JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION VOL. 33, NO.5 AMERICANWATER RESOURCES ASSOCIATION OCTOBER 1997

DISTINGillSHING BETWEEN WATERSHEDS AND ECOREGIONSl

JamesM. Omernik and Robert G. Bailey2

ABSTRACT: In an effort to adopt more holistic ecosystem To accomplishthis redirection, the need for a spa- approaches to resource assessment and management, many state and federal agencies have begun using watershed or ecoregion tial framework is obvious.The problem is which one. frameworks. Although few would question the need to make this Do we use existing frameworks, or do we need to move from dealing with problems and issues on a case by case or developone to fit this particular purpose?Many have point-type basis to broader regional contexts, misunderstanding of felt that watershedsprovide the spatial tool necessary each of the frameworks has resulted in inconsistency in their use for effective research, assessment,and management and ultimate effectiveness. The focus of this paper is on the clarifi- cation of both frameworks. We stress that the issue is not whether of ecosystems(Water Environment Federation, 1992; to use watersheds (or basins or hydrologic units) or ecoregions for Armitage, 1995; Montgomery et al., 1995; Parsons, needs such as developing ecosystem management and non-point 1985; USFWS, 1995; Cannon, 1994; Lotspeich, 1980; source pollution strategies or structuring water quality regulatory FEMAT, 1993; Maxwell et al., 1995; CoastalAmerica, programs, but how to correctly use the frameworks together. Defmi- 1994; USEPA, 1996a). However, publications recom- tions, uses, and misuses of each of the frameworks are discussed as well as ways watersheds and ecoregions can be and have been used mending use of the watershed framework do not all together effectively to meet resource management needs. agree on how and at what scales this use should be (KEY TERMS: ecoregions; basins; watersheds; hydrologic units; undertaken. For example, Montgomery et al. (1995) ecosystem management.) and the report by the Forest EcosystemManagement AssessmentTeam (FEMAT, 1993) suggestthe water- shed (or basin) framework is applicable at two middle hierarchical levels, with physiographic regions or eco- BACKGROUND logical regions (also known as ecoregions),such as we have developed (Omernik, 1995; Bailey, 1995a, Much of the recent popularity with the terms 1995b), being appropriate at the broadest level, and "watershed" and "ecoregion"has come about because project or site delineations being most useful at the of the attempt by government agencies at regional, more detailed levels (largest scales).Others, such as state, national, and international levels to adoptmore the U.S. Fish and Wildlife Service (USFWS), recom- holistic approaches to research, assess, monitor, mend use of watersheds,basins, or hydrologic units at inventory, and manage their resources. The intent all hierarchical levels (USFWS, 1995). appears to be to shift from dealing with single issues, There have been warnings regarding the potential point-sourceproblems, and micro scales,to a broader misuse and misunderstanding of watersheds for approachthat considers spatial patterns of the aggre- structuring ecological research and management gate of natural and anthropogenic interrelationships (Omernik and Griffith, 1991; Hughes and Omernik, involving ecosystemsand their components.This no 1981; Hughes et al., 1994), but unfortunately such doubt stems from a growing realization of the insidi- caveatsare often veiled (Cannon,1994; Born and Son- ous nature of increased human population and modifi- zogni, 1995; Water "Environment Federation, 1992). cation of environmental resources(Holling, 1994). At the Watershed'93 Conference,John Cairns (1994)

IPaper No. 96178 of the Journal of the American Water ResourcesAssociation (formerly Water ResourcesBulletin). Discussions are open until June 1, 1998. ZRespectively, Research Geographer, U.S. Environmental Protection Agency, National Health and Environmental Effects Research Labora. tory, Western Ecology Division, 200 SW 35th St., Corvallis, Oregon 97333; and Leader, Ecosystem Management Analysis Center, USDA For. est Service, 3825 East Mulberry St., Fort Collins, Colorado 80524.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 935 JAWRA Omernik and Bailey gave a plenary presentation titled "The Current State management strategies. We stress that it is not an of Watersheds in the United States: Ecological and "either/or" argument. Both frameworks have impor- Institutional Concerns." Except for its appearancein tant purposes and are complementary when used the title, the word "watershed" canbe found only once together correctly. in the proceedingsmanuscript. The bulk of the paper was aptly focused on the complexities of ecosystems, what we do not know about them, and ecosystem management and restoration scenarios. Cairns DEFINITIONS stressed that although lip service is given to it, little understanding exists of ecosystems and ecosystem Ecoregions management in a holistic sense.He stated that cur- rent efforts to emphasize watershed management remain focused on componentssuch as water quality. The most glaring difference betweenthe definitions When discussing the spatial extrapolation of ecosys- of watersheds and ecoregionsconcerns the degree of tem level restoration activities, Cairns referred to agreement on the definitions. Whereas the definition of the term watershed is fairly widely accepted,there ecoregions rather than watersheds. Repeatedly, Cairns made the point that there has been a reluc- are marked differences of opinion regarding ecore- tance to deal with that in which few have experience gions and how they can or should be delineated. and understanding, i.e., ecosystem-leveldecisions. Much of the difference in approachesto define ecore- Jonathan Cannon (1994), in another paper in the gions sterns from a lack of agreement on a definition of that which we are attempting to regionalize - proceedingsof the Watershed'93 Conference,strongly ecosystemsin aggregate (Born and Sonzogni, 1995). endorsed the U.S. Environmental Protection Agency Allen et al. (1993)claimed that the conceptof ecosys- (EPA)watershed approachand cited the "Water Qual- ity 2000" report which stated that watershedsprovide tem is both widely understood and diffuse and the appropriate spatial framework for total environ- ambiguous. Some question the concept itself (Calli- mental and economic planning (Water Environment cott, 1995;Fitzsimmons, 1996). In discussingthe com- plexities of this problem, Kay and Schneider (1994) Federation 1992).He acknowledgedthe caveatburied stated that most North American ecologicaljournals in "Water Quality 2000" stating: "In some water- sheds, planning and management activities may be (particularly U.S. journals) do not consider holistic more effective in attaining water quality goals if they ecosystemresearch a fit topic because it does not fol- low traditional scientific methods in that there are are organized by ecologicalregions (sub-watersheds). not observer-independent ways of defining ecosys- This is because the natural differences in climate, tems. The definition of "ecosystem" as it relates to geology,soil, land form, and vegetation may not con- form strictly to hydrologic regions.These features can regionalization and ecosystemmanagement (versus individual lakes, streams, wetlands, forests, etc., com- determine the ecological character of surface water and near-surfacegroundwater." Notice, however, that prising ecosystems)appears to be evolving (Haeuber and Franklin, 1996; Grumbine, 1994; Barnes, 1993). in this statement ecoregions are considered "sub- Originally the definition was centered on the biota watersheds," indicating at least an imprecise use of and then becamethought of as subsuming biotic and terms, if not a lack of understanding of the difference abiotic characteristics but in the absence of humans. between ecoregionsand watershedsand their hierar- More recently the term has taken on a more holistic chical nature. meaning that considershumans as part of the biota The purpose of this paper is to clarify the difference (McDonnell and Pickett, 1993; Barnes, 1993;Rowe, between watersheds and ecological regions and to 1990, 1992). The definition has also evolved some- explain some appropriate and inappropriate uses of what regarding scale. It is now commonto consider these spatial frameworks. We will not present a dis- ecosystemsin a multi-scale sense, from specific sites cussion of the different techniques for defining ecore- to global regions, as opposed to mostly relatively gions. Although we are not in complete agreement homogeneoussmall areas. Some of the difference in regarding the delineation of ecoregions,we share the definition is the result of our different educational concern that a spatial framework of watersheds, backgrounds and experiencesand differences in the basins, or hydrologic units has very different purposes missions of the agencieswe work within. than one of ecologicalregions, and that there is an Although the authors of this paper have employed urgent need to clarify the differences to reduce the dissimilar approachesin developingecological regions misuse of each framework. We also wish to address (Bailey, 1995a; Omernik, 1995), our objectives have some common misconceptions about watersheds, been similar, and as we revise our understandings of ecoregions,and hydrologic units that are germane to the meaning of the term "ecosystems"the products of their utility for the regionalization of ecosystem

