sign in sign up
<<
Home , Ecoregion

ROBERT G. BAILEY within a geographical framework of regions or zones. This paper ex. Rocky Mountain and Range Experiment Station plains the basis for the regions delineated on the 1976 map Ecore- USDA Forest Service gions of the United States. Four ecological levels are discussed- Fort Collins, Colorado 80526 domain, division, province, and section-based on climatic and ve- getational criteria. Statistical tests are needed to verify and refine ABSTRACT / As a means of developing reliable estimates of eco. map units. system , classification needs to be placed

Ecoregions are large of regional extent that tem classesmust be developed. These relationships are called contain a number of smaller ecosystems.They are geographical rules (Davis 1980). Rules take many forms, ranging from zones that represent geographical groups or associations of simple experienced-basedjudgments to multivariate regression similarly functioning ecosystems.Regional boundaries may be models to complex mathematical simulations. Application of delineated on the basis of detailed information about ecosys- these rules is firmly based upon conceptsof transfer by analogy. tems at the site level, or by analysis of the environmental factors Thus, the rules necessaryfor estimation of productivity are ex- that most probably acted as selective forces in creating variation trapolated from experimental sites or from management expe- in ecosystems. rience to analogous areas defined by classification. Using the latter approach, Bailey (1976) constructed a Such methods do not require any prior knowledge of 1:7,500,000-scale map of the United States' ecoregions,with a functional relationships between site parameters and the spe- brief narrative text that described the approach and develop- cific form of biological production, although such criteria may ment of the map. Bailey's map was developed as an inland be incorporated. The land as a holistic ecosystemor separate counterpart to the regionalization of marine and estuarine major component (for example, ) is classified on the basis of systems proposed by Cowardin and others (1979). An accom- a number of observedand measurable characteristics. panying manual (Bailey 1980) describes and illustrates the Subsequent evaluation is based on the hypothesis that all regions shown on the map. However, a detailed explanation of replications of a physically defined and characterized ecosys- the basis for the delineation of these regions was not included. tem or component class will respond in a similar way to This topic is addressedhere. management for any specified level of use. This hypothesis has The regionalization system described here is based on been questioned by a number of research workers on the published information. This paper does not report on a grounds that correlations between ecosystemand component data-supported study; but is, instead, an exposition of a classesand production are generally low. rationale for data synthesis using existing small-scale maps. It One of the ways to establish reliable site-production rela- presents one system for recognizing and displaying ecosystems tionships is to identify homogeneousgeographical strata where with respect to geographical distribution. This system and similar ecosystemshave developed on materials having similar others are undergoing rigorous evaluation to determine the properties (Rowe 1962). For example, similar sites (those most appropriate procedure for delineating ecosystemassocia- having the same , slope, parent material, and drainage tions (Driscoll and others 1981). characteristics) may be found in several climatic regions. Within a region, these sites will support the same communities, but in other regions vegetation on the sites will be Why Ecoregions Are a Necessary Classification different. Thus, beach ridges in the region support The purpose of ecological land classification is to divide the low-growing shrubs and forbs, whereas beaches in the subarc- landscapeinto variously sized ecosystemunits that have signifi- tic region usually have dense growth of black spruce or jack cance both for development of resourcesand for conservation of pine. display similar trends, as the kind and development environment. More specifically, such units are the base for of soil properties vary from region to region on similar sites. estimating ecosystem productivity and the probable responses The depth of the thawed layer and the form and kind of surface to management practices. To make such estimates, relation- expression of permafrost on similar sites also vary between ships between the needed production information and ecosys- regions but remain relatively constant on comparable sites within a region. Ecoregions, therefore, define broad areas KEY WORDS: Ecoregions; Ecosystems; Ecological land classification; Map- where one can expect to find the same kinds of vegetation and ping; Site production soil associations on similar sites.~

Environmental Management, Vol. 7, No.4, pp. 365-373 @ 1983 Springer-Verlag New York Inc 366 A.G. Bailey

