HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Related articles in JEQ Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Shirazi, M. A. Articles by Griffith, G. E. Search for Related Content PubMed PubMed Citation Articles by Shirazi, M. A. Articles by Griffith, G. E. GeoRef GeoRef Citation Agricola Articles by Shirazi, M. A. Articles by Griffith, G. E. Related Collections Spatial Distribution Watershed and Landscape Processes Spatial Variability Ecosystem Management Other Models

TECHNICAL REPORTS
Landscape and Watershed Processes

Quantitative Soil Descriptions for Ecoregions of the United States

^a Western Ecology Division, National Health and Environmental Effects Research Laboratory (NHEERL), USEPA
^b USDA Natural Resources Conservation Service, Watershed Science Institute
^c Indus Corporation, 200 SW 35th Street, Corvallis, OR 97333

^* Corresponding author (safa{at}mail.cor.epa.gov)

Received for publication January 21, 2002.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Researchers have defined and mapped ecological regions of the United States based on similar patterns of ecosystems such as deserts, forests, and croplands. These studies are useful in regional research, monitoring, and environmental management because data can be more readily extrapolated within the same ecoregion and to regions with similar characteristics. The description of ecoregions is largely holistic and qualitative. Conversely, quantitative information for soil are abundant and soil is an important ecosystem component related to many ecoregion properties. We used the nationwide State Soil Geographic database (STATSGO) to describe the soils of 84 Level III ecoregions in the United States. Among the 24 soil characteristics studied were texture, rock fragments, available water capacity, bulk density, and organic matter content. For each ecoregion we developed ranks to describe (i) its similarity to the U.S. average soil characteristics, (ii) the accuracy of predicting those characteristics, (iii) how well the soil map unit boundaries fit within ecoregion boundaries, (iv) the spatial relationship of soils across neighboring ecoregion boundaries, and (v) the homogeneity of texture–rock patterns. We present a national map of soil texture and rock fragments and five soil ranks for each ecoregion, and examine relationships between soils and other ecological components for selected ecoregions. Because soils relate to other ecosystem components such as vegetation, geology, and land use, the soil ranks complement and enrich the qualitative ecoregion descriptions. Similar analyses of physical or biological components of ecoregions will expand the understanding of the ecosystem patterns.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
WISE MANAGEMENT and policy decisions regarding the environment require assessments of complex ecosystems and translations of theoretical constructs into applications (Hoag et al., 1998). The environment consists of the physical, chemical, and biological factors that interact with each other. At a biological community level, these relationships define an inseparable entity called an ecosystem (Odum, 1959; Warren, 1971). Ecosystems, such as forests, deserts, and agricultural land are considered natural units for studying the environment. Ecoregions are composed of ecosystems that are relatively homogeneous, or exhibit a particular pattern, within a defined boundary, that is different from the neighboring regions (Bailey et al., 1985; Rowe and Sheard, 1981; Omernik and Bailey, 1997; Omernik et al., 2000). Ecoregions represent geographic units for regional studies (Commission for Environmental Cooperation, 1997) and as with ecosystems, no single component of an ecoregion can completely describe it. However, investigations of components begin the process of translation between theoretical and applied knowledge.

Decisions regarding environmental monitoring or assessment are often made at the spatial scale of ecoregions. Karlen et al. (1997) discussed the use of soil data at the regional scale but very few quantitative data have been developed for ecoregions. This type of data is needed because resource utilization often targets particular features, such as minerals, water, or lands productive for agriculture or forests. Further, lessons learned within an ecoregion are more readily transferable in the same ecoregion and to other similar ecoregions.

Soils are an important ecoregion component that has been used in environmental assessments with ecological, social, and economic contexts (Hoag et al., 1998). Others have used soil characteristics and models to describe the influence of soil on water quality (Shirazi et al., 2001b,c; Johnson et al., 1991; Larson and Pierce, 1994). In a discussion of environmental assessment using soil data (e.g., the State Soil Geographic database [STATSGO]; Soil Survey Staff, 1991), Lammers and Johnson (1991)(p 154) noted that "soil map units can be linked to soil properties needed to support process models." These authors emphasized that "a geographic framework is needed ... to ensure spatial congruence of data" and cited Omernik (1987) as a suitable framework for the United States. Our study combines soil data from STATSGO with a national framework (e.g., ecoregion maps) to facilitate regional environmental assessments in the United States.