JAWRA 936 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Distinguishing BetweenWatersheds and Ecoregions our efforts to refine the ecoregion frameworks are category.More than one of these conditions occur in tending to look more alike. In broad terms, ecological many areas making the problem more complex. regions, at any scale, can be defined as areas with rel- Many parts of the country that have been affected ative homogeneityin ecosystems.Our intent has been by continental glaciation are pocked with lakes, pot- to depict regions within which the mosaic of ecosys- holes, swamps, and marshes where surface water tem components(biotic and abiotic as well as terres- doesnot drain directly into streams. Although most of trial and aquatic) is different than that of adjacent the Midwest, much of the Northeast, and the north- regions. ern fringe of the western United States contain these characteristics, they are most commonin North Dako- ta, Minnesota, Wisconsin, and Michigan (Figure 2). Watersheds Delimiting the watershed boundaries of large rivers can be approximated fairly accurately in much of this Unlike ecoregions, there is little disagreement region, but that is not the case with many of the regarding the definition of watersheds. Quite simply, smaller streams. Other parts of the country where they are topographic areas within which apparent watershed delimitation is particularly problematic surface water runoff drains to a specific point on a include the nearly level, karst and sand dominated stream or to a waterbody such as a lake. There is an state of Florida, the Sand Hills of Nebraska, the semi- infinite number of points from which topographic arid karst and playa lands of west Texas and Okla- watersheds can be delineated, although regarding homa and eastern New Mexico, and the deserts of streams, confluencesare normally used. Large water- western and southwesternUnited States. shedsare commonlytermed basins (e.g.,the Colorado River Basin or the Susquehanna River Basin). The hierarchical classification of hydrologic units as USE AND MISUSE OF ECOREGIONS mapped by the U.S. GeologicalSurvey (Seaberet ai., 1987)is made up of watersheds or segmentsof water- sheds often with adjacent interstices (areas in Ecoregionsare intended to provide a spatial frame- between). However, at each level of classification,the work for ecosystemassessment, research, inventory, majority of these hydrologic units are not true topo- monitoring, and management. These regions delimit graphic watersheds. large areas within which local ecosystems reoccur Much of the apparent usefulness of watershedsas more or less throughout the region in a predictable study units arises from the general understanding pattern. By observing the behavior of the different that the quantity and quality of water at a point on a kinds of systems within a region it is possible to pre- stream reflects the aggregateof the characteristics of dict the behavior of an unvisited one. This affords the the topographic area upgradient from that point. extrapolation mechanism for identifying areas from However, the conclusionthat becauseof this a frame- which site specific knowledge on ecosystembehavior work of watersheds,basins, or hydrologic units is ide- can be applied. As such, they also suggestareas with- ally suited for spatially organizing ecosystem in which similar responsesand managementstrate- management, or even water quality management,is gies are applicable (Bailey, 1987). Ecoregionsshould flawed for at least three major reasons. First, and be thought of as multi-purpose regions, designedto most important, the areas within which there is simi- show areas within which the aggregate of all terres- larity in the aggregate of geographic characteristics trial and aquatic ecosystemcomponents is different related to the quality and quantity of environmental from or less variant than that in other areas. They resources seldom if ever correspond to patterns in provide a commonspatial framework for the various topographic watersheds. Second, in many xeric resource managementagencies responsible for differ- regions of the country where watersheds can be ent aspectsof the environment (e.g., forests, fish and defined and "influent" streams predominate (where wildlife, wetlands, water quality, and agriculture) to streams feed the groundwater, as comparedto "efflu- organize their activities as they move toward a more ent," where the groundwater feedsthe streams),topo- holistic ecosystemapproach that requires considera- graphic watersheds do not encompass the same tion of all aspectsof the environment. integrating processesas in mesic and hydric areas. Applications of ecoregionsare appropriate at vari- Third, in many areas watersheds are difficult or ous scales. International applications include the impossible to define (Hughes and Omernik, 1981). analysis of types and distributions of protected areas These types of areas comprise roughly a third of the acrossNorth America, the evaluation of the represen- conterminous United States (Figure 1). Regions of tativeness of these areas,and the assessmentof cross- continental glaciation, deep sand, karst topography, boundary environmental impacts related to the North flat plains, and xeric climates all fall into this American Free Trade Agreement(NAFTA){Wiken and

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 937 JAWRA Omernik and Bailey

" fZi~,-__~-.~ ~; ) "'"9~ ~

IF I ~ '1.-- -.,---, J I ~~';,~~ ~~ '- / / C'1-.\;;> . //1 I ~ D \. ri~~l w ~~ ~T-~-U-~--~~--'fI:fJ -~{.7d ., i -"\;",,:,~/"1I f-'I'; ,. I -- w Jr!' ' ~---~-",~~- ~, '"" M' ~ J ~ " '~~ ~ ~ J ~r~__) " ~~~~~ "'-- ..0 I~F~ \---~' ,~, \ h I : , !4\-~-~ ; J """""'--"-1 ;; r-' -4 ~ 'J ,--_LJ"~~ -- s ~l ,I{ '! .I '\ (--, I , ~ Karstlands ~ ~ I' 'J' ,.j ~~~~ ~ Glaciatedareas "\: r \, [illill] Excessivelyaridareas with '\ W 200 aMi 200 400 lessthan 20 cm precipitation ~ 400 400 800'J'" Sandyareas, more than 50percent covered bysand

Figure 1. CharacteristicsThat Hinder or PrecludeWatershed Delineation. (Adapted from Hughesand Omernik, 1981.;

Lawton, 1995; Commission for Environmental Coop- the framework has been central to structuring envi- eration Working Group, 1996). At national levels, ronmental resourceregulatory programs and manage- existing monitoring networks of research sites, such ment strategies. The effectiveness of ecoregions for as those of the Long Term EcologicalResearch (LTER) stratifying stream water quality information has been network, can be comparedto ecoregionmaps to deter- demonstrated in many states including Arkansas mine where representation is inadequate and (Rohm et aI., 1987), Iowa (Wilton, 1996), Nebraska additional sites are needed (Bailey, 1995a). In recom- (Bazata, 1991), Ohio (Larsen et aI. 1986, 1988), mending a national aquatic ecosystem restoration Oregon (Hughes et aI., 1987; Whittier et aI., 1988), strategy for the United States, the National Research Texas (Hornig et aI., 1995), and Washington (Plot- Council stated that restoration goals and assessment nikoff, 1992).State resource managementagencies in strategies should be established for each ecoregion these states have used ecoregions primarily to set (National ResearchCouncil, 1992). Canada'suses of water quality standards and to developbiological cri- ecoregionsinclude reporting on the state of the envi- teria and nonpoint source pollution management ronment in that country, developing protected area goals. Davis et aI. (1996)reported that as of 1995, 15 strategies, and developingregional indicators of forest states were using ecoregionalreference conditions in disturbance and biodiversity (Governmentof Canada, their biological assessmentprograms and another 24 1991; EcologicalStratification Working Group, 1995). states were in the processof developing similar pro- The most common usage of ecoregions within grams. the United States has been at the state level where

JAWRA 938 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION kr Distinguishing BetweenWatersheds and Ecoregions

Figure 2. In Many RegionsAffected by Continental Glaciation,Such as This Area in Northwestern Wisconsin,Streams are Lacking, Many Lakes HaveNo Inlets or Outlets,and the ThpographicWatersheds of Rivers Can Only beApproximated.