Ecoregions have at least two important functions for man- to climate is termed zonal. The term azonal describes processes agement. First, a map of such regions suggestsover what area or features that occl,ir in several zones. For example, the productivity relationships (rules) derived from experiments are not associatedwith a particular climatic zone. and experience can be applied without too much adjustment. Efforts to divide the world into ecological regions have been Second, they provide a geographical framework in which based primarily on the distribution of climate-vegetation zones similar responses may be expected within similarly defined (for example, Herbertson 1905, James 1951, Biasutti 1962, sites. Methods of site classification involving such a geograph- Udvardy 1975). Recently, Walter (1977, Walter and Box ical framework have been employed with successfor over 50 1976) presented a scheme for classifying the world into a years in . The Baden-Wiirttemberg system of south- hierarchy of ecosystemsfrom a climatic viewpoint. In Russia, western as described by Spurr and Barnes (1980) is Berg (Isachenko 1973) detailed zones based on perhaps the best example. Similar methods have proliferated climate, while similar work was developed by Passarge (Troll widely in Canada. 1971) in Germany, and Galoux (Delvaux and Galoux 1962) in Belgium. Some systems for the classification of climates Criteria Used in Delineating Ecoregions (Koppen 1931, Thornthwaite 1931, 1948) seek to define climatic units that will correspond to major vegetation units. A Because information about variation among ecosystems number of authors (for example, Merriam 1898, Hopkins throughout the country is limited, a national zoning system 1938) have sought to define life zones primarily on the basis of must be based primarily on criteria that reflect causal environ- climate; the system of Holdridge (1947, Tosi 1964) employs a mental factors. For broad-scale subdivision of a into a complex classification of zones by both temperature and mois- small number of large units, the large ecological climate-zones ture conditions. present an obvious means of approach. Becausethe formation In Canada, the concept of forest ecosystemregions (there of soils and vegetation types and, to a lesser degree, of is called site regions) was developed by Hills (1960) based on determined primarily by the climate. The macroclimate is the macroclimate. Simliar work has been done in other parts of best reflection of zonality. Surface configuration, which is Canada (Crowley 1967, Burger 1976). Krajina (1965) has partly the result of internal forces, is less influenced by the delineated the biogeoclimatic zones of British Columbia. Cli- climate than either vegetation or soil; but the influence is great matic regionalization is used in biophysical or ecological land enough for the minor features to reflect the climate of the area classification throughout Canada (Wiken and Ironside 1977). where they are found. Of the variety of classifications available, the one devised by For further subclassification beyond climate, the macrofea- Crowley (1967) has been adopted as most suitable for the tures of the vegetation appear to be the appropriate criteria for current purpose. A hierarchical order is established by defining defining secondary divisions (Damman 1969, Kuchler 1973). successivelysmaller ecosystemswithin larger ecosystems(Fig- Although only a result, vegetation is important as a criterion in ure 1). First, subcontinental areas,termed domains, are identi- the delineation of geographical zones becauseit affords a very fied on the basis of broad climatic similarity, such as having dry delicate index of climate. The predominance of vegetation in climates. Climate is emphasized at the broadest level becauseof the landscape also ensures its consideration in any scheme of its overriding effect on the composition and productivity of zoning. Usually, the boundaries of vegetationally defined ;.cosystems from region to region. The domains are quite regions coincide with those of major relief units; this strength- heterogeneousand are further subdivided, again on the basis of ens the primary division. However, the surface features are climatic criteria, into divisons. The divisions correspond to more useful at lower levels, that is, for subdividing the areas having definite vegetational affinities (prairie or forest) biotically circumscribed areas. and falling within the same regional climate, generally at the The concept of climate as expressed by vegetation has been level of the basic climatic types of Koppen (1931) or of used frequently as the basis for delineating broad-scale ecologi- Thornthwaite (1931, 1948); usually the zonal soils are also cal regions. Of immense significance for the development of the related. Dry climates, for example, are separated into semiarid concept was the work of Dokuchaev (1899). This author and arid desert. A major exception to this is that the dry pointed out that natural conditions are characterized by many western side of the humid continental and subtropical climates, common features within the limits of extensive areas (zones) which is very extensive in central , is distin- and that these features change markedly in passing from one guished as subhumid prairie (Borchert 1950). zone to another. In subsequent studies, Grigor'yev and Budyko Within a division, one or several climatic gradients may (Grigor'yev 1961) established that the boundaries of geograph- affect the potential distribution of the dominant vegetation ical zones are determined to a considerable extent by climatic strata. Within the arid zone, for example, deserts that receive factors. Every feature with a distribution that broadly conforms onlv winter rain () can be distinguished from Delineation of Ecosystem Regions 367