Omernik and others defined ecoregions at the continental scale (Level I), national and regional scales (Levels II and III), and for 32 of the 48 conterminous United States (USEPA, 2000) at a subregional scale (Level IV). Level III ecoregions are widely used in environmental management and they incorporate many components such as geology, vegetation, soils, and land use to describe them (USEPA, 2000). Ecoregions are periodically updated with new information and 84 Level III ecoregions are currently defined for the conterminous United States. They range in size from 15 000 to 365 000 km² and may be discontinuous.

The scale of STATSGO is suitable for this study because it is designed to be "useful for understanding the soil resources and for planning broad use in a state or region" (Soil Survey Staff, 1991). In STATSGO, map units define geographic land areas by the characteristics of similar soil series. There are 10 483 soil map units in the United States with an average area of 750 km². Each map unit may include many dispersed polygons (mean number of polygons = 7.5, standard deviation = 17.7, minimum = 1, maximum = 405), and because soil characteristics of separate polygons are not distinguished from those of the entire map unit at the STATSGO scale, the map unit is the minimum spatial entity.

Shirazi et al. (2001c) aggregated soil characteristics of all soil layers above bedrock for each map unit in the STATSGO database and developed five models. The models were based on relationships among the map units in the texture classes (USDA5): cr = coarse, mocr = moderately coarse, mecr = medium coarse, mofn = moderately fine, and fn = fine. They partitioned the map units according to similar ranges of rock fragments in each texture class and used the appropriate model to predict the mean soil characteristics. Because the map units for all 48 conterminous U.S. states were used, we refer to these models collectively as the U.S. model. The authors presented a detailed list of soil characteristics and later applied the models to water quality prediction (Shirazi et al., 2001b, c).

The goal of our study is to quantify soil properties of Level III ecoregions of the United States. Specifically, our objectives are to (i) produce a map of 20 texture–rock classes along with the Level III ecoregions for the United States; (ii) use the U.S. model and 24 selected STATSGO variables to estimate the mean soil characteristics of the United States and each ecoregion; (iii) describe the agreement between soil and ecoregion boundaries, spatial separation across ecoregion borders, and the diversity of textures in each ecoregion; and (iv) summarize the above soil information as quantitative ranks to examine them in relation to other ecoregion characteristics.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
We followed the methods of Shirazi et al. (2001a) to classify all soil map units in the United States by texture. Each USDA5 texture class was also divided into four sets based on percent rock fragments, thus forming 20 classes of soil map units. The data were stored in a geographic information system (GIS) to create a map of the 20 texture–rock classes, and GIS techniques were used to intersect the soil map units with each ecoregion and determine the percent of map unit area inside the ecoregion border.

We determined the mean value for each of the 24 selected soil characteristics (see Table 1 for names and abbreviations) from all map units in a texture–rock class. Thus, there were 24 x 20 = 480 "observed" means derived from the soil characteristics of STATSGO map units nationwide. The means were also estimated using the U.S. models and standard errors were calculated by regression with the observed means. These values were needed for subsequent comparisons with ecoregion soil characteristic means, which were estimated by the same procedure.

The conceptual and mathematical definitions for the "agreement" or "fit" between ecoregion and soil map unit boundaries are based on their intersection areas. Level III ecoregions intersect an average of 196 map units (standard deviation = 150, minimum = 22, maximum = 738). The boundaries of the map units frequently extend beyond an ecoregion boundary, producing what we call an exclusion zone. This zone is made up of parts of the intersecting map units that are outside the ecoregion and in contrast to the inclusion area inside the boundary. When the soil exclusion area is large relative to the ecoregion and its circumscribed inclusion area, the "fit" to the ecoregion boundary is considered poor. Conversely, a small exclusion area, with an inclusion area nearly equal to the ecoregion area, indicates a high level of spatial coincidence. Ranks describing the soil–ecoregion boundary agreements based on intersection areas are derived below.