As with any framework, ecoregions can be and issue here is recognition of the most appropriate hier- have been misused and misunderstood. Ecoregions archical level of ecologicalregionalization. However, have not been designedfor regionalization of a partic- in this case and others, the bulk of the misuse and ular characteristic. A number of studies have misunderstanding of ecotegions centers on recogni- attempted to evaluate the usefulness of ecoregionsby tion of their purposeand appropriate methods of eval- comparing them to patterns of specificfactors such as uating them. Ecoregions are generally useful for fish assemblages, wildlife communities, particular structuring the research, assessment,and manage- hydrologic characteristics, and macroinvertebratedis- ment of all environmental resources,but may not be tributions (e.g., Lyons, 1989; Inkley and Anderson, the best framework for anyone particular resource. 1982; Poff and Allan, 1995; Poff and Ward, 1989; Most important, however,ecoregions provide the spa- Spindler, 1996). Although in nearly every case the tial tool necessary to address the health, integrity, ecoregionsand subject of study were determinedto be and quality of the aggregate of environmental generally correlated, it is not surprising that other resources. spatial characteristics were often found to be more helpful. For example, elevation and watershed size were found to be more useful than a coarse level of ecoregionsin explaining differences in macroinverte- USE AND MISUSE OF WATERSHEDS brate community structure in Arizona streams (Spindler, 1996). Had a lower, more detailed hierar- Watersheds have been and will continue to be a chical level of ecoregionsbeen available for Arizona critical spatial framework for scientific study of the that would have reflected zonal (without the effectsof effects of natural and anthropogenic phenomena on landform) differences in vegetation, precipitation, water quality and quantity. Hundreds, or perhaps soils, and hydrologic characteristics, ecoregionswould thousands, of studies on the effects of agricultural have shown a stronger correlation. Hence, part of the and silvicultural practices have been based on data

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 939 JAWRA Omernik and Bailey generated in small (betweena hectare or two to tens similarity in the combination of geographic phenome- of square kilometers in size) watersheds.Much of the na that cause and reflect regional differences in the LTER work has been conducted in watersheds at quality of ecosystems and ecosystem components places such as Coweeta, Hubbard Brook, and H. J. (Omernik, 1995; Bailey, 1995a). Watersheds and Andrews (Franklin et al., 1990; Risser et al., 1993). basins do not correspondto these patterns. The authors have depended on watershed data for However, we stress that basins and watersheds evaluating the effectivenessof ecologicalregions (e.g., are the appropriate spatial unit for resourcemanage- Bailey, 1984; Hughes et al., 1986)and stream buffer ment agenciesto assessthe relative contribution of strips (Omemik et al., 1981), for clarifying nonpoint human activities to the quality and quantity of water source -stream nutrient level relationships at specific points on streams and on particular water (Omernik, 1977), and for mapping sensitivity of sur- bodies. Becausethey integrate the surface and sub- face waters to acidification (Omernik and Powers, surface flow of water upgradient from the point at 1983; Omemik and Griffith, 1986)and lake phospho- which measurements are made, watersheds allow rus regions (Rohm et al., 1995; Omemik et al., 1988). drainage basin-specificaccountings to be made offac- However, for the effective extrapolation of these tors such as point and nonpoint source pollutants, watershed data one must know the larger regions whose transport is associatedwith the movement of within which similar characteristics exist. Theselarg- water. Watershedsare essentialfor these purposes. er regions seldomcorrespond to watersheds or basins. In developinga spatial framework for implement- Whereasthe watershedsserve as the study units, eco- ing an ecosystemapproach to fish and wildlife conser- logical regions, rather than watersheds, provide the vation, the USDI Fish and Wildlife Service adapted extrapolation mechanism. the U.S. Geological Survey's hydrologic unit map Much of the use of a watershed as a spatial unit (USFWS, 1995). One cannot argue the importance of stems from our basic understanding of watershed drainage divides in helping to explain spatial differ- functions (Black, 1997) and how they relate to flood ences in fish assemblagesand abundance, although forecasting and prevention or water supply for agri- ecoregions have also been shown to be important cultural, urban, and industrial use. However, even (Hughes et ai., 1994;Hughes et ai., 1987;Lyons, 1989; though the purpose of watersheds for tracking water Whittier et ai., 1988; Pflieger, 1971). On the other supply is clear, use of the framework in a social-sci- hand, the watershed framework does not seemappro- ence context is not self evident (Ciriacy- Wantrup, priate if the intent of the USFWS was to adopt a 1961).The physical and economicconditions relative holistic approach to assessmentand management. to watershed functions have little correlation with Such an approach would recognize patterns in the patterns of consumption (Ciriacy-Wantrup, 1961) or quality, quantity, and similarity in interactions of all with the distributions of most geographicphenomena ecosystem components. Most of these components that affect or reflect spatial differences in ecosystem have little or no associationwith drainage divides. health, integrity, and quality (Omemik, 1995). A recent publication by the U.S. Forest Service Watershed-type frameworks are widely used by titled "A hierarchical framework of aquatic ecological resource management agencies. Although a few are units in North America (Nearctic Zone)," (Maxwell et changing, or considering changing, to ecoregions, ai., 1995) outlines a classification scheme for what most state environmental resource management they term "aquatic ecological units." It describes a agencies presently use basins or hydrologic units rather complex hierarchical framework of aquatic (Seaber et al., 1987) to organize their semi-annual units under an umbrella of "zoogeographic" zones, 305b reports to Congress [in responseto Section305 subzones,regions, and subregions (which are differ- (b) of the Federal Water Pollution Control Act of 1972 ent hierarchical levels of watersheds and hydrologic (Public Law 92-500)] on the status of water quality in units). Although the publication attempts to establish their state. Following guidance from the USEPA a linkage with ecoregions (which they term "geocli- Office of Water, a number of states have recently matic settings"), we believe it errs by inferring that adopted river basin or watershed approaches to watersheds,groups of watersheds, or hydrologic units address their resource assessmentand management form the primary framework for dividing ecosystems. needs (USEPA, 1995, 1996a, 1996b).We believe it is Moreover,the reader could easily draw a seconderro- inappropriate to promote the use of watershedsas an neous inference that aquatic and terrestrial ecosys- extrapolation tool for these purposes.State, regional, tems need to be consideredseparately. An "ecosystem and national level managementstrategies, particular- approach" recognizesthat ecosystemcomponents do ly those involving ecosystemmanagement, require not function as independent systems, rather they a spatial framework that considers the regional exist only in association with one another (Bailey tolerancesand capacities of the landscape.Ecoregions 1995a). were designedto fill that need and identify areas with