Figure 1. A hierarchy to the fourth level for ecosystemregions within the Humid Temperate Domain. those that receive only summer rain (). The vertical arrangement of the ecological systems in moun- Within the steppe zone, a semiarid steppe (short-grass prairie) tainous areas, with their various altitudinal belts, poses a climate that has a dry summer seasonand occasional problem. Such systems do not have the same climate as the can be distinguished from an arid semidesert (sagebrush) adjacent lowlands, but they do have the same climatic regime climate that has a very pronounced drought seasonplus a short (cycle of weather phenomena). These systems do not belong to humid season. A southern (coniferous forest) climate and the zones; rather, each forms a certain ecologi- northern (forest-tundra) climate can be distinguished within cal unit in itself, which depends on the relations within the the Subarctic Division of the Polar Domain. zone(s) in which it is located. This results in a certain zonal These climatic subzonesare evident as areas covered by the relationship, which also affects the vertical climate-vegetati°n.- same dominant vegetation on sites supporting climax vegeta- soil zonation within the mountain range. The earlier view that tion (Weaver and Clements 1938). Such sites are uplands, with the upward sequence of altitudinal belts in the mountains well-drained surface, moderate surface slope, and we!l-devel- represents a short repetition of the zonal arrangement from oped soils. The climax vegetation corresponds to the major south to north (in the Northern Hemisphere) rests only on very formation (for example, deciduous forest) characterized superficial comparisons. The climate of an altitudinal belt in by uniformity both in physiognomy and in the structure of the the mountains is always different from the climates of a climax type. Each climatic subzone comprises both the climax northerly climate zone. The latitude is different, as are the day formation and all the successionalstages within its geograph- length, the solar declination, the length of the season,and the ical area. precipitation pattern, the level of which generally increases Divisions are subdivided into provinces on the basis of the upward in the mountains as a result of the ascending rain. climax plant formation that geographically dominates the Between the individual altitudinal belts, there is a lively upland area of the province. Boundaries drawn on the basis of exchange of materials: water and the products of erosion move this broad criterion are often coincident with the major soil down the mountains; updrafts and downdrafts carry dust and zones which, therefore, serve as supplemental criteria for pieces of organic matter; animals can move easily from one belt delimiting provinces. Provinces are further subdivided into into the next; seeds are easily scattered by the wind or sections on the basis of differences in the composition of the propagated by birds. The vertical belts, as a result, are not climax vegetation type. Thus, the summer green deciduous always as sharply separated from each other as are the climatic forest of eastern North America is fairly homogeneous in its zones. The geographical area over which a sequence of belts main structural features from east to west and north to south; extends is consideredto be a large ecological unit, designated by but, five discrete climax associations can be recognized on the the term highland province. In this system, the montane forest basis of floristic composition: oak-hickory, beech-maple, belt is not treated as a separate zone. The montane belt is only Appalachian oak, mixed mesophytic, and maple-basswood. one member of the total sequenceof altitudinal belts. Montane The sections correspond generally to the potential natural belts in mountainous areas of different climatic zonesare just as vegetation types of Kuchler (1964). distinct from one another as the montane belt is from other 368 R. G. Bailey

