sciences.

Freshwater Ecoregions of theWorld: A New Map of Biogeo-graphic Units for FreshwaterBiodiversity Conservation

ROBIN ABELL, MICHELE L. THIEME, CARMEN REVENGA, MARK BRYER, MAURICE KOTTELAT, NINA BOGUTSKAYA,BRIAN COAD, NICK MANDRAK, SALVADOR CONTRERAS BALDERAS, WILLIAM BUSSING, MELANIE L. J. STIASSNY,PAUL SKELTON, GERALD R. ALLEN, PETER UNMACK, ALEXANDER NASEKA, REBECCA NG, NIKOLAI SINDORF,JAMES ROBERTSON, ERIC ARMIJO, JONATHAN V. HIGGINS, THOMAS J. HEIBEL, ERIC WIKRAMANAYAKE, DAVIDOLSON, HUGO L. LÓPEZ, ROBERTO E. REIS, JOHN G. LUNDBERG, MARK H. SABAJ PÉREZ, AND PAULO PETRY

We present a new map depicting the first global biogeographic regionalization of Earth’s freshwater systems. This map of freshwater ecoregions isbased on the distributions and compositions of freshwater fish species and incorporates major ecological and evolutionary patterns. Coveringvirtually all freshwater habitats on Earth, this ecoregion map, together with associated species data, is a useful tool for underpinning global andregional conservation planning efforts (particularly to identify outstanding and imperiled freshwater systems); for serving as a logical frameworkfor large-scale conservation strategies; and for providing a global-scale knowledge base for increasing freshwater biogeographic literacy. Preliminarydata for fish species compiled by ecoregion reveal some previously unrecognized areas of high biodiversity, highlighting the benefit of looking at theworld’s freshwaters through a new framework.

Keywords: freshwater, ecoregions, biogeography, fish, mapping

www.biosciencemag.org May 2008 / Vol. 58 No. 5 • BioScience 403

http://www.biosciencemag.org

new freshwater species are described. For South Americaalone, about 465 new freshwater fish species have been de-scribed in the last five years (Eschmeyer 2006), a figure thatcorresponds to a new species every four days. The presenceof species confined to small ranges is also unusually high infreshwater ecosystems; for example, 632 animal species havebeen recorded as endemic to Lake Tanganyika (Groombridgeand Jenkins 1998).

Despite this combination of extraordinary richness, highendemism, and exceptional threat, few broadscale conserva-tion planning efforts have targeted freshwater systems and theirdependent species. This relative inattention derives in partfrom an acute lack of comprehensive, synthesized data on thedistributions of freshwater species (Revenga and Kura 2003).Themost exhaustive recent global inventory of freshwater taxaacknowledges serious survey gaps and assigns species distri-butions only to the level of continent (Lévêque et al. 2005).Such inventories are valuable for highlighting research pri-orities and providing a global picture of how taxonomic di-versity compares across continents, but they have limitedutility for conservation planning efforts, for which the largestplanning unit is often the river basin or ecoregion.

A global freshwater regionalizationEcoregions are awidely recognized and applied geospatial unitfor conservation planning,developed to represent the patternsof environmental and ecological variables known to influencethe distribution of biodiversity features at broad scales (Groveset al. 2002). Building on the work of Dinerstein and col-leagues (1995),we define a freshwater ecoregion as a large areaencompassing one ormore freshwater systemswith a distinctassemblage of natural freshwater communities and species.The freshwater species, dynamics, and environmental con-ditionswithin a given ecoregion aremore similar to each otherthan to those of surrounding ecoregions, and together forma conservation unit. Ecoregion boundaries are not necessar-ily determined by the turnover of species ranges (McDonaldet al. 2005) but are intended to describe broad patterns ofspecies composition and associated ecological and evolu-tionary processes.

Ecoregion delineation benefits from the best available datadescribing species and systems ecology, but can proceed withimperfect information (Wikramanayake et al. 2002). Globalecoregion frameworks have already been developed for theterrestrial and, more recently, marine realms, both of whichare characterized by their own data limitations (Olson et al.2001, Spalding et al. 2007). In this article we demonstrate howthe ecoregion concept has been applied to freshwater systems,and present the first global map of freshwater ecoregions—a starting point for conservation planning anywhere on Earth.

Ecoregions have typically been delineated to representpatterns of potential vegetation (Olson et al. 2001) and haveat times been used to characterize regional differences inwater quality as well (Omernik 1987). Terrestrial ecoregionsare delineated largely on the basis of climate, physiography,and vegetation types, but different features are often domi-

nant in shaping the broadscale distributions of freshwaterspecies. As Tonn (1990) described, the species occurring ina given river reach, lake, spring, or wetland will be a functionof a hierarchy of continental-scale filters (including moun-tain building, speciation, and glaciation) that have defined largebiogeographic patterns; regional-scale filters (such as broadclimatic and physiographic patterns, and dispersal barrierssuch as regional catchments); and subregional and finer-scale habitat filters (e.g., distinct physiographic types andmacrohabitats) acting on the regional species pool. Freshwaterecoregions capture the patterns generated primarily by con-tinental- and regional-scale filters.

Of these filters, dispersal barriers in the form of catch-ment divides (also called watersheds) are distinctive to fresh-waters. Unlike terrestrial species or those with aerial orwind-dispersed life stages, obligate freshwater species—thoseconfined to the freshwater environment and unable to movevia land, air, or sea—generally cannot disperse from one un-connected catchment to another. Furthermore, all speciesdependent on freshwater systems,whether or not they are con-fined to the aquatic environment, are to some extent affectedby the hydrological and linked ecological processes of thecatchments where they live. As a result, catchments stronglyinfluence broad freshwater biogeographic patterns inmost re-gions. There are exceptions, however. Tectonic movementshave in some cases separated once-joined catchments, al-lowing for further speciation.Also,natural drainage evolutionover geological time includes river piracy, which severs con-nections and provides new interdrainage links that reconformsystems.The freshwater ecoregions of theworld presented herereflect both the hydrological underpinning of freshwater fishspecies distributions as well as historical shifts in landmassesand consequent evolutionary processes.

Ecoregion delineation and species list compilationNoglobal biogeographic framework for freshwater species wasavailable as the foundation for our map. The applicability ofWallace’s (1876) and Udvardy’s (1975) zoogeographic realmstomost freshwater taxa is unresolved (Berra 2001,Vinson andHawkins 2003), and these divisions are too large for conser-vation planning endeavors. Several examinations of globalfreshwater biogeography (e.g., Banarescu 1990) provided in-formation at somewhat finer scales but could not be clearlytranslated into seamless ecoregion delineations. Whereappropriate, we adapted previous continental efforts. ForNorthAmerica,Africa, and Madagascar,we updated region-alizations outlined in two previously published volumes(Abell et al. 2000, Thieme et al. 2005), but we excluded aprior delineation for Latin America and the Caribbean (Ol-son et al. 1998) because the approach differed markedly fromour current methodology, and data have improved substan-tially since its development (e.g., Reis et al. 2003).We exam-ined but chose to exclude the 25 European regions of Illies’simpressive Limnofauna Europaea (1978) because the ap-proach for delineating those regions differed considerably fromours: those regions were based on the distributions of 75

Articles

404 BioScience • May 2008 / Vol. 58 No. 5 www.biosciencemag.org


different taxonomic groups andwere drawnwithout referenceto catchments.Moreover, neither ecological nor evolutionaryprocesses figured in those delineations.A complete list of allreferences and experts consulted in the process of delineat-ing ecoregions is available online (www.feow.org).