JAWRA 940 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Distinguishing BetweenWatersheds and Ecoregions

Boundaries of ecologicalsignificance emerge from CLARIFYING HYDROLOGIC UNITS studies that reveal corresponding change in ecosys- tem components.This is different from attempting to There is also a common misunderstanding that synthesize ecosystemunits by addition of components initially defined as things in themselves, with no hydrologic units do in fact comprise watersheds. For whole unit in mind. If we follow the rationale for sep- example,in a well meaning effort to implement a spa- tial and more integrated approach to environmental arate but equal systems, one could argue that we resource management, particularly regarding water need separate maps for soils, timber, range, water, fisheries, etc. Analyzing resourceinteractions is diffi- quality programs, the EPA has recently advocateda "watershed approach" (USEPA, 1995, 1997). This cult because each discipline selects its own unit of land for analysis (e.g., stands of trees for forestersand approachemphasizes "managing by watersheds"and recommendsuse of U.S. GeologicalSurvey hydrologic watersheds for water-quantity analysts). Decadesof units (Seaberet al., 1987) which are claimed to "pro- research and field operations by a host of practition- vide a common national framework for delineating ers have produced spatial classifications that deal with resources as singular and independent items. watersheds and their boundaries at a number of dif- ferent geographic scales" (USEPA, 1995). Another What is needed now is a new approachthat provides EPA Office of Water publication, accompaniedby a set a basis for a firm understanding of the relationships of maps, is directed toward characterizing "the aquat- and interactions of resourceson the sameunit of land. ic condition and vulnerability of each of the 2,150 An ecosystemapproach to resource evaluation stress- es the interrelationships among components rather watersheds in the continental United States" than treating each as separate characteristics of the (USEPA, 1997). Watersheds in this publication are defined as the eight digit cataloging units of the U.S. landscape. It provides a basis for making decisions about resource interaction. To accomplish this, we GeologicalSurvey hydrologic unit system. Aside from providing additional examples of misuses of water- believe that the terrestrial and aquatic components sheds, the problem here is that it is not possible to must be consideredtogether at the sametime, rather divide the country, any state, or as far as we know, than after separate classifications have been devel- any county, into a finite number of topographicwater- oped. sheds. For instance, in Tennessee only four of the Like the USFWS (1995)framework, the alignments eleven "accounting level" hydrologic units covering of the ecologicalregions suggestedby Maxwell et al. the state are true topographic watersheds (Seaberet (1995) follow drainage divides. Three hierarchical lev- al., 1987)(Figure3). Most are segments of watersheds els of these regions, termed zoogeographicsettings, with adjacent interstices. Less than half (26 of 54) of were proposed. In providing a rationale for using the "cataloging units," the next more detailed hierar- watersheds and drainage divides, Maxwell and his chicallevel of hydrologic units, that are partly or com- colleaguesclaimed that watershedsprovide a natural pletely in Tennessee,are true topographic watersheds nested hierarchy for ecological stratification over a (U.S. Department of the Interior, GeologicalSurvey, wide range of scales, and that watersheds integrate 1974). many physical, chemical, and biological processes One reason given by the USFWS for their decision affecting the form and function of both aquatic and to use what they term "watershed based units," was terrestrial ecosystems.However, spatial patterns of that "watersheds are discrete physical units that pro- most of the principal factors that determine the vide widely recognized and generally well-defined integrity of surface water resources (Karr, 1991; boundaries" (USFWS 1995). Another reason given Yoder, 1995) generally do not correspondwell to pat- was "watershedsprovide a vehicle to considerthe crit- terns of watersheds, but they do correspondto ecore- ical linkages between upstream and downstream gions (Hughes et al., 1994; Larsen et al., 1988;Brown effects." The implication here is that the hydrologic and Brown, 1994). Factors such as physical habitat, units they adapted and the "watershed based units" water chemistry, energy sources,and biota are direct- are at least similar to true topographic regions where ly associated with the aggregate of factors (climate, the limits of apparent surface water runoff can be geology,physiography, soils, and land covercharacter- clearly defined, and that the stream system within istics including vegetation) that are used to define these spatial units provides the pathways for interac- ecoregions. tions of the ecosystemcomponents. However, of the 41 USFWS "watershed based units" defined for the con- terminous United States, only 17 percent are actually topographic watersheds.One of these units, the Mis- sissippi Headwaterstrallgrass Prairie straddles a con- tinental divide; the northwestern portion drains north

JOURNAl OF THE AMERICAN WATER RESOURCES ASSOCIATION 941 JAWRA Omernik and Bailey

Truewatershed 0 100 Hydrologicunits I I I Miles Major streams 051302 Accountingunit code

Figure 3. AccountingLevel HydrologicUnits of TennesseeThat Are True Watersheds.(Adapted from Seaberet ai., 1987.) to the Hudson Bay, whereas the southeasternpart is THE NATIONAL NETWORK OF in the Mississippi drainage system. More important, HYDROLOGIC BENCH MARKS from a standpoint of suitability for framing ecosystem or water quality assessmentand management, this The U.S. GeologicalSurvey's Network of Hydrolog- "watershed based unit" covers distinctly different eco- ic Bench Marks, as envisioned by Luna Leopold, and logical regions, which are based on entirely different Leopold's proposal for an international "Vigil N et- criteria. The unit encompassesa large part of a region work," reflect the need to couple watersheds with eco- of nutrient poor soils and high quality lakes and logical regions (Leopold,1962a). Although Leopolddid forests in northwestern Wisconsin and northeastern not use the terms ecosystemsand ecoregionsspecifi- and central Minnesota, as well as sections of several cally (those terms were not as commonlyused in the formerly prairie but now largely agricultural regions late "50s" and early "60s" as they are now), the -the Corn Belt in southern Minnesota, the flat Lake implication is clear. Leopold recognizedthe need for a Agassiz Plain (Red River Valley), and the semi-arid systemof regionally representative, small watersheds prairie pothole country of North Dakota. Hence, not for which "field observations of the same kind in only do hydrologic units lack spatial correspondence many placeswould improve our understanding of the to areas within which there is similarity in the quali- hydrologic and biologic aspects" of ecosystems ty and quantity of surface waters and the factors (Leopold, 1962a). These measurements he stated associated with spatial differences in quality and would "enhance our opportunity to distinguish quantity, most are not true topographicwatersheds. betweenthe effects of man and nature on the environ- ment" (Leopold,1962b). The key element of Leopold's vision was that eachof these small watersheds"would be chosen to represent a typical area in a general