altitudinal belts in the same zone. For example, the montane nounced or significant compared with adjacent areas. A line coniferous belt appears in the subarctic zone as spruce-fir drawn in this way is only an approximation showing where forest, while in the steppe it is represented by a fir and pine most of the changes take place; many local variations become forest. obvious when smaller areas are mapped in more detail. For this The structure of the stratification (that is, the number and reason, a particular soil name or vegetation type may be sequenceof strata or belts) within a single zone is usually quite applied to two adjacent ecoregions when site conditions are uniform. The variant types of vertical stratification are distin- similar in these two regions. This situation usually exists only guished, in the main, not by the number of belts in the sequence within a limited area on either side of a regional boundary line; but by the existence of certain plant species in individual belts. as the distance from the boundary line increases, climatic Thus, all mountain systemsthat lie within the steppezone have change is sufficient to produce major, significant differences in a uniform system of belts, yet they reveal certain partial similar site conditions. differences in character. In the forest of the northern Rocky Because vegetation serves an an indicator of climate, its Mountains, the montane belt consists of grand fir-Douglas-fir geographical distribution servesas a primary recognition crite- forest; within the central region, only of Douglas-fir; and in the rion for regional boundaries. It is important to note again that south, of ponderosa pine-Douglas-fir forest. The term high- usually variation in vegetation is essentially continuous, and land section is applied to these different areas (forest types). degreesof difference form a full spectrum from insignificant to Highlands are distinguished where, becauseof the influence complete. Consequently, the placement of the boundaries is of altitude, the climatic regime differs substantially from that in inherently subjective. adjacent lowlands so that there is complex vertical climate- The vegetation of the United Statesis described and mapped vegetation-soil zonation. As a result, the levels appear as in "Potential Natural Vegetation of the Conterminous United follows: States" (Kuchler 1964) at a scale of 1:3,168,000. Kuchler's map was used in conjunction with maps of climatic regions as Domain the base map for delineating ecoregions. As a first step in regionalizing Kuchler's map, all lands were divided into Division lowlands and highlands. The pattern of montane vegetation as I I exhibited on the map was sufficient to identify these areas as a Province- Highland Province first approximation. These areas were further refined by reference to Hammond's (1964) land-surface form map (scale 4. Lowland Section 4. Highland Section 1:5,000,000). Hammond's high mountains (less than 50"10of All levels are needed to denote a particular area within the area gently sloping with local relief over 900 m) correlate well complete hierarchical system. For example, the Colorado with the patterns of montane vegetation shown on Kuchler's Piedmont is part of the Grama-Buffalograss Section of the map. However, the vegetation boundaries and the orographic Great Plains Short-grass Prairie Province within the Steppe boundaries do not always coincide in detail. For example, the Division of the Dry Domain. The Cascade Range represents vegetation of the basal plain commonly does not stop at the foot the Silver Fir-Douglas-fir Forest Section of the Pacific Forest of the mountain but extends up the lower slopes. In this case, Province, which belongs to the Marine Regime Highlands the orographic boundary is the one that stands out most clearly Division of the Humid Temperate Domain. and is, therefore, treated as the ecoregion boundary, even The ecoregions of the United States listed in Table 1 are though this choice is essentially arbitrary. shown in Figure 2, which also shows the distribution of ten For the purpose of the ecoregion map (Bailey 1976), the marine and estuarine provinces proposed by Cowardin and working definition of highlands that was adopted was an area others (1979). in which the relief must be high enough to bring about a differentiation of the vegetal cover with elevation. In addition, this differentiation must be great enough to cause a change in life form, for example, from desert shrub to forest. The basic Mapping Procedures idea was to identify regions based not on orography alone, but When considering ecoregion differences, one should realize on vegetation as the result and index of orography and climate. that regional boundaries indicate where significant ecological Such a definition would include high mountains such as the changes are taking place, often over a transitional zone. Colorado Rocky Mountains but would exclude the lower relief Therefore, the definitions of the limits will always be some- Appalachians. Although the widely spaced ranges of the what subjective. A boundary is placed where changes in the southwest exhibit such zonation, the region as a whole is climate-vegetation-soil conditions appear to be most pro- occupied largely by gently sloping plains. Thus, at the level of