We used A_e, A_m, and A_i to denote, respectively, the areas of an ecoregion, a map unit, and the part of an intersecting map unit inside the ecoregion boundary. These three areas were combined into two types of dimensionless areas _m = A_m/A_e and _i = A_i/A_e and we referred to them as map unit and inclusion areas. Next, we sorted the inclusion areas from the largest to the smallest such that:

[1]

where t is the total number of intersections and 1 < s < t. The number, s, is a derived property that defines the area of "fit" (Eq. [4], below). We calculated sorted cumulative areas from Eq. [1] and used superscripts for summation as follows:

[2]

The quantity ^t_i is the total soil area within an ecoregion covered by t map units. Theoretically, ^t_i = 1 for ecoregions containing no surface waters. However, ^t_i is generally less than 1 and is used as the limit for cumulative inclusion area. These statistics range from 0.3178 to 0.9341 and, for the USA as a whole, ^t_i = 0.9830.

Like the inclusion areas, we summed the entire area of intersected map units into:

[3]

where ^t_m is the total of map unit areas. A unique exclusion area can be calculated for any sequential map unit number based on the difference between Eq. [2] and [3]. For example, ¹_m - ¹_i is the exclusion area for the first map unit, and the exclusion zone for all map units is ^t_m - ^t_i.

The cumulative areas in Eq. [2] and [3] each form a line in a coordinate system with the number of map units as the horizontal axis and as the vertical axis. The two lines never intersect, but the inclusion versus exclusion area relationships are optimized at map unit number s, such that:

[4]

In other words, the solution for s is obtained when the direction of the inequality changes from < to > between map unit s and s + 1. Equation [4] delimits this point of "switching" or "crossover" beyond which the cumulative map unit area exceeds the inclusion area limit. The resulting ^s_i value mathematically defines the "fit." We sorted the 84 ecoregion ^s_i values from the largest to the smallest to create a rank (R_mu) that reflects its place along a good-to-poor agreement sequence.

To determine how much of the soils in an ecoregion, referred to as the host, are shared with their neighbors, we used map units 1 through s of the host (i.e., up to and including the switch point). Recall that parts of these map units may exist in neighboring ecoregions, which are not necessarily directly adjacent to the host ecoregion. The host inclusion area was divided by the sum of corresponding map unit areas in the neighboring ecoregions to produce a ratio. Then, the reciprocals of the ratios for the 84 ecoregions were sorted from the smallest to largest to form a neighbor–host rank called R_nh. Low ranks indicate more separation of soils between the host and neighboring ecoregions and high values denote more overlap.

We used the 20 texture–rock classes to describe soil homogeneity within ecoregions. Theoretically, an ecoregion's soil texture can be completely heterogeneous if covered uniformly by all 20 classes or completely homogeneous if covered by a single class. That is, homogeneity or heterogeneity = (maximum area of any class)/(mean area of all possible classes). The values are obtained by separating ^t_i into areas for each class and determining the class with the greatest total area:

[5]

where ß_k is the total area in class k and the superscript tk denotes summation over the number of map units in the class. The two theoretical limits for any ecoregion are:

[6]

[7]

Equation [6] defines the uniform distribution of all 20 classes and Eq. [7] indicates that only class j covers the entire soil area of the ecoregion. For the first limit, ^t_i/20 = 0.05, and the maximum to mean ratio is ß/0.05 = 1 in Eq. [6]. For ^t_i = 1 in Eq. [7], the ratio is ß_j/0.05 = 20. We calculated the texture–rock class areas for each of the 84 ecoregions, divided the maximum by the real ^t_i/20 (i.e., mean area of all possible texture classes) for the ecoregion, sorted the 84 ratios from the largest to the smallest, and assigned a rank (R_tx). Thus, low R_tx ranks indicate homogeneous texture and higher ranks reflect greater heterogeneity. For example, for the whole USA, ^t_i = 0.9830, ß₁₇ = 0.1205, 0.9830/20 = 0.0492, and 0.1205/0.0492 = 2.45, which indicates a high rank and a heterogeneous texture.

Ecoregions can be compared using the appropriate individual ranks. However, a multivariate analysis is needed to determine relationships among the ranks. We used a minimal spanning tree analysis (Becker et al., 1988) based on the original five soil statistics, for the 84 ecoregions and the United States, to classify ecoregions relative to the U.S. values.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Figure 1 displays the spatial relationship of texture and rock fragment classes for the conterminous United States overlaid with Level III ecoregion boundaries. The texture–rock fragment classes and ecoregion lines were produced independently but are shown together to facilitate visual comparison.