JAWRA 942 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Distinguishing BetweenWatersheds and Ecoregions region -that is typical in its general geology,vegeta- resource management are unfortunately often divi- tion, slopes, topography, and land use" (Leopold, sive, arguing for one structure over the other. Howev- 1962a). er, although the two frameworks have very different Regrettably, the eventual design of the National purposes,they can be used together to effectivelyhelp Network of Hydrologic Bench Marks did not ade- assessand manage environmental resources.In areas quately result in a systemof regionally representative where watersheds are relevant and can be defined, watersheds.Leopold's vision regarding regionalrepre- both ecoregions and watersheds are necessary for sentativeness differed from that of Langbein and developing a system of regional reference sites Hoyt (1959), who originally proposed the network, enabling an understanding of the attainable quality, and that vision was not reflected in the subsequent integrity, and health of ecosystemsand their compo- design (Biesecker and Leifeste, 1975; Briggs, 1978). nents. A true ecosystemapproach for reaching these Whereas Leopold stressed that the bench marks be objectives requires consideration of the mosaic of typical of the regions they represent, yet relatively biotic and abiotic componentsin both the aquatic and undisturbed, Langbein and Hoyt were mainly con- terrestrial environment. Watersheds, in part, facili- cerned that the bench marks be located in places tate this meshing. For those areas where watersheds where anthropogenicimpacts and future modification cannot be defined, where they are of little relevance would be minimal. "National and state parks or public (e.g., in xeric areas), and where there are few if any lands purposely withdrawn from entry" were to be streams (e.g., parts of the Prairie Pothole region and consideredfirst for selectionof stations (Langbeinand the western part of the Nebraska Sand Hills), rele- Hoyt, 1959). Because many of our parks contain vant representative "areas" rather than watersheds anomalouslandscapes and were selectedfor preserva- must be chosen.For the bulk of the country, however, tion due to these characteristics rather than that they ecoregionalreference sites will comprise watersheds. typified the larger ecological regions they occupy, We therefore define these reference sites as water- many of the bench mark stations are also anomalous. sheds or areas that are representative of the ecore- Crater Lake in the Cascade Mountains is probably gions they occupy,but that are relatively unimpacted the best example of a bench mark which is unique (Hughes,1995; Omernik, 1995;Hughes et al., 1986). and not at all representative of the larger region it Watershedsand ecoregionscan be complementary occupies. The North Fork of the Quinault River in tools. The characteristics of the portions of large Washington is another example. The headwaters of watersheds that cover more than one ecoregionwill the Quinault are in the Olympic Mountains where be different in each of the ecoregionsthey cover(Bai- glaciers produce milky colored streams atypical of ley, 1995a). For example, the portions of the Platte streams in the lower elevation Coast Range where River watershed that are in the largely forested there are generally mild winter conditions and no RockyMountains contribute differently to the quality glaciers. Streams draining the high Olympic Moun- and quantity of the Platte River than portions in the tains, although situated in the Coast Range, tend to much drier Great Plains. Depending on the ecoregion be more like those in the North Cascade Range of their watershedsare within, streams flowing into the Washingtonand British Columbia. Platte have very different thermal characteristics, The National Network of Hydrologic Bench Marks gradients, aeration, and resultant biota (Figure 4). also missed its intended goal regarding coverageof Watershedsthat are within similar ecoregionstend to sites and maximum watershed size.The eventualnet- be similar to one another and different from those work contained considerablyless than the 100 water- of other ecoregions (Omernik and Griffith, 1991). sheds thought necessary by Langbein and Hoyt Although no two watersheds are alike, regarding (1959), and the size criteria that each watershed be their quality and quantity of water and mosaic of less than 100 square miles was not followed. Bieseck- ecosystemcomponents, the variation in characteris- er and Leifeste (1975)reported the network to contain tics between watersheds within the same ecoregion 57 stations and Briggs (1978) reported 51, of which will tend to be less, or of a similar nature, as com- only 39 drained watersheds less than 100 square pared to other ecoregions. For example, within the miles. United States, all watersheds that are completely within the Great Plains are similar to one another, as comparedto those of the easterntemperate forests, as comparedto those of the xeric basin and range coun- USING WATERSHEDSAND try in the west, and so on. Each of these broad ecore- ECOREGIONSTOGETHER gions contains a particular mosaic of geographic phenomena(including climate, physiography,geology, Discussions of the appropriateness of watershed soils, and land cover) that cause or reflect differences and ecoregion frameworks for environmental in capacities and potentials of ecosystems. These

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 943 JAWRA Omernik and Bailey

Figure4. Locationsof Streams That SupportWarm-Water and Cold.WaterFish and StreamsThat Do Not Support Fishery (asmapped by Funk, 1970)within the Platte River Watershed(as mapped by the U.S. GeologicalSurvey, 1979).

ecoregions can also be used for the assessment of Maxwell et at. (1995) attempted to address a way ecosystemhealth, quality, and integrity. Lower level, watersheds or basins and ecoregions (which they smaller ecoregionsthat are nested within the larger termed geoclimatic units) could be used in a comple- ecoregionsalso contain similarities in these phenome- mentary fashion. In our view, there are at least two na, but they generally contain less variability. Hence flaws encompassed in their suggestion: first, they watersheds that are completely within each of these have selected basins or watersheds as the primary smaller ecoregionswill be more similar to one another meaningful integrating spatial units; and second, than those of other parts of the larger ecoregions they recognized the importance of ecoregions at only within which they are nested.Watersheds that strad- one hierarchical level. The example used by Maxwell dle two different ecoregionswill reflect the character- et at. (1995) centered on defining fish distributions istics of both ecoregions. and their general habitats, but the overall purpose of