2.Lowland Delineation of Ecosystem Regions 369 Table 1 Scheme for a fourth-order ecosystem regionalization of the United States. 1000 Polar Domain 1200 Tundra Division 2530 Tall-grass Prairie Province 1210 Arctic Tundra Province 2531 BluestemPrairie Section 1220 Bering Tundra Province 2532 Wheatgrass-Bluestem-Needlegrass Highlands) Section M1210 Brooks Range Province 2533 Bluestem-Grama Prairie Section 1300 Subarctic Division 2600 Mediterranean Division 1310 Yukon Parkland Province 26to GrasslandProvince 1320 Yukon Forest Province Highlands! M2610 Sierran ForestProvince Highlands! M2620 California Province M1310 Alaska Range Province 3000 Dry Domain 2000 Humid Temperate Domain 3100 SteppeDivision 2100 Warm Continental Division 3110 Great Plains Short-grassPrairie Province 2110 Laurentian Mixed Forest Province 3111 Grama-Needlegrass-Wheatgrass 2111 Spruce-fir Forest Section 2112 Northern Hardwoodsc-Fir Forest Section 3112 Wheatgrass-NeedlegrassSection Section 3113 Grama-BuffalograssSection 2113 Northern Hardwoods Forest Section 3120 PalouseGrassland Province 2114 Northern Hardwoods-Spruce 3130 IntermountainSagebrush Province Forest Section 3131 Sagebrush-Wheatgrass Section Highlands! 3132 LahontanSaltbush-Greasewood M2110 Columbia ForestProvince Section M2111 Douglas-fir ForestSection 3133 Great BasinSagebrush Section M2112 Cedar-Hemlock-Douglas-fir 3134 BonnevilleSaltbush-Greasewood ForestSection 2200 Hot ContinentalDivision Section 3135 PonderosaShrub ForestSection 2210 EasternDeciduous Forest Province 3140 Mexican Highlands Shrub SteppeProvince 2211 Mixed MesophyticForest Section Highlands! 2212 Beech-MapleForest Section M3110 Rocky Mountain ForestProvince 2213 Maple-BasswoodForest + Oak M3111 Grand Fir-Douglas-fir Forest SavannaSection Section 2214 AppalachianOak ForestSection M3112 Douglas-fir ForestSection 2215 Oak-Hickory ForestSection M3113 PonderosaPine-Douglas-fir 2300 SubtropicalDivision ForestSection 2310 Outer CoastalPlain ForestProvince M3120 Upper Gila Mountains ForestProvince 2311 Beech-Sweetgum-Magnolia-Pine-Oak P 3130 ColoradoPlateau Province ForestSection P3131 juniper-Pinyon + 2312 SouthernFloodplain ForestSection Sagebrush-SaltbushMosaic 2320 SouthernMixed ForestProvince Section 2400 Marine Division P3132 Grama-GalletaSteppe+ 2410 Willamette-PugetForest Province juniper-Pinyon Woodland Highlands! Section M2410 Pacific ForestProvince A3140 Wyoming BasinProvince M2411 Sitka Spruce-Crear-Hemlock A3141 Wheatgrass-Needlegrass- ForestSection SagebrushSection M2412 RedwoodForest Section A3142 Sagebrush-Wheatgrass Section M2413 Cedar-Hemlock-Douglas-fir 3200 Desert Division ForestSection 3210 ChihuahuanDesert Province M2414 California Mixed Evergreen 3211 Grama-Tobosa Section -ForestSection 3212 Tarbush-CreosoteBush Section M2415 Silver Fir-Douglas-fir Forest 3220 AmericanDesert (Mojave-Colorado-Sonoran) Section Province 2500 Prairie Division 3221 CreosoteBush Section 2510 Prairie Parkland Province 3222 CreosoteBush-Bur SageSection 2511 Oak-History-BluestemParkland 4000 Humid Tropical Domain Section 4100 Tropical SavannaDivision 2512 Oak + BluestemParkland Section 4110 EvergladesProvince 2520 Prairie BrushlandProvince 4200 RainforestDivision 2521 Mesquite-BuffalograssSection Highlands! 2522 Juniper-Oak-Mesquite Section M4210 Hawaiian IslandsProvince 2523 Mesquite-AcaciaSection