We assembled our global map of freshwater ecoregionsusing the best available regional information describing fresh-water biogeography, defined broadly to include the influ-ences of phylogenetic history, palaeogeography, and ecology(Banarescu 1990).We restricted our analyses to informationdescribing freshwater fish species distributions, with a fewexceptions for extremely data-poor regions and inland seas,where some invertebrates and brackish-water fish wereconsidered, respectively. We focused on freshwater fish forseveral reasons. On a global scale, fish are the best-studiedobligate aquatic taxa. Detailed information exists for otherfreshwater taxa in regions likeNorthAmerica and Europe, butthe consideration of such groups in a global analysis wouldbe difficult, given the wide variation in available data (Balianet al. 2008). Freshwater dispersant fish species—those unableto cross saltwater barriers—are better zoogeographic indicatorsthan freshwater invertebrates, which can often disperse overland, survive in humid atmospheres outside water, or betransported between freshwaters (Banarescu 1990). Finally,the distributions of obligate aquatic invertebrate groups ingeneral respond to ecological processes at localized scalesthat are too small to be meaningful for ecoregion delineation(Wasson et al. 2002). Therefore, fish serve as proxies for thedistinctiveness of biotic assemblages.We recognize that analy-ses of other taxonomic groups would almost certainly revealdifferent patterns for some regions, and that our results arescale dependent (Paavola et al. 2006). Our near-exclusivefocus on fish is a departure from earlier continental eco-regionalization exercises (Abell et al. 2000, Thieme et al.2005), and we have updated the ecoregion delineationsaccordingly.

The available data for describing fish biogeography varywidely. In the United States, it is possible to map presence/absence data for all freshwater fish species to subbasins aver-aging about 2025 square kilometers (km2) in size (NatureServe2006). But for many of the world’s species, occurrence dataare limited to a small number of irregularly surveyed systems.Large parts of the massive Congo basin remain unsampled,for instance,withmost sampling occurring nearmajor townsand most taxonomic studies of the region dating from the1960s. Problems with taxonomy and species concepts ham-per broadscale analyses even where systems have been rea-sonably well sampled (Lundberg et al. 2000). Althoughaddressingmany of these problems is beyond the scope of thisproject, in our analyses we have attempted to minimizenomenclatural errors by normalizing species names withEschmeyer’s Catalog of Fishes (2006; www.calacademy.org/research/ichthyology/catalog/).

Freshwater fish patterns were analyzed separately fordifferent regions of the world to account for data variability.The geographic scope of major information sources largelydefined those regions (table 1). Information sources weretypically taxonomic works, some of which included bio-geographical analyses. Leading ichthyologists delineatedecoregions primarily by examining the distributions of en-demic species, genera, and families against the backdrop ofan area’s dominant habitat features and the presence of eco-logical (e.g., large concentrations of long-distance migratoryspecies) and evolutionary (e.g., species flocks) phenomena.More than 130 ichthyologists and freshwater biogeographerscontributed to the global map by either delineating or re-viewing ecoregions.

Data gaps and biogeographic drivers resulted in the use ofslightly different criteria among and evenwithin some regions(table 2, box 1). Where fish species data were reasonablycomprehensive and available at subbasin or finer scales, weattributed species distributions to catchments to facilitateevaluation of biogeographic patterns in a bottom-up

Articles


Table 1. Regional information sources used for ecoregion delineations.

Region Primary information source

Africa Roberts 1975, Skelton 1994, Lévêque 1997, Thieme et al. 2005Middle East No regional information sources available.Former USSR No regional information sources available.Remainder of Eurasia For Europe: Kottelat and Freyhof 2007; no regionwide information sources for Asia.Australasia McDowall 1990, Allen 1991, Unmack 2001, Allen et al. 2002Oceania Keith et al. 2002Canada Scott and Crossman 1998United States Maxwell et al. 1995, Abell et al. 2000Mexico Contreras-Balderas 2000, Miller et al. 2005Central America Bussing 1976, CLOFFSCA (Reis et al. 2003)Caribbean Rauchenberger 1988, Burgess and Franz 1989South America CLOFFSCA (Reis et al. 2003), Menni 2003

Note: In many cases, these same sources were used to compile species lists. A full bibliography with additional publications,which along with unpublished data often constituted the greater part of inputs to ecoregion delineations and species lists, isavailable at the Web site www.feow.org. Every region also benefited from expert input; individual contributors are listed in theacknowledgments section and at the Web site. Regions in some cases correspond to politically rather than biophysicallydefined units to take advantage of existing information sources and expertise.


approach. For example, a new high-resolution hydrographicdataset (HydroSHEDS;www.wwfus.org/freshwater/hydrosheds.cfm) for SouthAmerica provided fine-scale catchment mapsthat, in conjunctionwith newly synthesized species data (Reiset al. 2003), aided in the assessment of biogeography. Inregions without extensive species data, or where major basinssupport highly similar faunas as a result of recent glaciation,a top-down analysis used qualitative expert knowledge ofdistinctive species and assemblages to map major bio-geographic patterns (table 2). Ecoregional boundaries result-ing from either approach, therefore, largely coincide withcatchment boundaries.

Whereas overall there is correspondence between catch-ments and ecoregion boundaries, unconnected neighboringcatchments were in some cases grouped together, wherestrong biogeographic evidence indicates that landscape orother features overrode contemporary hydrographic integrity.For example, owing to historic drainage evolution andsimilarities in fauna, Africa’s southern temperate highveldcombines headwaters of coastal basins that drain to theIndian Ocean with those of the Atlantic-draining Orangebasin. Considerable faunal exchange of the headwaters ofthe Orange River system with that of the coastal systems mayhave occurred as the coastal rivers eroded their basins at a fasterrate than the adjacentOrange tributaries (Skelton et al. 1995).

These and other examples demonstrate that historical geo-graphic events and current hydrology may have conflictingeffects on the fish fauna of a particular region and therebyargue for different boundaries. The decision to weigh someeffects more strongly than others was made on a case-by-casebasis, and it is acknowledged that additional data may favoralternative delineations.

With the exception of islands, individual freshwater eco-regions typically cover tens of thousands to hundreds ofthousands of square kilometers (Maxwell et al. 1995). Eco-region size varies in large part because of landscape history.Regionswith depauperate faunas resulting from recent glacia-tion events tend to have large ecoregion sizes, as do those dom-inated by very large river systems (e.g., much of SouthAmerica). Regions with recent tectonic activity or smaller,more isolated freshwater systems often are divided into smallerecoregions. For example, central Mexico has experiencedintermittent isolation and exchange between basins owing toactive mountain-building processes leading to small, frag-mented systems with distinct faunas. We acknowledge thatdata quality may also influence the size of ecoregions; for in-stance, the entire Amazon is currently divided into only 13ecoregions, but better data on species occurrences withinmajor subbasins wouldmost likely support finer delineations.

Articles


Table 2. Basic ecoregion delineation approaches for individual regions.

Region Delineation approach

Africa Using Roberts (1975) as a starting point, ecoregions were delineated using a top-down qualitative assessment thatincorporated expert knowledge and divisions of major river basins. In a few cases where basin divides do not circumscribespecies distributions or where basins contain internal barriers to dispersal, ecoregions straddle or divide basins.

Middle East Species lists were generated for whole drainage basins, which were then either combined with smaller catchments thatwere very similar faunistically (minor desert basins, for example) or subdivided on the basis of different ecologies (e.g., theTigris-Euphrates with lowland marshes and upland streams).

Former USSR A species/genera/family presence/absence matrix was compiled for a hierarchy of hydrographic units, and cluster analysisand ordination techniques (Primer v.6 statistics software) were employed to assess biotic similarities among hydrographicunits and to identify major faunal breaks.

Remainder of Eurasia For Southeast Asia and southern Europe, a bottom-up approach employing both published and unpublished field data andexpert assessment was used. East Asian, northern European, and eastern European ecoregions were delineated through atop-down process using major basins as a starting point and incorporating traditionally recognized zoogeographic patternswhere appropriate.

Australasia For Australia, ecoregions were adapted from Allen and colleagues’ (2002) and Unmack’s (1991) “freshwater fish biogeo-graphic provinces”; provinces were derived through similarity analyses, parsimony analysis, and drainage-based plots ofspecies ranges. For New Guinea, “subprovinces” of Allen (1991) were modified (primarily combined) on the basis of expertinput. For New Zealand and other islands and island groups, islands were placed in ecoregions on the basis of expertinput.

Oceania Islands and island groups were placed in ecoregions on the basis of distinctive (endemic or near-endemic) fish faunas.

Canada Separate cluster analyses were conducted on fish occurrence in the secondary watersheds in each of the nine primarywatersheds in Canada.

United States The “subregions” of Maxwell and colleagues (1995) were adopted, with relatively small modifications made following inputby regional specialists, especially the Endangered Species Committee of the American Fisheries Society.