JAWRA 944 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Distinguishing BetweenWatersheds and Ecoregions the framework they proposed was much broader, are relative to disturbed, relatively unimpacted, or encompassing the classification of aquatic systems historic conditions. Basin wide management strate- and ecological analysis of these systems. Although gies and planning, while an improvement over deal- they stated "watersheds can be characterized and ing with pollution problems on point or site bases, assessed on the basis of the geoclimatic setting in may be misleading in that the plans and strategies which they are found," which could be interpreted as often do not fit the regional potentials of the land and being in agreement with the explanation we present- water, unless ecoregionaldifferences are determined. ed in the precedingparagraph, the spatial framework Basin or watershed studies, where the interest is on they recommendedis clearly one of hydrologic units the quality or quantity of water at a specific point, or basins, and not ecoregions. Ecoregions as useful can be conductedrelatively quickly if ecoregionsand spatial units were only recognizedat what they term ecoregionalreference conditions have been previously the "subbasin" level. Here, Maxwell et al. (1995)stat- determined. ed that basins "may be divided into subbasinsbased To developa network (regional, national, or inter- on physiographic criteria that define different national) of ecoregionalreference sites requires build- physical-chemical habitat patterns inhabited by dis- ing on Leopold's(1962a, 1962b)vision of the National tinct species groups." Watershed patterns, they Network of Hydrologic Bench Marks, with the key claimed, are defined by similarity in settings and fea- aspects being regional representativeness and rela- tures using information on geoclimatic pattern, zoo- tive lack of disturbance. A third critical elementis the geographic pattern, morphology, and disturbance. number of sites necessary.In order to provide an ade- However, as we have previously noted, areas within quate understanding of the natural variability within which there is similarity in geoclimatic and biologic and between ecologicalregions, determine the relative characteristics, as well as anthropogenic distur- importance of human impacts, and developscenarios bances, seldomcorrespond to topographicwatersheds. regarding attainable quality of ecosystemsand their To illustrate the way watersheds and ecoregions components,a large number of reference sites will be can be used together we have adapted a figure from needed. Sets of watersheds/areas that are, for the Maxwell et at. (1995), and have substituted the term most part, completely within ecoregionswill be need- ecoregion for subbasins (Figure 5). In general, the ed for each ecoregion.The number and size of refer- aquatic and terrestrial characteristics of watersheds ence watersheds or areas needed will vary from one C, D, and E will be similar to one another and dissim- ecoregionto another and are dependent on the size ilar to characteristics of watersheds A and B as well and heterogeneity of each ecoregion(Hughes, 1995; as those of F, G, and H. Because the Santee and Omernik, 1995,Bailey, 1991). Savannah basins cover the same ecoregions, the water quality near the mouth of both rivers can be expected to be similar, unless human impacts within the basins are dissimilar. The groups of basins shown CONCLUSION on Figure 5 comprise one aquatic ecologicalunit (the South Atlantic Subregion) of the framework suggest- The current interest in adopting a more holistic ed by Maxwell et al. (1995). However, from an ecosys- ecosystem approach to resource assessment and man- tem management standpoint, this unit has little agement has agencies, programs, and individuals meaning. Information for aquatic resource needs, scurrying to map shelves for spatial frameworks to such as water quality standards or biological criteria, help implement the approach. The recent U.S. Gener- that are based on information gathered in reference al Accounting Office (GAO) report on ecosystem man- watersheds/areas in the Coastal Plain will not be agement emphasized the need for delineating applicable in the Piedmont or Blue Ridge portions of ecosystem boundaries and noted that management this hydrologic unit. However,this information canbe must be along ecological rather than political or extrapolated elsewherein the sameecoregion, regard- administrative boundaries (U.S. GAO, 1994). In com- less of whether it fits within the South Atlantic Sub- menting on the GAO report, Jack Ward Thomas, then region. Chief of the U.S. Forest Service, stressed that the pro- Hence, although the watershed framework must cess of agreeing on an ecological classification system often be used to determine regional reference condi- "should not be one of selecting a watershed approach tions, resource management agencies must first rec- over an ecoregion approach, but how to best use ognize the ecoregions within which ~cosystemsand these tools to assess the condition of the nation's the quality and quantity of environmental resources ecosystems." Although ecoregions and watersheds are similar. An ecoregionframework, versus one of are intended for different purposes, they can be com- basins or watersheds, provides a more effective tool plementary. Ecoregions provide the spatial frame- for extrapolating reference conditions, whether they work within which the quality and quantity of

JOURNAl OF THE AMERICAN WATER RESOURCES ASSOCIATION 945 JAWRA Omernik and Bailey

LocatorMap \#

~

~

-Cape Fear Basin

--Pee Dee Basin

SanteeBasin

EdistoBasin

SavannahBasin

Ogeechee Basin

"'" AltamahaBasin -Blue Ridge Ecoregion D PiedmontEcoregion !ill CoastalPlain Ecoregion Wid Selecte;dwatersheds within ecoreglons

Figure 5. Ecoregions,Basins, and SelectedWatersheds Within the SouthAtlantic Subregionas Defined by Maxwell et ai. (1995).(Adapted from Maxwell et ai., 1995.) ~.JAWRA 946 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Distinguishing BetweenWatersheds and Ecoregions environmental resources,and ecosystemsin general, Black, P. E., 1997.Watershed Functions. Journal of the American can be expectedto exhibit a particular pattern. Where Water ResoureesAssociation 33(1):1-11. Born, S. M. and W. C. Sonzogni,1995. Integrated Environmental watershedsare relevant and can be defined, they are Management: Strengthening the Conceptualization. Environ- necessary for studying the relationships of natural mental Management19(2):167-181. and anthropogenicphenomena with water quality, as Briggs,J. C., 1978.Nationwide Surface Water Quality Monitoring well as for providing the spatial unit for reference Networks of the U.S. GeologicalSurvey. In: Establishment of areas within ecoregions at all scales. Where basins Water Quality Monitoring Programs, L. G. Everett and K. D. (large watersheds) are needed to determine the con- Schmidt (Editors). American Water Resources Association, Herndon,Virginia, pp.49-57. tributions to the quality and quantity of water at a Brown,D. D., andT. C. Brown, 1994.An Evaluation of Land Classi- specificpoint, ecoregionalreference site data can help fication Systems for Large-Scale Water Quality Assessments. in making these determinations. When used together In: Effects of Human-InducedChanges on Hydrologic Systems. correctly, these two frameworks provide a powerful Proceedingsof the AWRA 1994 Annual Summer Symposiumof the American Water ResourcesAssociation. AWRA Technical mechanism necessary for meeting the specific PublicationSeries TPS-94-3, Herndon, Virginia, pp.939-948. resource managementgoals such as those outlined in Callicott,J. B., 1995.A Reviewof SomeProblems with the Concept the Clean Water Act, as well as the broader ecosystem of EcosystemHealth. EcosystemHealth 1(2):101-112. managementobjectives currently being soughtregion- Cannon, J. Z., 1994. Geographic Approaches to Environmental ally, nationally, and internationally. Management:Bioregionalism Applied. In: Proceedingsof Water- shed '93, a National Conference on Watershed Management, Alexandria, Virginia, pp. 281-286. Cairns, J., 1994. The Current State of Watershedsin the United ACKNOWLEDGMENTS States: Ecologicaland Institutional Concerns.In: Proceedingsof Watershed'93, a National Conferenceon Watershed Manage- We gratefully acknowledgeDouglas Brown and Glenn Griffith ment, Alexandria, Virginia, pp.11-17. for their assistancein reviewing the literature on watershedsand Ciriacy-Wantrup,S. V., 1961. Philosophy and Objectivesof Water- ecoregions.We also thank Gerald McMahon,Glenn Griffith, Walter shedPolicy. In: Economicsof WatershedPlanning, G. Tolleyand Russell,Denis White, and three anonymousreviewers for their crit- F. Riggs (Editors). Iowa State University Press,Ames, Iowa, pp. ical commentson earlier drafts of the manuscript, and we thank 1-12. SandraAzevedo for her help in preparing the graphics. Coastal America, 1994. Toward a Watershed Approach: A Frame- work for Aquatic EcosystemRestoration, Protection,and Man- agement.Coastal America Partnership Project. Commissionfor Environmental CooperationWorking Group,1996. LITERATURE CITED An Ecological Framework of North America. Commissionfor Environmental Cooperation,Montreal, Quebec,Canada. Allen, T. F. H., B. L. Bandurski, and A. W. King, 1993.The Ecosys- Davis, W. S.,B. D. Snyder,J. B. Stribling, and C. Stoughton,1996. tem Approach: Theory and EcosystemIntegrity. International Summary of State Biological AssessmentPrograms for Streams Joint Commission,United States and Canada,ISBN 1-895085- and WadeableRivers. EPA 230-R-96-007,U. S. Environmental Protection Agency,Office of Policy, Planning and Evaluation, 78-0. Armitage, D., 1995. An Integrative MethodologicalFramework for Washington,D.C. Sustainable Environmental Planning and Management.Envi- EcologicalStratification Working Group, 1995. A National Ecologi- cal Framework for Canada. Agriculture and Agri-food Canada, ronmental Management19(4):469-479. Bailey,R. G., 1984.Testing an EcosystemRegionalization. Journal Centre for Land and Biological ResourcesReseareh, and Envi- of Environmental Management19:239-248. ronment Canada, State of the Environment Directorate. Bailey,R. G., 1987.Mapping Ecoregionsto Manage Land.In: 1987 Ottawa/Hull, Canada,125 pp. Yearbookof Agriculture. U.S. Department of Agriculture. Wash- Fitzsimmons, A. K., 1996. Sound Policy or Smoke and Mirrors: Does EcosystemManagement Make Sense?Water Resources ington, D.C., pp 82-85. Bailey, R. G., 1991. Design of EcologicalNetworks for Monitoring Bulletin 32(2):217-227. Global Change.Environmental Conservation18(2):173-175. Forest Ecosystem Management Assessment Team (U.S.), 1993. Bailey, R. G., 1995a. EcosystemGeography. Springer-Verlag, New Forest EcosystemManagement: An Ecological,Economic, and SocialAssessment: Report of the Forest EcosystemManagement York, New York. Bailey, R. G., 1995b.Descriptions of the Ecoregionsof the United AssessmentTeam/United States Department of Agriculture, States (2nd Edition). MiscellaneousPublication No. 1391,Map ForestService et aI. scale 1:7,500,000,U.S. Department of Agriculture, Forest Ser- Franklin, J. F., C. S. Bledsoe,and J. T. Callahan, 1990. Contribu- tions of the Long-TermEcological Researeh Program. BioScience vice, 108pp. Barnes,B. V., 1993. The LandscapeEcosystem Approach and Con- 40(7):509-524. servation of EndangeredSpaces. Endangered Species UPDATE Funk, J. L., 1970.Warm-Water Streams. In: A Century of Fisheries in North America, N. G. Benson(Editor). American Fisheries 10(3and 4):13-19. Bazata, K., 1991Nebraska Stream ClassificationStudy. Nebraska Society,Washington, D.C., pp.141-152 Department of Environmental Control, Lincoln,Nebraska. Governmentof Canada,1991. The State of Canada'sEnvironment. Biesecker,J. E. and D. K. Leifeste, 1975,Water Quality of Hydro- Ministry of the Environment, Ottawa, Canada. Grumbine,R. E., 1994. What is EcosystemManagement? Conser- logic BenchMarks -An Indicator of Water Quality in the Natu- ral Environment. Geological Survey Circular 460-E, U.S. vation Biology8(1):27-38. Department of Interior, GeologicalSurvey. Reston, Virginia, 21 Haeuber, R. and J. Franklin, 1996. Perspectives on Ecosystem Management.Ecological Applications 6(3):692-693. pp.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 947 JAWRA Omernikand Bailey