IKey to letter symbols: M-Mountain, P-Plateau, A-Altiplano. 370 R. G. Bailey

.§

~~ "Q. .->< "- oS .- ~~ E-- tj C/) .-,,; ~ ci5 ] .2 :::>.- ...-= .,0 I: 0 'Qj.- 8 ~ N

Q) ~ ~0) Delineation of Ecosystem Regions

generalization of the map, they are considered lowland Table 2. Regional climates, based on the Koppen system of regions. classification (1931), as modified by Trewartha (1943). In addition to mountain systems,the extreme high-altitude phases of plains (altiplano) and plateaus should be distin- Koppengroups and types guished so that the distinctive position of such plains as the A-Tropical humid climates Wyoming Basin and the Colorado Plateau may be appre- (Af) ciated. Tropical (Aw) Within the lowland series, section and province boundaries B-Dry climates were established by using the diagnostic criteria of Crowley Dry domain Steppe(BSk) Steppedivision (1967) to generalize Kuchler's map. The generalization was of Desert(BWh, BWk) Desert division two types. First, boundaries were simplified by elimating enclavesand peninsulas that were deemedto small to show on C-Subtropical climates Humid temperatedomain the finished small-scale map. (The smallest region shown on Mediterranean(Csa) Mediterranean division Humid subtropical(Cfa) Subtropical division the map is about 11,100 km2 (4300 sq mi).) Second,there was a Prairie division I reduction in the number of Kuchler types. Some of Kuchler's Marine west coast (Cfb) Marine division mapping units show the presence of large areas of azonal soils such as sand plains and peat deposits. These units were D-Temperate climates Humid continental, warm combined with surrounding zonal types in delineating sections Hot continentaldivision summer (ilia) Prairie division) and provinces; this relegated edaphically controlled ecosystems Humid continental, cool Warm continentaldivision to a lower level of classification and to more detailed maps. summer (Dfb) Prairie division) Lowlands and highlands were then grouped into larger climatic regions (domains and divisions) following the Koppen Polar domain Subarctic(Dfc, Dfd) Subarcticdivision (1931) system. Koppen's systemis simple, is basedon quantita- tive criteria, and correlates well with the distribution of many E-Polar climates natural phenomena, such as vegetation and soils. In the Tundra (ET) Tundra division Koppen system,summarized in Tables 2 and 3, eachclimate is Ice cap (EF) defined according to assigned values of temperature and pre- cipitation, computed in terms of annual and monthly values. It has become the most widely used climatic classification for geographical purposes. Particularly useful in delineating cli- matic regions were the climatic map of the world, modified Table 3. Definitions and boundaries of Koppen system. from the Koppen system (Trewartha 1943) at a scale of A climates; without frost. Coolest month warmer 1:75,000,000, and the climatic map of North America (Thorn- than 18°C (65°F). thwaite 1931) at a scale of 1:20,000,000. Boundaries of the B Dry climates; evaporation exceedsprecipitation. climatic regions were altered in some cases to make them BS-Steppe or semiarid climate. BW-Desert or arid climate. conform to vegetation boundaries. C Subtropical climates; 8 months or more warmer than 10°C (50°F); coolest month warmer than DoC (32°F) but colder than 18°C (65°F). Testing and Validation D Temperate forest climates; 4-8 months warmer than 10°C (50°F); coldest month cooler than DoC (32°F). As Rowe and Sheard (1981) point out, maps are hypotheses E Polar climates, warmest month colder than 10°C (50°F). to be tested and improved. The unit areas delineated are ET -Tundra climate; warmest month colder than 10°C (50°F) but warmer than DoC (32°F). hypotheses that arise from a knowledge of what is ecologically EF-Perpetual frost; all months colder than DoC (32°F). important. The regionalization of the United States, as just a-average temperature of warmest month warmer than 22°C described, is the product of an ecological concept. A key (72°F). hypothesis is that the large divisions, bounded by changes in b-average temperature of warmest month colder than 22°C (72°F). climatic regime, are functionally different in important ways. c-fewer than 4 months warmer than 10°C (50°F). d-same as c, but coldest month cooler than -38°C (-36°F). Furthermore, it is hypothesized that similar sites within map f-no dry season; rainfall throughout the year. units should respond in a similar way to management. h-hot and dry; all months warmer than DoC (32°F). Some land management agencies are currently using the k-cold and dry; at least one month colder than DoC (32°F). ecoregion concept without formal validation, much as others s-dry seasonin summer. have used Kuchler's (1964) map of potential natural vegeta- w-dry seasonin winter.