Mexico Ecoregion delineations were based on qualitative similarity/dissimilarity assessments of major basins, using the standardadministrative hydrographical regions of the Mexican federal government. Subregions within major basins were recognizedas separate ecoregions when the fish fauna was sufficiently distinctive.

Central America Fish provinces from Bussing (1976) were revised and subdivided on the basis of the application of the similarity index tosubbasin fish presence/absence data.

Caribbean Ecoregions from Olson and colleagues (1998) were modified on the basis of similarity analyses of island-by-island specieslists and expert input.

South America Ecoregion delineations were based on qualitative similarity/dissimilarity assessments of catchments, resulting in aggrega-tion/disaggregation. See box 1 for additional information.

Note: Some of the variations resulting from differences in data quality and biogeographic drivers across and within regions are noted. For some regions,subecoregions (described at www.feow.org) were delineated to capture finer-scale patterns than could be represented by ecoregions.


The process of delineating ecoregions required compilingand synthesizing information on the distributions of fishspecies. A logical and practical extension of the delineationswas the compilation of fish species lists for each ecoregion.For theUnited States,NatureServe provided presence/absencedata for individual species, coded to eight-digit hydrologic unitcodes (HUCs); these HUC occurrences were then translatedto ecoregions, and the data were manually cleaned of erro-neous occurrences derived from introductions and prob-lematic records. These species lists were then merged withthose from Canada and Mexico for transnational ecoregions.For all other ecoregions, data came from the published liter-ature, as well as from gray literature and unpublished sources(see table 1; a full bibliography is available atwww.feow.org).In all cases, experts served as gatekeepers of these data to en-sure that lists were based on the best available information,both in terms of distributions and nomenclature. Introducedspecies were removed from the tallies presented here, as were

undescribed species. Confirmed extinct species (Ian J. Har-rison,American Museum of Natural History,NewYork, per-sonal communication, 29 March 2007) were excluded, butextirpated species were included to acknowledge restorationopportunities. Endemic species, defined as those occurringonly in a single ecoregion,were identified first by experts andcross-checked using a species database constructed for thisproject, which includes more than 14,500 described fishspecies. Species were coded as freshwater, brackish, or marineusing data from FishBase (www.fishbase.org), and specieswith only brackish or marine designations were omittedfrom the richness and endemism totals reported here.

Freshwater ecoregional map and species resultsOurmapof freshwater ecoregions contains 426 units, coveringnearly all nonmarine parts of the globe, exclusive of Antarc-tica, Greenland, and some small islands (figure 1; a full leg-end is available at www.feow.org). There is large variation inthe area of individual ecoregions. Large ecoregions, such asthe dry Sahel (4,539,429 km2), tend to be found in more de-pauperate desert and polar regions exhibiting low speciesturnover. Smaller ecoregions are typically found in noncon-tinental settings where systems are by nature smaller andspecies turnover is higher, as in the Indo-Malay region. Thesmallest ecoregion, at 23 km2, is Cocos Island (Costa Rica);the average ecoregion size is 311,605 km2. Ecoregions rangedfrom those encompassing only 1 country to those straddling16 countries (central and western Europe ecoregion).

In total,we assignedmore than 13,400 described freshwaterfish species to ecoregions, of which more than 6900 wereassigned to single ecoregions (i.e., endemic). Examination ofthe fish species data synthesized by ecoregion confirms somewell-knownpatterns and highlights others unknown tomanyconservationists, managers, and policymakers working atregional or global scales (figures 2a–2d). In agreement withprevious global assessments (Groombridge and Jenkins 1998,Revenga et al. 1998), our analysis identifies as outstanding forboth fish richness and endemism systems that include largeportions of Africa’s Congo basin, the southernGulf of Guineadrainages, and LakesMalawi,Tanganyika, andVictoria;Asia’sZhu Jiang (Pearl River) basin and neighboring systems; andlarge portions of South America’s Amazon and Orinocobasins. Areas confirmed for globally high richness includeAsia’s Brahmaputra,Ganges, andYangtze basins, aswell as largeportions of the Mekong, Chao Phraya, and Sitang and Ir-rawaddy; Africa’s lower Guinea; and SouthAmerica’s Paranáand Orinoco. When richness is adjusted for ecoregion area,additional systems such as the Tennessee, Cumberland,Mo-bile Bay,Apalachicola, andOzark highlands in the southeasternUnited States; portions of Africa’s Niger River Basin; theislands of New Caledonia,Vanuatu, and Fiji; China’s HainanIsland; and large parts of Sumatra and Borneo, among manyother areas, are also especially noteworthy.

Numerous systems previously identified as highly endemicfor fishwere confirmed, asmeasured by either numbers of en-demic species or percentage endemism. A subset includes

Articles


The delineation process for South America followed a step-wise process of subdivision of the continent’s major drainagesystems. Delineation started with the historically recognizedmajor ichthyographic provinces exemplified in Gery (1969)and Ringuelet (1975) and proceeded with subdivision at finerscales using regionalized data on fish distributions.

The criteria for determining the merit of delineating an eco-region were not uniform across the continent as a result oflocalized faunistic differences. In some areas, delineationswere based on family-level data, whereas in others, faunisticturnover at lower taxonomic levels was the criterion. Forinstance, astroblepid catfishes are distinct components of high-elevation freshwaters along the Andes forefront, and that fami-ly’s distribution was critical to informing the delineation of thehigh Andean ecoregions. On the other side of the continentalong the Atlantic coast, we used the presence or absence ofendemic assemblages of the genus Trichomycterus, several gen-era of the subfamily Neoplecostomatinae, and the presence orabsence of annual killifish genera and species to distinguishdistinct drainage complexes from one another.

In the piedmont zones and in contact areas between lowlandsand geologic shield areas, we used indicator groups to deter-mine where along the elevation/slope gradient the fauna waschanging. The distribution of lowland forms was matchedwith forms found in higher-gradient systems to establishwhere one group was dropping out and the other startedoccurring. This transition zone was then established as theoperational boundary between connecting ecoregions.

For areas like Patagonia, the Titicaca altiplano, and the Mara-caibo basin, the uniqueness of the fauna, often occurring with-in clearly defined geographic areas, permitted reasonablystraightforward delineations. In the larger river basin systemswhere there are no clear boundaries, the ecoregional limits arethe best approximation, given the current data.

Box 1. Example of criteria applied to ecoregiondelineation: South America.


Articles


Nor

thAm

eric

a101

Alas

kanCoa

stal

102

Upp

erYu

kon

103

Alas

kaan

dCan

adaPa

cific

Coa

stal

104

Upp

erMackenzie

105

Lower

Mackenzie

106

Cen

tral

Arctic

Coa

stal

107

Upp

erSas

katche

wan

108

MiddleSas

katche

wan

109

English–

Winnipe

gLa

kes

110

Sou

thernHud

sonBay

111

Wes

tern

Hud

sonBay

112

Can

adianArctic

Archipelag

o113

EasternHud

sonBay–U

ngava

114

Gulfof

St.LawrenceCoa

stal

Drainag

es115

Can

adianAtlantic

Island

s116

Lauren

tianGreat

Lakes

117

St.Lawrence

118

Northea

stUSan

dSou

thea

stCan

ada

Atlantic

Drainag

es119

Scotia

–Fun

dy120

Colum

biaGlaciated

121

Colum

biaUng

laciated

122

Upp

erSna

ke

123

Orego

nan

dNorthernCalifo

rnia

Coa

stal

124

Orego

nLa

kes

125

Sacramen

to–S

anJoaq

uin

126

Laho

ntan

127

Bon

neville

128

Dea

thVa

lley

129

Vega

s–Virgin

130

Colorad

o131

Gila

132

Upp

erRio

Grand

e–Bravo

133

Pecos

134

Rio

Con

chos

135

Lower

Rio

Grand

e–Bravo

136

Cua

troCiene

gas

137

Rio

Salad

o138

Rio

San

Juan

(Mexico)