Holling, C.S., 1994. An Erologist View of the Malthusian Conflict. Maxwell,J. R., C. J. Edwards,M. E. Jensen,S. J. Paustian,H. Par- In: Population, Economic Development,and the Environment, rot, and D. M. Hill, 1995.A Hierarchical Framework of Aquatic K. Lindahl-Kiessling and H. Landberg(Editors). Oxford Univer- Ecological Units in North America (Nearctic zone). General sity Press,New York, New York, pp 79-104. TechnicalReport NC-176,United States Department of Agricul- Hornig, C. E., C.W. Bayer,S. R. Twidwell,J. R. Davis,R. J. Klein- ture, Forest Service,North Central Forest Experiment Station, sasser,G. W. Linam, and K. B. Mayes, 1995. Developmentof St. Paul, Minnesota,72 pp. Regionally Based Biological Criteria in Texas. In: Biological McDonnell, M. J. and S. A. Pickett (Editors), 1993. Humans as Assessmentand Criteria: Tools for Water ResourcePlanning Components of Ecosystems: The Ecology of Subtle Human and DecisionMaking. W. Davis and T. Simon (Editors). Lewis Effects and Populated Areas. Springer-Verlag,New York, New Publishers,Boca Raton, Florida, pp 145-152. York. Hughes, R. M., 1995. Defining Acceptable Biological Status by Montgomery,D. R., G. E. Grant, and K. Sullivan, 1995.Watershed Comparing with Reference Conditions. In: Biological Assess- Analysis as a Framework for Implementing EcosystemManage- ment and Criteria: Toolsfor WaterResource Planning and Deci- ment. WaterResources Bulletin 31(3):369-386. sion Making, W. Davis and T. Simon(Editors). Lewis Publishers, National ResearchCouncil, 1992. Restoration of Aquatic Ecosys- BocaRaton, Florida, pp.31-47. tems: Science,Technology, and Public Policy. National Academy Hughes, R. M., S. A. Heiskary, W. J. Matthews, and C. 0 Yoder, Press,Washington, D.C. 1994.Use of Ecoregionsin BiologicalMonitoring. In: Biological Omernik, J. M., 1977. Nonpoint Source-Stream Nutrient Level Monitoring of Aquatic Systems,S. L. Loeb and A. Spacie (Edi- Relationships: A Nationwide Study. EPA/600/3-77/105,U.S. tors). Lewis Publishers,Boca Raton,Florida, pp.125-151. Environmental Protection Agency, Environmental Research Hughes,R. M., D. P. Larsen, and J. M. Omernik, 1986,Regional Laboratory,Corvallis, Oregon. Reference Sites: A Method for Assessing Stream Potentials. Omernik,J. M., 1995.Ecoregions: A Spatial Framework for Envi- Environmental Management10(5):629-635. ronmentalManagement. In: Biological Assessmentand Criteria: Hughes,R. M., E. Rexstad,and C. E. Bond, 1987.The Relationship Tools for Water Resource Planning and Decision Making, of Aquatic Ecoregions, River Basins, and Physiographic W. Davis and T. Simon(Editors). Lewis Publishers,Boca Raton, Provinces to Ichthyogeographic Regions of Oregon. Copeia Florida, pp.49-62. 1987(2):423-432. Omernik, J. M., A. R. Abernathy, and L. M. Male, 1981. Stream Hughes, R. M. and J. M. Omernik, 1981. Use and Misuse of the Nutrient Levels and Proximity of Agricultural and Forest Land Terms Watershedand Stream Order. American Fisheries Soci- to Streams: SomeRelationships. Journal of Soil and Water Con- ety, Warmwater Stream Symposium,Bethesda, Maryland, pp. servation36(4):227-231. 320-326. Omernik,J. M. and C. F. Powers,1983. Total Alkalinity of Surface Inkley, D. B. and S. H. Anderson,1982. Wildlife Communitiesand Waters -A National Map. Annals of the Associationof Ameri- Land ClassificationSystems. In: Transactions47th North Amer- can Geographers73(1):133-136. ican Wildlife and Natural Resource Conference, K. Sabol OmernikJ. M. and G. E. Griffith, 1986. Total Alkalinity of Surface (Editor). Wildlife Management Institute, Washington, D.C. Waters:A Map of the WesternRegion. Journal of Soil and Water pp.73-81. Conservation41(6):374-378. Karr, J. R., 1991. Biological Integrity: A Long-NeglectedAspect of Omernik,J. M., C. M. Rohm, S. E. Clark, and D. P. Larsen, 1988. Water ResourceManagement. EcologicalApplications 1(1):66- SummerTotal Phosphorusin Lakes: A Map of Minnesota,Wis- 84. consin, and Michigan, Environmental Management 12(2):815- Kay, J. K. and E. Schneider, 1994. Embracing Complexity: The 825 Challengeof the EcosystemApproach. Alternatives 20(3):32-39. Omernik,J. M. and G. E. Griffith, 1991. EcologicalRegions Versus Langbein,W. B. and W. G. Hoyt, 1959.Water Facts for the Nation's Hydrological Units: Frameworks for Managing Water Quality. Future: Uses and Benefits of Hydrologic Data Programs. The Journal of Soil and Water Conservation46(5):334-340. Ronald Press Company,New York, NewYork. Parsons,J. P., 1985. On "Bioregionalism" and "Watershed Con- Larsen, D. P., D. R. Dudley, and R. M. Hughes, 1988. A Regional sciousness."The ProfessionalGeographer 37(1):1-6. Approach for AssessingAttainable Surface Water Quality: An Pflieger, W. L., 1971. A Distributional Study of Missouri Fishes. Ohio Case Study. Journal of Soil and Water Conservation University of Kansas Publication, Museum of Natural History 43(2)171-176. 20:225-570. Larsen, D. P., J. M. Omernik, R. M. Hughes, C. M. Rohm,T. R. Plotnikoff, R. W., 1992.Timber/Fish/Wildlife Ecoregion Bioassess- Whittier, A. J. Kinney, A. L. Gallant, and D. R. Dudley, 1986. ment Pilot Project. Washington Department of Ecology. The CorrespondenceBetween Spatial Patterns in Fish Assem- Olympia,Washington. blages in Ohio Streams and Aquatic Eroregions.Environmental Poff,N. L. andJ. DAllan, 1995.Functional Organization of Stream Management10: 815-828. Fish Assemblagesin Relation to Hydrologic Variability. Ecology Leopold,L. B., 1962a.The Man and the Hill. GeologicalSurvey Cir- 76(2):606~27. cular 460-A. United States Department of the Interior, Geologi- Poff, N. L. and J. V. Ward, 1989.Implications of StreamflowVari- cal Survey,Washington, D.C. ability and Predictability for Lotic Community Structure: A Leopold, L. B., 1962b. A National Network of Hydrologic Bench RegionalAnalysis of StreamflowPatterns. Canadian Journal of Marks. GeologicalSurvey Circular 460-B.United States Depart- Fisheriesand Aquatic Sciences46:1805-1818. ment of the Interior, GeologicalSurvey, Washington, D.C. Risser,P. G.,J. Lubchenco,N. L. Christenson,P. J. Dillon, L. D. Lotspeich, F. B., 1980. Watersheds as the Basic Ecosystem:This Gomez,D. J. Jacob,P. L. Johnson,P. Matson,N. A. Moran, and ConceptualFramework Provides a Basis for a Natural Classifi- T. Rosswall, 1993. Ten-Year Review of the National Science cation System.Water ResourcesBulletin 16(4):581-586. Foundation Long-TermEcological Research (LTER) Program. Lyons, J., 1989. CorrespondenceBetween the Distribution of Fish National ScienceFoundation, Arlington, Virginia. Assemblagesin WisconsinStreams and Omernik's Ecoregions. Rohm,C. M., J. W. Geise,and C. C. Bennett, 1987.Evaluation of an AmericanMidland Naturalist 122(1):163-182. Aquatic EcoregionClassification of Streams in Arkansas. Jour- nal of FreshwaterEcology 4:127-140.