371 372 R. G. Bailey

tion, on the assumption that nothing better is available. Thus, Burger: D. .1976. The concept of ecosystem region in forest site while on the surface it may appear that the ecoregion conceptis classification. Pages 1-6 in Proceedings XVI IUFRO World definitive, in fact the concept and map must be tested and Congress, Oslo, Norway. validated before long-range analysis and planning are under- Coward.in, ~. M., V. Carter, F. C. Golet, and E. T. LaRoe. 1979. taken. Classification of wetlands and deepwater of the United If adequate data on productivity are collected from the States. FWS/OBS-79/31. US Fish and Wildlife Service, Washing- ton, DC. 103 pp. regions and assembled,the hypothesis underlying the regional- ization can be evaluated statistically, and the validity of the Crowley, J. M. 1967. . Can. Geogr. 11:312-326. regional structure (map) objectively evaluated. A program of Damman, A. W. H. 1969. The role of vegetation analysis in land classification. For. Chron. 55:175-182. research to statistically test the ecoregion construct by using Davis, L. S. 1980. Strategy for building a location-specific, multi- data to verify and refine the boundaries of ecoregions, the purpose information system for wildland management. j. For. internal homogeneity of the regions and their subdivisions, and 78:402-408. the general applicability of the ecoregion conceptto nationwide Delvaux, J., and A. Galoux. 1962. Les territoires ecologiques du resource assessmentand planning is being developed. sud-est beIge. Centre d'Ecologie generale. Travaux hors serie. 311 pp. Dokuchaev, V. V. 1899. On the theory of natural zones. Sochineniya Summary (Collected Works) v. 6. Academy of Sciences of the USSR, Moscow-Leningrad, 1951. The map of the ecoregions of the United States depicts Driscoll, R. S., D. L. Merkel, J. S. Hagihara, and D. L. Radloff. ecosystemsof regional extent according to the Crowley classifi- 1981. A component land classification for the United States: status and plans. US Department of , Forest Service, Fort cation, with climate and vegetation as indicators of the extent of Collins, CO. 77 pp. (mimeo). each unit. Four ecological levels are shown. The broadest, Grigor'yev, A. A. 1961. The heat and moisture regime and geographic domains, are based on observable differences that have devel- zonality. Soviet : Review and Translation 2:3-16. oped largely because of prevailing climatic conditions. Then, Hammond, E. H. 1964. Classes of land-surface form in the forty-eight on the basis of further climatic criteria, domains are broken states, U.S.A. Ann. Assoc. Am. Geogr. v 54, Map Supplement down into categories called divisions which, on the basis of the No.4. climax plant formation that geographically dominates the area Herbertson, A. J. 1905. The major natural regions: an essay in of the province, are subdivided into provinces. Provinces are systematic geography. Geogr. J. 25:300-312. subdivided into sections which differ in the floristic composition Hills, G. A. 1960. Regional site research. Forestry Chronicle 36:401- of the climax plant formation. Highland provinces and sections 423. are distinguished where, as a result of the influence of altitude, Holdridge, L. R. 1947. Determination of world plant formations from the climatic regime differs substantially from adjacent lowlands simple climatic data. Science 105:367-368. to cause complex vertical climate-vegetation-soil zonation. Hopkins, A. D. 1938. Bioclimatics: a science of life and climate relations. Miscellaneous Publication 280. US Department of Agri- culture, Washington, DC. 188 pp. Isachenko, A. G. 1973. Principles of landscape scienceand physical- Acknowledgments geographic regionalization. Melbourne University Press, Carlton, I wish to thank J. F. Franklin, A. W. Kuchler, and J. S. Victoria, Australia. 311 pp. Rowe for reviewing the manuscriptand for making several James, P. E. 1951. A geography of man. Ginn and Co., Boston, MA. valuablesuggestions, which I haveincorporated into the text. 375 pp. Koppen, W. 1931. Grundriss der Klimakunde. Walter de Gruyter Co., Berlin. 388 pp. Literature Cited Krajina, V. J. 1965. Biogeoclimatic zones and classification of British Columbia. Pages 1-17 in V. J. Krajina, ed. of western Bailey, R. G. 1976. Ecoregionsof the United States (map). US North America. University of British Columbia Press, Vancouver, Departmentof Agriculture, ForestService, Ogden, UT. Canada. Bailey, R. G. 1980.Description of the ecoregionsof the United States. Kuchler, A. W. 1964. Potential natural vegetation of the conterminous US Department of Agriculture, MiscellaneousPublication 1391, United States (map and manual). American Geographical Society Washington,DC. 77 pp. Special Publication 36. New York. 116 pp. Biasutti, R. 1962. II paesaggioterrestre, 2nd ed. Unione Tipografico, Kuchler, A. W. 1973. Problems in classifying and mapping vegetation Torino, Italy. 586pp. for ecological regionalization. Ecology 54:512-523. Borchert,J. F. 1950. The climate of the central North American Merriam, C. H. 1898. Life zones and crop zones of the United States. .Ann. Assoc.Am. Geogr.40:1-39. US Department of Agriculture. Bull. Div. Biol. SUTV.10:1-79. Delineation of Ecosystem Regions 373