View larger version (201K): [in this window] [in a new window]   Fig. 1. Soil texture (USDA5) and rock contents of State Soil Geographic database (STATSGO) soil map units of the conterminous United States and the boundaries of 84 Level III Omernik ecoregions. The yellow color represents coarse (cr) soils with colors darkening through moderately coarse (mocr), medium coarse (mecr), moderately fine (mofn), and fine (fn) soils in the reddest shade. A higher percentage of rocks adds coarseness to a soil and is depicted as decreasing color intensity within the texture class.

Ecoregion Ranks and Soil Map Unit Data
Although data are available for all ecoregions, we explain soil characteristics and the ranks using the centrally located Ecoregion 39 (the Ozark Highlands). This ecoregion is intersected by 79 soil map units and has a moderately fine (mofn) mean texture (Table 2). The soil characteristics are unlike the U.S. mean values as indicated by R_sc = 72 and the detailed comparison in Fig. 2a . Their errors are shown in Fig. 2b, and together provide a good predictability rank of R_er = 8.

View larger version (27K): [in this window] [in a new window]   Fig. 2. Comparison of U.S. and Ecoregion 39 soil characteristics. (a) The baseline represents each mean soil characteristic for the USA and the bars above and below represent the amount of difference for Ecoregion 39. For example, the cation exchange capacities (cath, catl), percent clay, and liquid limits (liqh, liql) are greater than the national averages whereas soil permeability (perh, perl) and percent sand are less. (b) The bars indicate error differences between the U.S. model and the standard errors of prediction for the Ozark Highlands. Values for percent clay and percent rocks are above the U.S. average in this ecoregion, and perl is below average. Rsc, compares mean characteristics to the national average; Rer, error of prediction relative to the U.S. models.

For the Ozark Highlands, 36 of the 79 intersected map units cover 93.4% of its total area (Fig. 3a) , which produces the best boundary agreement among the 84 ecoregions and the rank of R_mu = 1. Figure 3a illustrates the increase in inclusion area and total map unit area with increasing numbers of intersected map units. Figure 3b shows the spatial distribution of the cumulative areas inside and outside Ecoregion 39 at the switch point.

View larger version (35K): [in this window] [in a new window]   Fig. 3. Analysis of Ecoregion 39. (a) The relationships between map unit areas (solid lines, Eq. [3]) and inclusion areas (dashed lines, Eq. [2]) for Ecoregion 39. The vertical axis represents dimensionless areas, . From left to right, the circular markers locate map unit areas 1m, sm, and tm on the solid lines, and 1i, si, and ti on the dashed lines. The 1 denotes the largest single contributor of area, s is the "switch point" from inside to outside the ecoregion, t is the total number of map units, and the subscript i refers to the inclusion zone. The vertical difference between the solid and dashed lines represents the size of the exclusion zone for the ecoregion. At 1, the exclusion zone (1m - 1i) is minimal whereas the inclusion area, 1i, is the largest for a single map unit. (b) The spatial relationships of intersecting map units of Ecoregion 39 demonstrate the "switch point" between the inclusion and exclusion zones. Map unit areas < s are dark gray, parts of which are outside the ecoregion boundary. Although the light gray areas extend far beyond the ecoregion boundary, only those within the ecoregion boundary and the dark gray areas outside the boundary determine the boundary agreement and the neighbor–host ranks (Rmu and Rnh).

Ecoregion Ranks and Geographic Characteristics
Level III ecoregions represent the distribution of distinct patterns of ecosystems in the United States. The patterns naturally differ because of relationships between many ecological characteristics. For example, the mutual occurrence of cool and moist soils in a mountainous terrain, wet or moist forest vegetation, and forestry land use, may be used to define a homogeneous region. If these correlated geographic phenomena are spatially extensive and produce sharp boundaries with the neighboring ecoregions, the region is "distinct" as well. The relationships between ecological components in a heterogeneous ecoregion are variable or form a spatially fragmented pattern (mosaic). When the characteristics gradually merge with the neighboring regions, it is a transitional ecoregion. An understanding of these patterns helps determine the extent to which data may be extrapolated within an ecoregion or between similar regions and is aided by quantitative assessments of one or more components. The reliability of data extrapolation is improved when components of the ecoregion are concordant.