139

Wes

tTexasGulf

140

East

TexasGulf

141

Sab

ine–

Galveston

142

Upp

erMisso

uri

143

MiddleMisso

uri

144

USSou

thernPlains

145

Oua

chita

Highlan

ds146

Cen

tral

Prairie

147

Ozark

Highlan

ds148

Upp

erMississ

ippi

149

Lower

Mississ

ippi

150

Teays–Old

Ohio

151

Cum

berla

nd152

Tenn

esse

e153

Mob

ileBay

154

Wes

tFloridaGulf

155

Apalachicola

156

FloridaPe

nins

ula

157

Appa

lachianPied

mon

t158

Che

sape

akeBay

159

Sou

thernCalifo

rnia

Coa

stal–B

aja

Califo

rnia

160

Son

ora

161

Guzman

–Sam

alayuca

162

Sinaloa

163

Mayran–

Vies

ca164

Rio

San

tiago

165

Lerm

a–Cha

pala

166

Llan

osEl

Salad

o167

Panu

co168

Ameca–

Man

antla

n169

Rio

Balsa

s

170

SierraMad

rede

lSur

171

Papa

loap

an172

Coa

tzacoa

lcos

173

Grijalva–U

sumacinta

174

Upp

erUsu

macinta

175

Yucatan

176

Bermud

a

Cent

ralA

mer

ica

201

Chiap

as–Fon

seca

202

Quintan

aRoo

–Motag

ua203

Mos

quitia

204

Estero

Rea

l–Tempisq

ue205

San

Juan

(Nicarag

uaan

dCos

taRica)

206

Chiriq

ui207

Isthmus

Caribbe

an208

San

taMaria

209

Cha

gres

210

Rio

Tuira

211

Cub

a–Cayman

Island

s212

Bah

amaArchipelag

o213

Jamaica

214

Hispa

niola

215

Puerto

Rico–

Virgin

Island

s

216

Windw

ardan

dLe

ewardIsland

s217

Cocos

Island

(Cos

taRica)

Sout

hAm

eric

a301

North

Ande

anPa

cific

Slope

s–Rio

Atrato

302

Mag

dalena

–Sinu

303

Maracaibo

304

Sou

thAm

ericaCaribbe

anDrainag

es–Trin

idad

305

Orin

ocoHighAn

des

306

Orin

ocoPied

mon

t307

Orin

ocoLlan

os308

Orin

ocoGuian

aShield

309

Orin

ocoDelta

andCoa

stal

Drainag

es310

Esse

quibo

311

Guian

as312

Amazon

asHighAn

des

313

Wes

tern

Amazon

Pied

mon

t314

Rio

Neg

ro315

Amazon

asGuian

aShield

316

Amazon

asLo

wland

s317

Ucayali–Uruba

mba

Pied

mon

t318

Mam

ore–

Mad

rede

DiosPied

mon

t319

Gua

pore–Itene

z

Figu

re1.Map

offreshw

ater

ecoregionsof

theworld,inwhich

426ecoregionsaredelineated.Aninteractiveversionof

thismap

that

includesad

dition

alinform

ationis

availableat

ww

w.f

eow

.org.