JAWRA 948 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATK)N Distinguishing BetweenWatersheds and Ecoregions

Rohm, C. M., J. M. Omernik, and C. W. Kiilsgaard, 1995.Regional Patterns of TotalPhosphoros in Lakes of the NortheasternUnit- ed States.Lake and ReservoirManagement 11(1):1-14. Rowe,J. S., 1990.Home Place: Essayson Ecology.Canadian Parks and WildernessSociety. Henderson Book Series No.12,NeWest Publishers Ltd., Edmonton,Alberta, Canada. Rowe, J.S., 1992. Regionalization of Earth Spaceand its Ethical Implications. Unpublished paper presented at the Landscape Ecology Symposium,Corvallis, Oregon,April 10,1992. Seaber,P. R.,F. P. Kapinos, and G. L. Knapp, 1987.Hydrologic Unit Maps. U.S. GeologicalSurvey Water-SupplyPaper 2294,United States Department of the Interior, GeologicalSurvey, Denver, Colorado. Spindler, P., 1996. Using Ecoregionsfor Explaining Macroinverte- brate Community Distribution Among Reference Sites in Arizona, 1992. Final Report, Arizona Department of Environ- mental Quality, Water Quality Division, Phoenix,Arizona. U. S. Department of the Interior, Fish and Wildlife Service,1995. An Ecosystem Approach to Fish and Wildlife Conservation. ConceptualDocument. U.S. Fish and Wildlife Service,Washing- ton, D.C. U. S. Department of Interior, GeologicalSurvey, 1979. Accounting Units for the National Water Data Network. U.S. Geological Survey,Washington, D.C. U. S. Department of Interior, GeologicalSurvey, 1994. Hydrologic Unit Map -1994: State of Tennessee.U.S. GeologicalSurvey, Reston,Virginia. U. S. Environmental Protection Agency, 1995. Watershed Protection: A Statewide Approach. EPA841-R-95-004,U. S. Environmental Protection Agency,Office of Water, Washington, D.C. U. S. Environmental Protection Agency, 1996a. Watershed ApproachFramework. EPA840-S-96-001,U. S. Environmental ProtectionAgency, Office of Water,Washington, D.C., 16pp. U. S. Environmental Protection Agency, 1996b. The Watershed Approach and Reinvention -"Learn From the Past, Look Toward the Future". In: EPA WatershedEvents: A Bulletin on Sustaining Aquatic Ecosystems.U. S. Environmental Protection Agency,Office of Water,Washington, D.C. U. S. Environmental Protection Agency,1997. The Index of Water- shed Indicators. EPA841-R-97-010,U. S. Environmental Protec- tion Agency,Office of Water, Washington,D.C. U. S. General Accounting Office, 1994. EcosystemManagement: Additional Actions Needed to Adequately Test a Promising Approach. GAO/RCED-94-111,U.S. GeneralAccounting Office, Washington,D.C., 87 pp. Water Environment Federation, 1992. Water Quality 2000: A National Water Agenda for the 21st Century. Water Environ- ment Federation,Alexandria, Virginia. Whittier, T. R., R. M. Hughes, and D. P. Larsen, 1988.The Corre- spondenceBetween Ecoregions and Spatial Patterns in Stream Ecosystems in Oregon. Canadian Journal of Fisheries and Aquatic Sciences45(7): 1264-1278. Wiken, E. B. and K. Lawton, 1995. North American Protected Areas: An EcologicalApproach to Reporting and Analysis. The GeorgeWright Forom 12(1):25-34. Wilton, T. F., 1996. Final Report: Pilot Study of Biocriteria Data Collection Procedures for Wadeable Streams in Iowa. Iowa Department of Natural Resources,Des Moines,Iowa. Yoder, C.O., 1995. Policy Issues and ManagementApplications of Biological Criteria. In: Biological Assessment and Criteria: Tools for Water Resource Planning and Decision Making, W. Davis and T. Simon (Editors). Lewis Publishers,Boca Raton, Florida, pp. 49-62.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 949 JAWRA