Soils vary considerably in different ecoregions and the five soil ranks define distinctiveness, texture homogeneity, and transitional characteristics of ecoregions with respect to their soils. The relationships among ranks form patterns similar in concept to concordant ecoregion components, except that the ranks are quantitative and relative to the entire United States. The numerical statistics derived for all but the soil texture rank are bounded by the U.S. values as the upper limits.

To facilitate the discussion of ecoregions spanning a diverse array of soils, the ecoregions were sorted based on the relationships among all five soil rank statistics with respect to the U.S. values. The sorted list was then divided into three equal groups. We used the groups to illustrate the predictability of soil characteristics as shown in Fig. 4 , plotting the smallest total errors on the left and progressively larger errors to the right. As the distance from the U.S. error baseline increases from the left to the right and from Group A to Group C, the predictability of the ecoregion soil characteristics decreases.

View larger version (26K): [in this window] [in a new window]   Fig. 4. A summary of the standard errors of predicting soil characteristics of the State Soil Geographic database (STATSGO) map units of the United States and three groups of Level III ecoregions. The groups were based on the relationships among the five soil ranks compared with the U.S. values. Group A has the highest predictability (lowest total error) and Group C has the lowest predictability (highest errors). Group A = Ecoregions 8, 32, 37, 54, 24, 34, 65, 70, 26, 52, 35, 40, 47, 73, 43, 46, 6, 72, 67, 55, 7, 30, 42, 71, 22, 74, 39, and 61. Group B = Ecoregions 17, 79, 59, 69, 62, 82, 3, 48, 60, 81, 31, 36, 53, 20, 56, 51, 33, 23, 21, 9, 45, 10, 29, 66, 78, 68, 19, and 28. Group C = Ecoregions 49, 76, 77, 75, 84, 50, 12, 63, 58, 2, 1, 15, 14, 18, 5, 27, 80, 11, 64, 41, 38, 13, 16, 57, 44, 83, 4, and 25. Soil characteristics and standard errors of ecoregions in italic type are listed in Table 3.

The ecoregions in Table 3 (also marked in italic type in the Fig. 4 caption) represent the entire range of our soil ranks and were used to examine relationships between soils and other ecological components. This simple classification could be improved with quantitative information for other ecoregion components, and ultimately determine the ability to extrapolate data within the same ecoregion or between ecoregions in the group. However, discussions of associations between the soil ranks and other ecoregion components will remain qualitative until methods are developed to quantify those components.

Ecoregions Most Distant from the USA
Northern Minnesota Wetlands (49)
With ranks R_sc = 84, R_er = 84, R_mu = 80, R_nh = 82, and R_tx = 38, the soils in this ecoregion are the least similar to the U.S. mean soils, the least predictable, and have indistinct borders and moderately uniform textures. The high organic content of this ecoregion is a distinguishing soil characteristic. Ecologically, it is a transitional region. The central portion of Ecoregion 49 is dominated by swamps with boreal forest vegetation, sparse human population, and wildlife habitat as a major land use. On the east, the ecoregion gradually merges into the Northern Lakes and Forests (50) with its mix of spruce fir, pine, and hardwood vegetation and the land uses of forestry, recreation, and some mining. On the west, agriculture becomes more common, as is typical in the adjacent Lake Agassiz Plain (48) ecoregion.

The Nebraska Sand Hills (44)
With very prominent ranks R_sc = 83, R_er = 69, R_mu = 3, R_nh = 4, and R_tx = 3, this ecoregion stands out noticeably because of its soils. It is one of the most distinctive and ecologically homogenous ecoregions in North America as well as one of the largest areas of grass-stabilized sand dunes in the world. Coarse soils predominate and it generally lacks cropland agriculture and trees, except for some riparian trees in the north and east. Grazing is the main land use, but the density of cattle is less than in adjacent Ecoregions 47, 27, and 25. The latter regions are arable with irrigation and 47 is a productive cropland.

Ecoregions of Intermediate Distance to the USA
North Central Appalachians (62)
The ranks R_sc = 68, R_er = 52, R_mu = 84, R_nh = 84, and R_tx = 2 indicate highly diffuse soil boundaries and a very homogeneous texture. Most overlaps of the soil map units are with Ecoregions 67, 69, and 70 to the south and east. Ecoregions 60 and 62 to the north and west were affected by glaciation whereas most of this ecoregion was not, therefore a geomorphically distinct boundary separates these ecoregions. Ecoregion 62 is generally more heavily forested than its neighboring ecoregions and has very little agriculture. The neighboring ecoregions have dairy agriculture (61 and 70), a mosaic of forest and croplands (60 and 67), or are similarly densely forested (69). Ecoregion 62 is ecologically distinct, but the soil ranks indicate a poor fit of soil map units in the ecoregion and considerable overlap with adjacent regions.