Articles


320

Tapa

jos–Juruen

a321

Mad

eira

BrazilianShield

322

Xing

u323

Amazon

asEs

tuaryan

dCoa

stal

Drainag

es324

Tocantins–Arag

uaia

325

Parnaiba

326

Northea

sternCaa

tinga

andCoa

stal

Drainag

es327

S.Fran

cisco

328

Northea

sternMataAtlantica

329

Paraibado

Sul

330

Ribeira

deIgua

pe331

Sou

thea

sternMataAtlantica

332

Lower

Urugu

ay333

Upp

erUrugu

ay334

Lagu

nado

sPa

tos

335

Tram

anda

i–Mam

pituba

336

Cen

tral

Ande

anPa

cific

Slope

s337

Titic

aca

338

Atacam

a339

Mar

Chiqu

ita–S

alinas

Grand

es340

Cuyan

–Des

agua

dero

341

Sou

thAn

dean

Pacific

Slope

s342

Cha

co343

Paragu

ay344

Upp

erPa

rana

345

Lower

Parana

346

Igua

ssu

347

Bon

aerens

eanDrainag

es348

Patago

nia

349

ValdivianLa

kes

350

Galap

agos

Island

s351

Juan

Fernan

dezIsland

352

Flum

inen

se

Euro

pean

dM

iddl

eEa

st401

Icelan

d–JanMayen

402

NorthernBritishIsles

403

Can

tabric

Coa

st–Lan

gued

oc404

Cen

tral

andWes

tern

Europe

405

Norweg

ianSea

Drainag

es406

NorthernBaltic

Drainag

es407

Baren

tsSea

Drainag

es408

Sou

thernBaltic

Lowland

s409

Lake

One

ga–LakeLa

doga

410

Volga–

Ural

411

Wes

tern

Cas

pian

Drainag

es412

Wes

tern

Iberia

413

Sou

thernIberia

414

EasternIberia

415

Gulfof

Venice

Drainag

es416

ItalianPe

nins

ulaan

dIsland

s417

Upp

erDan

ube

418

Dnies

ter–Lo

wer

Dan

ube

419

Dalmatia

420

Sou

thea

stAd

riatic

Drainag

es421

Ionian

Drainag

es422

Vardar

423

Thrace

424

Aege

anDrainag

es425

Dniep

er–S

outh

Bug

426

Crim

eaPe

nins

ula

427

Don

428

Kub

an429

Wes

tern

Anatolia

430

NorthernAn

atolia

431

Cen

tral

Anatolia

432

Sou

thernAn

atolia

433

Wes

tern

Tran

scau

casia

434

Kura–

Sou

thCas

pian

Drainag

es435

Sinai

436

Coa

stal

Levant

437

Orontes

438

Jordan

River

439

Sou

thwes

tern

Arab

ianCoa

st440

Arab

ianInterio

r441

Lower

Tigris

andEu

phrates

442

Upp

erTigris

andEu

phrates

443

Oman

Mou

ntains

444

Lake

Van

445

Orumiyeh

446

Cas

pian

Highlan

ds

447

Nam

ak448

Kaviran

dLu

tDes

erts

449

Esfaha

n450

TuranPlain

451

NorthernHormuz

Drainag

es452

Cas

pian

Marine

453

VolgaDelta–N

orthernCas

pian

Drainag

es

Afric

aan

dM

adag

asca

r501

Atlantic

Northwes

tAfric

a502

Med

iterran

eanNorthwes

tAfric

a503

Sah

ara

504

Dry

Sah

el505

Lower

Niger–B

enue

506

Niger

Delta

507

Upp

erNiger

508

Inne

rNiger

Delta

509

Sen

egal–G

ambia

510

Fouta–

Djalon

511

NorthernUpp

erGuine

a512

Sou

thernUpp

erGuine

a513

Mou

ntNim

ba514

Eburne

o515

Asha

nti

516

Volta

517

Bight

Drainag

es518

NorthernGulfof

Guine

aDrainag

es–B

ioko

519

Wes

tern

Equa

toria

lCraterLa

kes

520

Lake

Cha

d521

Lake

Victoria

Bas

in522

Upp

erNile

523

Lower

Nile

524

Nile

Delta

525

Ethiop

ianHighlan

ds526

Lake

Tana

527

Wes

tern

Red

Sea

Drainag

es528

NorthernEa

sternRift

529

Hornof

Afric

a530

Lake

Turkan

a531

She

belle–Jub

a532

Ogo

oue–

Nyang

a–Kou

ilou–

Niari

533

Sou

thernGulfof

Guine

aDrainag

es534

San

gha

535

Sud

anic

Con

go–O

uban

gi536

Uele

537

Cuvette

Cen

trale

538

Tumba

539

Upp

erCon

goRap

ids

540

Upp

erCon

go541

Albe

rtineHighlan

ds542

Lake

Tang

anyika

543

Malag

aras

i–Moyow

osi

544

Ban

gweu

lu–M

weru

545

Upp

erLu

alab

a546

Kas

ai547

Mai

Ndo

mbe

548

Maleb

oPo

ol549

Lower

Con

goRap

ids

550

Lower

Con

go551

Cua

nza

552

Nam

ib553

Etos

ha554

Karstveld

SinkHoles

555

Zambe

zian

Hea

dwaters

556

Upp

erZa

mbe

ziFloo

dplains

557

Kafue

558

MiddleZa

mbe

zi–Lua

ngwa

559

Lake

Malaw

i560

Zambe

zian

Highveld

561

Lower

Zambe

zi562

Mulan

je563

EasternZimba

bweHighlan

ds564

Coa

stal

East

Afric

a565

Lake

Ruk

wa

566

Sou

thernEa

sternRift

567

Tana

,Athi,an

dCoa

stal

Drainag

es568

Pang

ani

569

Okavang

o570

Kalah

ari

571

Sou

thernKalah

ari

572

Wes

tern

Orang

e573

Karoo

574

Drakens

berg–M

alotiH

ighlan

ds

575

Sou

thernTempe

rate

Highveld

576

Zambe

zian

Lowveld

577

Amatolo–

WinterbergHighlan

ds578

Cap

eFold

579

Wes

tern

Mad

agas

car

580

Northwes

tern

Mad

agas

car

581

Mad

agas

carEa

sternHighlan

ds582

Sou

thernMad

agas

car

583

Mad

agas

carEa

sternLo

wland

s584

Com

oros

–Mayotte

585

Seychelles

586

Mas

carene

s587

S.Tomean

dPrincipe

–Ann

obon

Nor

ther

nAs

ia601

Irgyz–Turga

i602

Ob

603

Upp

erIrtysh

604

Chu

ya605

Yenise

i606

Lake

Baika

l607

Taim

yr608

Lena

609

Kolym

a610

Anad

yr611

East

Chu

kotka

612

Koryakia

613

Kam

chatka

andNorthernKurils

614

Okh

otsk

Coa

st615

Coa

stal

Amur

616

Lower

Amur

617

MiddleAm

ur618

Argu

n619

Shilka(Amur)

620

Son

ghua

Jian

g621

Inne

rMon

golia

Endo

rheicBas

ins

622

Wes

tern

Mon

golia

623

Dzung

aria

624

Balka

sh–A

laku

l625

Tarim

626

Lower

andMiddleSyr

Darya

627

Lake

IssykKul–U

pper

Chu

628

NorthernCen

tral

AsianHighlan

ds629

Aral

Sea

Drainag

es630

MiddleAm

uDarya

631

Upp

erAm

uDarya

632

Qaida

n633

Upp

erHua

ngHe

634

Upp

erHua

ngHeCorrid

or635

Hua

ngHeGreat

Ben

d636

Lower

Hua

ngHe

637

Liao

He

638

EasternYellowSea

Drainag

es639

Sou

thea

sternKorea

nPe

nins

ula

640

Ham

gyon

g–San

mae

k641

Sak

halin,H

okka

ido,

andSikho

te–

Alin

Coa

st642

Hon

shu–

Shikoku

–Kyush

u643

BiwaKo

Sout

hern

Asia

701

Baluchistan

702

Helman

d–Sistan

703

Lower

andMiddleIndu

s704

Yagh

istan

705

Indu

sHim

alayan

Foothills

706

Upp

erIndu

s707

Tibe

tanPlatea

uEn

dorheicDrainag

es708

Nam

uda–

Tapi

709

Gan

gesDelta

andPlain

710

Gan

gesHim

alayan

Foothills

711

Upp

erBrahm

aputra

712

MiddleBrahm

aputra

713

NorthernDeccanPlatea

u714

Sou

thernDeccanPlatea

u715

Wes

tern

Gha

ts716

Sou

thea

sternGha

ts717

SriLa

nkaDry

Zone

718

SriLa

nkaWet

Zone

719

ChinHills–Arak

anCoa

st720

Sita

ng–Irawad

dy721

Upp

erSalwee

n722

Lower

andMiddleSalwee

n

723

Inle

Lake

724

Upp

erLa

ncan

g(M

ekon

g)725

ErHai

726

Lower

Lancan

g(M

ekon

g)727

Kho

ratPlatea

u(M

ekon

g)728

Kratie

–Stung

Tren

g(M

ekon

g)729

Mekon

gDelta

730

Sou

thernAn

nam

731

EasternGulfof

Thailand

Drainag

es732

Cha

oPh

raya

733

Mae

Khlon

g734

Malay

Penins

ulaEa

sternSlope

735

NorthernCen

tral

Sum

atra–W

estern

Malaysia

736

Aceh

737

Indian

Ocean

Slope

ofSum

atra

andJava

738

Sou

thernCen

tral

Sum

atra

739

Sou

thernSum

atra–W

estern

Java

740

Cen

tral

andEa

sternJava

741

Kap

uas

742

Northwes

tern

Borne

o743

Borne

oHighlan

ds744

Northea

sternBorne

o745

EasternBorne

o746

Sou

thea

sternBorne

o747

Maluk

ku748

Less

erSun

daIsland

s749

Sulaw

esi

750

Malili

Lakes

751

Lake

Poso

752

Minda

nao

753

Lake

Lana

o755

NorthernPh

ilipp

ineIsland

s756

Palawan

–Bus

uang

a–Mindo

ro757

Wes

tern

Taiwan

758

EasternTaiwan

759

Haina

n760

NorthernAn

nam

761

Son

gHon

g762

Yunn

anLa

kes

763

XiYian

g764

Upp

erYang

tze

765

MiddleYang

tze

766

Lower

Yang

tze

767

Coa

stal

Fujian–

Zeijang

768

Anda

man

Island

s769

Nicob

arIsland

s

Aust

ralia

and

Paci

fic801

Sou

thwes

tern

Australia

802

Pilbara

803

Kim

berle

y804

Paleo

805

Arafura–

Carpe

ntaria

806

Lake

Eyre

Bas

in807

EasternCoa

stal

Australia

808

Murray–Darling

809

Bas

sStraitDrainag

es810

Sou

thernTasm

ania

811

New

Zealan

d812

Voge

lkop

–Bom

berai

813

New

Guine

aNorth

Coa

st814

New

Guine

aCen

tral

Mou

ntains

815

Sou

thwes

tNew

Guine

a–Tran

s-FlyLo

wland

816

Papu

anPe

nins

ula

817

Bismarck

Archipelag

o818

Solom

onIsland

s819

Vanu

atu

820

New

Caled

onia

821

Fiji

822

Wallis–Futun

a823

Sam

oas

824

Society

Island

s825

Tubu

aiIsland

s826

Marqu

esas

Island

s827

Rap

a828

Haw

aiianIsland

s829

East

CarolineIsland

s830

Wes

tCarolineIsland

s


highland lakes in Cameroon along with Africa’s Lake Tana;northwestern and eastern Madagascar; freshwaters fromTurkey’s centralAnatolia region, the northern British Isles, thePhilippines, Sri Lanka, India’s western Ghats, the southwest-ern Balkans, and northwest Mediterranean; southwesternAustralia and nearly the entire island of NewGuinea; Eurasianlakes, including Baikal, Inle, and Sulawesi’s Lake Poso andMalili system;DeathValley in the United States and Mexico’sPánuco system; and South America’s Iguaçu River, Lake Tit-icaca, and the freshwaters of both the MataAtlántica and thecontinent’s northwestern Pacific coast. Additionally, newlyavailable data show that some systems previously recognizedfor high endemism, such as those of SouthAmerica’s Guianas,also exhibit exceptional richness.

Because our ecoregions cover all nonmarine waters, andbecause they often exist as subdivisions of major river basins,our results also highlight a number of smaller systems for thefirst time in global analyses.Using finer-resolution data allowedus to identify the high richness of the Congo’s Malebo Pooland Kasai basin. Cuba and Hispaniola stand out forendemism, along with the Amazon’s western piedmont and

the Tocantins-Araguaia systems. The Tocantins-Araguaia, aswell as the highly endemic São Francisco,were defined as unitsof analysis inRevenga and colleagues (1998),but fish datawereunavailable for those basins when that study was done.Systems never before analyzed globally but recognized inour results as exceptionally rich for fish include those of theMalay Peninsula’s eastern slope and Japan.A large number ofecoregions are identified for the first time for highly endemicfaunas,measured as percentage endemism.Newly identifiedecoregions with at least 50% endemism include Africa’sCuanza,Australia’s Lake Eyre Basin,Mexico’sMayrán-Viesca,and New Zealand, as well as a large number of highly depau-perate ecoregions such asAfrica’s karstveld sink holes,Turkey’sLake Van, the Oman Mountains, western Mongolia, andHawaii.