Rowe,J. S. 1962.Soil, siteand land classification.Forestry Chronicle 38:420-432. Rowe,J. S.,and J. W. Sheard.1981. Ecological land classification:a surveyapproach. Environ. Manage.5:451-464. Spurr,S. H., andB. V. Barnes.1980. Forest ecology. John Wiley and Sons,New York, NY. 687pp. Thornthwaite,C. W. 1931.The climatesof North Americaaccording to a newclassification. Ceogr. Rev. 21:633-~55. Thornthwaite,C. W. 1948.An approachtowards a rational classifi- cationof climate.Ceogr. Rev. 38:55-94. Tosi,J. S. 1964.Climatic controlof terrestrialecosystems: a report on the Holdridge model.Econ. Ceogr.40:173-181. Trewartha, G. T. 1943.An introduction to weatherand climate,2nd ed. McGraw-Hill, New York, NY. 545pp. Troll, C. 1971. (geoecology) and biogeocenology- a terminologystudy. Ceoforum8:43-46. Udvardy, M. D. F. 1975. A classificationof the biogeographical provincesof the world. OccasionalPaper 18. International Union for Conservation of and Natural Resources,Morges, Switzerland.48 pp. Walter, H. 1977. Vegetationszonenund Klima: die okologische Gliederungder Biogeosphare.Eugen Ulmer Verlag,Stuttgart. 309 pp. Walter, H., and E. Box. 1976. Global classificationof natural terrestrialecosystems. Vegetatio 32:75-81. Weaver, J. E., and F. E. Clements. 1938. Plant ecology,2nd ed. McGraw-Hill, New York, NY. 601 pp. Wiken, E. B., and G. Ironside. 1977.The developmentof ecological (biophysical)land classificationin Canada. LandscapePlanning 4:273-275.