Northern Appalachian Plateau and Uplands (60)
Ranks R_sc = 64, R_er = 48, R_mu = 74, R_nh = 74, and R_tx = 1 indicate that soils in this ecoregion are mainly distinctive in their texture homogeneity, which is similar to the neighboring North Central Appalachians (62), with R_tx = 2. Ecologically, Ecoregion 60 is transitional between the less hilly, more agricultural and urban terrain of the Eastern Great Lakes and Hudson Lowlands ecoregion (83) and the more mountainous, forested, and less populated Ecoregions 62 and 58. Although many areas of Ecoregion 60 contain dairy farms with pastures, hay, and grain fields, there are also large areas of oak and northern hardwood forests. Surrounding Ecoregions 83, 62, and 58 tend to have more homogeneous land cover and are more distinguishable in their combination of geographic characteristics.

Ecoregions Closest to the USA
Arizona–New Mexico Plateau (22)
The ranks R_sc = 1, R_er = 3, R_mu = 36, R_nh = 41, and R_tx = 84 indicate that soil characteristics are similar to the U.S. mean and highly predictable, with moderate agreement of soil–ecoregion boundaries and overlaps, and very heterogeneous textures. A substantial portion of the shared soils are in Ecoregion 23. Ecologically, Ecoregion 22 is transitional between semiarid grasslands in 26 to the east, the higher relief tablelands in 20 to the north, and the lower, hotter, and less vegetated deserts in the south (24) and west (14, 81). The mountainous regions around 22 are more moist and vegetated while the surrounding deserts are hotter and drier. The diversity of climate and terrain in this region is reflected in very heterogeneous patterns of soil characteristics and texture (R_sc = 1 and R_tx = 84).

Ozark Highlands (39)
The ranks R_sc = 72, R_er = 8, R_mu = 1, R_nh = 2 and R_tx = 22 define soils that are distinct in characteristics and with respect to segregation from neighboring ecoregions. It is also ecologically distinct because all bordering ecoregions, with the exception of the Boston Mountains (38), belong to separate Level II ecoregions. In particular, Ecoregion 39 differs in physiography, vegetation, and land use from adjacent Ecoregions 38, 40, 72, and 73. The majority of Ecoregion 39 is a dissected limestone plateau covered by oak or oak–pine forests. Less than one-fourth of its core has been cleared for pasture and cropland, but agriculture increases to about one-half or more at the periphery. However, the bordering Ecoregions 40, 72, and 73 are even more agricultural.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Conventional mapped soil texture information describes 12 USDA classifications separately for different clay, silt, sand, and rock contents. By contrast, the texture–rock map (Fig. 1) in this study was developed by combining these soil properties vertically and over the space of soil map unit groups (Shirazi et al., 2001a, c). Despite some discrepancies across state lines in STATSGO data, the overall patterns in the soil information help explain and identify Level III ecoregions. Shirazi et al. (2001b) demonstrated a link between soils and water quality predictions. This study found associations between ecoregion boundaries and soil information, thus confirming the importance of soils as an ecological property. In addition, it is another step in quantifying ecoregion characteristics. For a more complete analysis of ecoregions and to enhance the reliability of data extrapolations, similar assessments should be conducted on other components such as climate, geology, physiography, and flora and fauna.

	ACKNOWLEDGMENTS

 
Suggestions from Dr. Berman D. Hudson from the USDA Natural Resources Conservation Service, Dr. Alan Busacca from the Washington State University Department of Crop and Soil Sciences and Department of Geology, William W. Hargrove of the Oak Ridge National Laboratory, and the reviewers of JEQ significantly improved the clarity of our presentation. The USEPA through its Office of Research and Development funded the research. It has been subjected to USEPA review and approved for publication.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 