Each of the biodiversity analyses that we offer here em-phasizes different sets of ecoregions, suggesting that a singlemeasure of species diversity might overlook ecoregions ofimportant biodiversity value. In a comparative analysis ofbiodiversity value, ecoregions are probably best evaluatedagainst others within the same region,with similar historical

Articles


Figure 2. Preliminary freshwater fish species data for ecoregions: (a) species richness, (b) number of endemic species, (c)percentage endemism, and (d) species per ecoregion area. Numbers may be adjusted on the basis of an ongoing process tocorrect nomenclatural errors. Natural breaks (Jenk’s optimization) was the classification method used for figures (a)–(c).This method identifies breakpoints between classes using a statistical formula that identifies groupings and patternsinherent in the data.


and environmental characteristics, and of similar size to ac-count for the typically positive relationship between riverdischarge and fish species richness (Oberdorff et al. 1995).Nonetheless, some systems, such as theAmazon and many ofAfrica’s Rift Valley lakes, stand out by nearly any measure offish biodiversity and are indisputable global conservationpriorities.

Conservation applicationsThe ecoregion map and associated species data summarizedhere have a number of conservation applications. At globaland regional scales the ecoregion map can be used to distin-guish distinct units of freshwater biodiversity to be representedin conservation efforts. The Convention on Wetlands, forinstance, requires that sites nominated as wetlands of inter-national importance—with wetlands defined to include allfreshwaters—be evaluated against a “biogeographic region-alization” criterion (Ramsar Bureau 2006). Lack of a globalbiogeographic scheme has stalled the application of this cri-terion, but our global map and database may provide a nec-essary framework for identifying broadscale gaps in protection.Similarly, progress toward the establishment of representativenetworks of freshwater protected areas, as called for by the thirdIUCN World Conservation Congress, the fifth World ParksCongress, and the seventh Meeting of the Conference of theParties to the Convention on Biological Diversity, can now bemeasured using ecoregions as a proxy for finer-scale globalspecies or habitat distribution data. At a regional level, thefreshwater ecoregion map may be used as supplementary in-formation for implementation of the European Union’sWa-ter Framework Directive (2000/60/EC), which requires acharacterization of surface water bodies and currently usesregions defined by Illies (1978).

A primary use of ecoregions is as conservation planningunits (Higgins 2003). Our attribution of freshwater fishspecies data to ecoregions is an important first step for data-poor regions. Organizations or agencies with regional man-dates may choose to compare biodiversity values acrossecoregions in the process of setting continental priorities(Abell et al. 2000, Thieme et al. 2005). At the basin scale,ecoregions can help to introduce biodiversity information intowater-resource or integrated-basin management activities(Gilman et al. 2004).Where major basins are divided amongmultiple freshwater ecoregions, whole-basin exercises canuse ecoregions as stratification units to ensure adequate rep-resentation of distinct biotas.Where unconnected drainagesare combined into a single freshwater ecoregion,plannersmaychoose to consider a counterintuitive planning unit to in-corporate biogeographic patterns. Freshwater ecoregions de-fined in previous exercises have already been put to use by theNature Conservancy and WWF in numerous conservationplanning efforts acrossNorthAmerica (e.g.,UpperMississippi;Weitzell et al. 2003), South America (e.g., the Pantanal; deJesus 2003), and Africa (e.g., the Congo basin; Kamdem-Toham et al. 2003).

Caveats and limitationsEcoregions are delineated based on the best available infor-mation, but data describing freshwater species and ecologi-cal processes are characterized by marked gaps and variationin quality and consistency. Data quality is generally consid-ered high for NorthAmerica,Australia,New Zealand, Japan,western Europe, and Russia; moderate for Central America,the southern cone of South America, southern and westernAfrica, Oceania, and the Middle East; and poor for much ofsoutheastern Asia, central and eastern Africa, and SouthAmerica north of the Paraná River basin.

Freshwater ecoregions are not homogeneous units.Withinindividual ecoregions there will be turnover of species alonglongitudinal gradients of river systems and across differenthabitats such as flowing and standing-water systems. Theinclusion of multiplemacrohabitat types within a given fresh-water ecoregion is a marked departure from terrestrial ecore-gions, which typically encompass a single vegetation-definedbiome (e.g., deciduous forests, evergreen forests, or scrub;Wikramanayake et al. 2002).

Ecoregions are imperfect units for highlighting certainhighly distinct and highly localized assemblages occurring atsubecoregion scales. Examples include many peat swamps orsubterranean systems.Underground systems such as caves andkarstsmay require their own planning framework, as ground-water catchmentsmay not correspondwith the surface-watercatchments upon which our ecoregions are built.

For reasons of practicality and scale, our ecoregion frame-work does not take into account the distributions of freshwaterspecies such as invertebrates, reptiles, and amphibians. Thisis a limitation of the ecoregional approach presented here,which is especially problematic for places such as isolatedislands where freshwater fish provide little information toinform biogeographic delineation. We hope this taxonomicomission will serve as motivation for generating and syn-thesizing global data for other taxonomic groups to providecomplementary information for conservation planners, par-ticularly whenworking at subecoregional scales.We recognizethat improved information in the futuremaywarrantmap re-visions, and we highlight areas of greatest data uncertainty inpart to encourage enhanced research investment in thoseplaces.We believe that the critical state of freshwater systemsand species argues against waiting for ideal biodiversity datato be developed before generating urgently needed conser-vation tools like the ecoregion map.

Shifting transition zones for species are common, and werecommend that ecoregions be viewed as logical units formoredetailed analyses and strategies. Ecoregions are intended to de-pict the estimated original extent of natural communitiesbefore major alterations caused by recent human activities,but original distributions can be difficult to reconstruct. Asnew species are described, our understanding of distributionpatterns may also change. Ecoregional delineation is an iter-ative process, and changes to ecoregion boundaries should beincorporated as new information becomes available.

Articles



There is no definitive, error-free data source for classifyingfish species as freshwater, brackish, ormarine.We chose to usethe global FishBase habitat assignments, which are derivedfrom the literature, to ensure that any given species in our data-base would be classified consistently wherever it occurred.Werecognize that errors of omission or commission may derivefrom inaccuracies in the FishBase assignments as well asfrom the habitat plasticity of some species. All species in-formation provided to us by experts, regardless of habitatassignment, is retained in our database for future analyses.

The preliminary richness and endemism numbers pre-sented here are in some cases markedly different from exist-ing estimates in the literature. For example, our tally for LakeMalawi contains 431 described fish species, but other estimatesrun as high as 800 or more (Thieme et al. 2005). Our omis-sion of undescribed species, as well as the conservativeapproach taken by experts in using only robust speciesoccurrence data, account for many of these lower-than-expected numbers.Numbers of endemics may in some casesbe higher than expected because endemics were identifiedstrictly through a database query for unique occurrences,and many species lists are undoubtedly incomplete or usesynonyms. We anticipate that many tallies will change withfurther refinement of species lists but that the broad patternspresented here will hold.

ConclusionsThe newly available species data attributed to ecoregions hasimportant implications for prioritizing conservation invest-ments. As one illustration, in 2005 the Global EnvironmentFacility (GEF), which spends more than $1 billion each yearon environmental projects, adopted a new resource allocationframework.Terrestrial ecoregion maps and biodiversity datawere notable inputs to the framework, but parallel fresh-water information to help guide investments was lacking.The GEF framework fortunately leaves open the possibilityof incorporating freshwater ecoregions and biodiversity dataat a later date (GEF 2005).

In addition to providing data for scientific and conserva-tion purposes, we aim to give the largest possible number ofpeople access to the ecoregion-level information collected inassociation with the global map. The information will befreely available on the Internet (www.feow.org) as well as inbrochures, posters, and other publications. The freshwaterecoregion map covers virtually all land surfaces on Earth, sopeople around the globe will have the opportunity to learnabout the freshwater systems where they live.

For most policymakers, water resource managers, andeven conservationists, freshwater biodiversity is more of anafterthought than a central consideration of their work. Thefreshwater ecosystem services that support the lives and liveli-hoods of countless people worldwide are a far larger concern.Yet freshwater biodiversity and ecosystem services are linkedthrough ecological integrity, and better-informed efforts toconserve freshwater biodiversity should benefit human com-munities as well. The freshwater ecoregions of the worldmap

and associated species data begin to improve access to pre-viously dispersed and difficult to access freshwater biodiver-sity information.We hope that this set of products catalyzesadditional work toward a better understanding of freshwaterspecies distributions and—of equal if notmore importance—leads to a ramping up of freshwater conservation activityand success.