• Bailey, R.G., S.C. Zoltai, and E.B. Wiken. 1985. Ecological regionalization in Canada and the United States. Geoforum 16:265–275.
• Becker, R.A., J.M. Chambers, and A.R. Wilks. 1988. The S language. A programming environment for data analysis and graphics. Wadsworth & Books/Cole Computer Sci., Pacific Grove, CA.
• Commission for Environmental Cooperation. 1997. Ecological regions of North America, toward a common perspective. CEC, Montreal, QC, Canada.
• Hoag, D.L., J.S. Popp, and D.E. Hyatt. 1998. Sustainability and resource assessment: A case study of soil resources in the United States. EPA/600/R-98/038. USEPA Office of Research and Development, Research Triangle Park, NC.
• Johnson, M.G., P.W. Shaffer, D.L. Stevens, K.W. Thornton, and R.S. Turner. 1991. Predicting and forecasting surface water acidification: A plan for assigning data aggregation effects. EPA/600/3-9/024. USEPA, Corvallis, OR.
• Johnson, R.A., and D.W. Wichern. 1982. Applied multivariate statistical analysis. Prentice-Hall, Englewood Cliffs, NJ.
• Karlen, D.L., M.J. Mausbach, J.W. Doran, R.G. Cline, R.F. Harris, and G.E. Schuman. 1997. Soil quality: A concept, definition, and framework for evaluation (a guest editorial). Soil Sci. Soc. Am. J. 61:4–10.
• Lammers, D.A., and M.G. Johnson. 1991. Soil mapping concepts for environmental assessment. p. 149–160. In M.J. Mausbach and L.P. Wilding (ed.) Spatial variabilities of soils and landforms. SSSA Spec. Publ. 28. SSSA, Madison, WI.
• Larson, W.E., and F.J. Pierce. 1994. The dynamics of soil quality as a measure of sustainable management. p. 37–51. In Defining soil quality for a sustainable environment. SSSA Spec. Publ. 35. SSSA, Madison, WI.
• Odum, E.P. 1959. Fundamentals of ecology. 2nd ed. W.B. Saunders Co., Philadelphia, PA.
• Omernik, J.M. 1987. Ecoregions of the conterminous United States. Ann. Assoc. Am. Geogr. 77:118–125.
• Omernik, J.M., and R.G. Bailey. 1997. Distinguishing between watersheds and ecoregions. J. Am. Water Resour. Assoc. 33:935–949.
• Omernik, J.M., S.H. Chapman, R.A. Lillie, and R.T. Dumke. 2000. Ecoregions of Wisconsin. Trans. Wisconsin Acad. Sci. Arts Lett. 88:77–103.
• Rowe, J.S., and J.W. Sheard. 1981. Ecological land classification: A survey approach. Environ. Manage. 5:451–464.
• Shirazi, M.A., L. Boersma, and C. Burch Johnson. 2001a. Particle size distributions: Comparing texture systems, adding rock, and predicting soil properties. Soil Sci. Soc. Am. J. 65:300–310.[Abstract/Free Full Text]
• Shirazi, M.A., L. Boersma, C. Burch Johnson, and P.K. Haggerty. 2001b. Predicting physical and chemical water properties from relationships with watershed soil characteristics. J. Environ. Qual. 30:112–120.[Abstract/Free Full Text]
• Shirazi, M.A., L. Boersma, P.K. Haggerty, and C. Burch Johnson. 2001c. Spatial extrapolation of soil characteristics using whole-soil particle size distributions. J. Environ. Qual. 30:101–111.[Abstract/Free Full Text]
• Soil Survey Staff. 1991. State Soil Geographic data base (STATSGO), data user guide. Misc. Publ. 1492. U.S. Gov. Print. Office, Washington, DC.
• USEPA. 2000. Level III ecoregions of the continental United States. [Revision of Omernik, 1987.] USEPA Natl. Health and Environ. Effects Res. Lab., Western Ecol. Div., Corvallis, OR.
• Warren, C.H. 1971. Biology and water pollution control. W.B. Saunders Co., Philadelphia, PA.

Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 377-382. [Full Text]  

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Related articles in JEQ Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Shirazi, M. A. Articles by Griffith, G. E. Search for Related Content PubMed PubMed Citation Articles by Shirazi, M. A. Articles by Griffith, G. E. GeoRef GeoRef Citation Agricola Articles by Shirazi, M. A. Articles by Griffith, G. E. Related Collections Spatial Distribution Watershed and Landscape Processes Spatial Variability Ecosystem Management Other Models