AcknowledgmentsThe authors would like to thank the dozens of scientists whocontributed to development of the ecoregion map and syn-thesis of fish species data: E. K. Abbam, Vinicius Abilhoa,Angelo Agostinho, James Albert, Hector Samuel VeraAlcazar, Claudio Baigun, Eldredge Bermingham,Tim Berra,Vinicius Bertaco, Richard Biggins, Flavio Bockmann, PauloBuckup, Noel Burkhead, Brooks Burr, Mary Burridge,LaurenChapman,LindsayChatterton,BarryChernoff, LyndaCorkum, Ian Cowx, William Crampton, Alain Crivelli,Carolina Joana da Silva, Tim Davenport, Luc De Vos,Ignacio Doadrio, Carlos DoNascimiento, Luis FernandoDuboc, Brian Dyer, Carlo Echiverri, Jean Marc Elouard,Joerg Freyhof, Christopher Frissell, German Galvis, AngusGascoigne, Abebe Getahun, A. Gopalakrishnan, MichaelGoulding, Jon Harding, Tan Heok Hui, Liu Huanzhang,Leonardo Ingenito, Michel Jégu, Howard Jelks, Aaron Jenk-ins, Wolfgang Junk,Ad Konings, Friedhelm Krupp, PhilippeLalèyè, Carlos Alcala Lasso, Christian Lévêque, Flávio C. T.Lima, Cas Lindsey, Jorge Liotta, Marcelo Loureiro, CarlosLucena, Margarete Lucena, Paulo Henrique Lucinda, Anto-nioMachado-Allison,ChristopherMagadza, LuisMalabarba,Mabel Maldonado, Maria Cristina Dreher Mansur, LarryMaster, Don McAllister, Robert McDowall, J. D. McPhail,Geraldo Mendes dos Santos, Naércio A. Menezes, RobertoCarlos Menni, Jose Ivan Mojica, Peter Moyle, Thierry Ober-dorff, Javier Maldonado Ocampo, Mike K. Oliver, HernanOrtega, Mark Oswood, Vadim E. Panov, Carla SimonePavanelli,Christine Poellabauer,David Propst, Edson Pereira,Saul Prada, Francisco Provenzano, Gordon McGregor Reid,Anthony J. Ribbink, Francisco Antonio Rodrigues Barbosa,Ricardo S. Rosa, Norma J. Salcedo-Maúrtua, Jansen AlfredoSampaio Zuanon, Robert Schelly, Michael Schindel, UliSchliewen, Juan Jacobo Schmitter Soto, Martin Schneider-Jacoby,UweHorst Schulz,Lothar Seegers,Ole Seehausen,ScottSmith, John S. Sparks,Don Stewart,DonaldTaphorn,Christo-pher Taylor, Guy Teugels, Louis Tsague,Denis Tweddle, PaulVan Damme,D.Thys van denAudenaerde, Stephen J.Walsh,ClaudeWeber, RobinWelcomme, James D.Williams, PhillipWillink, and Stamatis Zogaris. Additionally, William Esch-meyer and Stan Blum provided critical support toward im-proving our fish species database.The importance of biologicalcollections and the work of taxonomists are basic to all bio-geographic mapping projects, and so we acknowledge andhighlight the fundamental contribution of collections and tax-onomy to this effort and to conservation generally. Institutionsand organizations that have generously provided data andassistance include theAmerican Fisheries Society’s Endangered

Articles



Species Committee, the American Museum of NaturalHistory, Belgium’s Royal Museum for Central Africa, theCaliforniaAcademy of Sciences, FishBase, Fundación La Sallede Ciencias Naturales, Instituto Nacional de Pesquisas daAmazônia, IUCN, Museu de Ciências e Tecnologia PUCRS,Museo de Zoologia de la Universidad Central de Venezuela,Museo de Zoologia de la Universidad Nacional de los LlanosOccidentales, Museu Nacional do Rio de Janeiro, Nature-Serve, South African Institute for Aquatic Biodiversity, andthe ZoologicalMuseumof theUniversity of Copenhagen.Eze-quiel Zamora,George Ledec,Douglas Graham, and GonzaloCastro were instrumental in the earliest stages of this project.We also thank Nasser Olwero for his development of theFEOWWeb site; Eric Dinerstein for his guidance and reviewof an earlier manuscript; and many additional former andcurrent WWF and Nature Conservancy staff acknowledgedon the FEOWWeb site, including but not limited to Jamie Pit-tock, Allison Pease, Brian Blankespoor, and Tucker Gilman.This work was supported in part by grants toWWF from theCoca-Cola Company and JohnsonDiversey Inc. Additionalsupport was generously provided to the Nature Conservancyby Bill Barclay, Ofelia Miramontes, and John Mordgridge.Work in South America was supported in part by the USAgency for International Development through award num-ber EDG-A-00-01-0023-00 for the Parks in Peril Program.

References citedAbell RA, et al. 2000. Freshwater Ecoregions of North America: A Conser-

vation Assessment.Washington (DC): Island Press.AllenGR.1991. FieldGuide to the Freshwater Fishes of NewGuinea.Madang

(Papua New Guinea): Christensen Research Institute.Allen GR, Midgley SH,Allen M. 2002. Field Guide to the Freshwater Fishes

of Australia. Perth (Australia): Western Australian Museum.Balian EV, Lévêque C, Segers H, Martens K, eds. 2008. Freshwater Animal

Diversity Assessment. Dordrecht (Netherlands): Springer.Banarescu P. 1990. Zoogeography of FreshWaters, vol. 1: General Distribu-

tion andDispersal of FreshwaterAnimals.Weisbaden (Germany):AULA.Berra TM. 2001. Freshwater Fish Distribution. San Diego: Academic Press.Burgess GH, Franz R. 1989. Zoogeography of the Antillean freshwater fish

fauna. Pages 263–304 inWoods CA, Sergile FE, eds. Biogeography of theWest Indies: Patterns and Perspectives. Boca Raton (FL): CRC.

BussingWA. 1976.Geographic distribution of the San Juan ichthyofauna ofCentral America with remarks on its origin and ecology. Pages 157–175inThorsonTB, ed. Investigations of NicaraguanLakes.Lincoln:Universityof Nebraska.

Contreras-Balderas S. 2000. Biogeografía mexicana de peces continentales.Mexicoa 2: 80–84.

de Jesus F, coord. 2003.Classification of Aquatic Ecosystems of the Pantanaland theUpper ParaguayWatershed.Brasilia (Brazil):NatureConservancy.

Dinerstein E, Olson DM, Graham DJ,Webster AL, Primm SA, BookbinderMP, Ledec G. 1995. A Conservation Assessment of the Terrestrial Eco-regions of Latin America and the Carribean. Washington (DC): WorldBank.

Dudgeon D, et al. 2006. Freshwater biodiversity: Importance, threats, statusand conservation challenges. Biological Reviews 81: 163–182.

EschmeyerWN. 2006.The Catalog of Fishes On-line (updated 7 November2006).CaliforniaAcademy of Sciences. (14March 2008;www.calacademy.org/research/ichthyology/catalog/fishcatsearch.html)

[GEF] Global Environment Facility. 2005. Technical Paper on the GEFResource Allocation Framework. (4 April 2008; www.gefweb.org/Operational_Policies/Resource_Allocation_Framework.html)

Gery J. 1969.The fresh-water fishes of SouthAmerica.Pages 828–848 in FitkauEJ, ed.Biogeography and Ecology in SouthAmerica.TheHague (Nether-lands): W. Junk.

Gilman RT, Abell RA, Williams CE. 2004. How can conservation biologyinform the practice of integrated river basinmanagement? Journal of RiverBasin Management 2: 135–148.

Groombridge B, Jenkins M. 1998. Freshwater Biodiversity: A PreliminaryGlobal Assessment. Cambridge (UK):World Conservation MonitoringCentre.

Groves CR, JensenDB,Valutis LL,RedfordKH,ShafferML, Scott JM,Baum-gartner JV, Higgins JV, Beck MW, Anderson MG. 2002. Planning forbiodiversity conservation: Putting conservation science into practice.BioScience 52: 499–512.

Higgins JV. 2003. Maintaining the ebbs and flows of the landscape:Conservation planning for freshwater ecosystems. Pages 291–318 inGroves C, ed.Drafting a Conservation Blueprint: A Practitioner’s Guideto Planning for Biodiversity. Washington (DC): Nature Conservancyand Island Press.

Illies J, ed. 1978. Limnofauna Europaea. New York: Gustav Fischer.Kamdem-Toham A, et al. 2003. Biological Priorities for Conservation in

theGuinean-Congolian Forest and Freshwater Region.Libreville (Gabon):WWF-CARPO.

Keith P, Vigneux E, Marquet G. 2002. Atlas des Poissons et des Crustacésd’Eau Douce de Polynésie Française. Paris:Muséum National d’HistoireNaturelle.

Kottelat M, Freyhof J. 2007. Handbook of European Freshwater Fishes.Switzerland: Steven Simpson Books.

Lévêque C, ed. 1997. Biodiversity Dynamics and Conservation: The Fresh-water Fish of Tropical Africa. Cambridge (UK): Cambridge UniversityPress.

Lévêque C, Balian EV, Martens K. 2005. An assessment of animal speciesdiversity in continental waters. Hydrobiologia 542: 39–67.

Lundberg JG, Kottelat M, Smith GR, Stiassny MLJ, Gill AC. 2000. So manyfishes, so little time: An overview of recent ichthyological discovery incontinental waters.Annals of the Missouri Botanical Garden 87: 26–62.

Maxwell JR, Edwards CJ, Jensen ME, Paustian SJ, Parrott H,Hill DM. 1995.A Hierarchical Framework of Aquatic Ecological Units in North Amer-ica (Nearctic Zone). St. Paul (MN):USDA Forest Service,North CentralForest Experiment Station. General Technical Report NC-176.

McDonald R, McKnight M, Weiss D, Selig E, O’Connor M, Violin C,MoodyA.2005. Species compositional similarity and ecoregions:Do eco-region boundaries represent zones of high species turnover? BiologicalConservation 126: 24–40.

McDowall RM.1990.NewZealand Freshwater Fishes:ANatural History andGuide. Auckland (New Zealand): Hinemann Reed.

[MEA] Millennium Ecosystem Assessment. 2005. Ecosystems and HumanWell-being: Synthesis.Washington (DC): Island Press.

Menni RC. 2003. Peces y ambientes en la Argentina continental. Mono-graph of the Museo Argentino de Ciencias Naturales 5: 1–316.

Miller RR,MinckleyWL,Norris S. 2005.Fishes of México.Chicago:Universityof Chicago Press.

NatureServe. 2006. NatureServe Explorer: An Online Encyclopedia of Life.Version 6.1.NatureServe. (17March 2008;www.natureserve.org/explorer)

Oberdorff T, Guegan JF, Hugueny B. 1995. Global scale patterns of fishspecies richness in rivers. Ecography 18: 345–352.

Olson DM, Dinerstein E, Canevari P, Davidson I, Castro G, Morisset V,Abell R,Toledo E. 1998. Freshwater Biodiversity of LatinAmerica and theCaribbean:A ConservationAssessment.Washington (DC): BiodiversitySupport Program.

Olson DM, et al. 2001. Terrestrial ecoregions of the world: A new map oflife on Earth. BioScience 51: 933–938.

Omernik JM. 1987. Ecoregions of the conterminous United States. Annalsof the Association of American Geographers 77: 118–125.

Paavola R, Muotka T,Virtanen R, Heino J, Jackson D, Maki-Petays A. 2006.Spatial scale affects community concordance among fishes, benthicmacroinvertebrates, and bryophytes in streams. Ecological Applications16: 368–379.

Articles



Ramsar Bureau. 2006. Strategic Framework for the List of Wetlands ofInternational Importance of the Convention on Wetlands (Ramsar,Iran, 1971), edition 2006. (17 March 2008;www.ramsar.org/key_guide_list2006_e.htm)

Rauchenberger M. 1988. Historical biogeography of Poeciliid fishes in theCaribbean. Systematic Zoology 37: 356–365.

Reis R, Kullander S, Ferraris C, eds. 2003. Check List of the FreshwaterFishes of South and Central America. PortoAlegre (Brazil): EDIPUCRS.

Revenga C, Kura Y. 2003. Status and trends of biodiversity of inland waterecosystems.Montreal (Canada): Secretary of theConvention onBiologicalDiversity. Technical Series no. 11.

Revenga C, Murray S, Abramovitz J, Hammond A. 1998.Watersheds of theWorld: Ecological Value and Vulnerability. Washington (DC): WorldResources Institute.

Revenga C, Campbell I, Abell R, de Villiers P, Bryer M. 2005. Prospects formonitoring freshwater ecosystems towards the 2010 targets. Philosoph-ical Transactions of the Royal Society B 360: 397–413.

Ringuelet R. 1975. Zoogeografía y ecología de los peces de aguas continen-tales de la Argentina y consideraciones sobre las áreas ictiológicas deAmérica del Sur. Ecosur 2: 1–122.

Roberts TR. 1975. Geographical distribution of African freshwater fishes.Zoological Journal of the Linnean Society 57: 249–319.

Scott WB, Crossman EJ. 1998. Freshwater Fishes of Canada. Oakville(Canada): Galt House.

Skelton PH. 1994.Diversity and distribution of freshwater fishes in East andSouthernAfrica.Annals of the Royal Central Africa Museum (Zoology)275: 95–131.

Skelton PH,Cambray JA, Lombard A, Benn GA. 1995. Patterns of distribu-tion and conservation status of freshwater fishes in South Africa. SouthAfrican Journal of Zoology 30: 71–81.

SpaldingMD,et al. 2007.Marine ecoregions of theworld:A bioregionalizationof coast and shelf areas. BioScience 57: 573–583.

Thieme ML,Abell R, Stiassny MLJ, Skelton P, Lehner B, Teugels GG,Diner-

stein E, Kamdem-Toham A, Burgess N,Olson D. 2005. Freshwater Eco-

regions of Africa and Madagascar: A Conservation Assessment.

Washington (DC): Island Press.

TonnWM.1990.Climate change and fish communities:A conceptual frame-

work. Transactions of the American Fisheries Society 119: 337–352.

Udvardy MDF. 1975.A Classification of the Biogeographic Provinces of the

World. Gland (Switzerland): IUCN. Occasional Paper no. 18.

Unmack PJ. 2001. Biogeography of Australian freshwater fishes. Journal of

Biogeography 28: 1053–1089.

Vinson MR, Hawkins CP. 2003. Broad-scale geographical patterns in local

stream insect genera richness. Ecography 26: 751–767.

Wallace AR. 1876. The Geographical Distribution of Animals. New York:

Harper.

Wasson J-G, Barrera S, Barrere B, Binet D,Collomb D,Gonzales I, Gourdin

F, Guyot J-L, Rocabado G. 2002. Hydro-ecoregions of the Bolivian

Amazon: A geographical framework for the functioning of river eco-

systems. Pages 69–91 in McClain ME, ed. The Ecohydrology of South

American Rivers and Wetlands. Wallingford (UK): International

Association of Hydrological Sciences.

Weitzell RE,KhouryML,GagnonP, Schreurs B,GrossmanD,Higgins J. 2003.

Conservation Priorities for Freshwater Biodiversity in the Upper

Mississippi Basin. Arlington (VA): NatureServe and The Nature

Conservancy.

Wikramanayake E,Dinerstein E, Loucks C,Olson D,Morrison J, Lamoreux

J, McKnight M, Hedao P. 2002. Ecoregions in ascendance: Reply to

Jepson and Whittaker. Conservation Biology 16: 238–243.

doi:10.1641/B580507Include this information when citing this material.

Articles



Freshwater Ecoregions of the World: A New Map of Biogeographic Units for Freshwater Biodiversity Conservation.

Documents

freshwater ecosystems

articlesnew freshwater

new freshwater fish

new species

freshwater habitats

freshwater realm

map of freshwater ecoregions

imperiled freshwater