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 Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (28) Citing Articles via Google Scholar Google Scholar Articles by Brejda, J. J. Articles by Allan, D. L. Search for Related Content PubMed Articles by Brejda, J. J. Articles by Allan, D. L. GeoRef GeoRef Citation Agricola Articles by Brejda, J. J. Articles by Allan, D. L.

DIVISION S-6-SOIL & WATER MANAGEMENT & CONSERVATION

Identification of Regional Soil Quality Factors and Indicators

II. Northern Mississippi Loess Hills and Palouse Prairie

^a USDA-ARS, Wheat, Sorghum, and Forage Res. Unit, 344 Keim Hall, Univ. of Nebraska, Lincoln, NE 68583 USA
^b USDA-ARS, National Soil Tilth Lab, 2150 Pammel Drive, Ames, IA 50011 USA
^c USDA-ARS, 215 Johnson Hall, Washington State Univ., Pullman, WA 99164-6421 USA
^d Dep. Soil, Water, and Climate, 1991 Buford Circle, Univ. Minn., St. Paul, MN 55108 USA

jbrejda{at}unlserve.unl.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 
Diversity of soil series present in a region may hinder identification of soil quality factors and indicators at a regional scale. Our objectives were (i) to identify soil quality factors for a diverse population of soils at the regional scale, (ii) to determine which factors vary significantly with land use, (iii) to select indicators from these factors that can be used with the National Resource Inventory (NRI) for monitoring soil quality, and (iv) to compare these results to a similar study involving only a single soil series. One hundred eighty-six points representing 75 soil series in the Northern Mississippi Valley Loess Hills and 149 points representing 58 soil series in Palouse and Nez Perce Prairies were sampled from a statistically representative subset of NRI sample points and analyzed for 20 soil attributes. Factor analysis was used to identify soil quality factors and discriminant analysis was used to identify factors and indicators most sensitive to land use within each region. In the Northern Mississippi Valley Loess Hills, five soil quality factors were identified. Discriminant analysis selected potentially mineralizable N (PMN), microbial biomass C (MBC), water stable aggregates (WSA), and total organic C (TOC) as the most discriminating attributes between land uses. In the Palouse and Nez Perce Prairies, six factors were identified. Discriminant analysis selected TOC and total N as the most discriminating attributes between land uses. The soil quality factors were similar among three of the four regions, but TOC was the only indicator common to all regions for distinguishing among land uses.

Abbreviations: CEC, cation-exchange capacity • CRP, Conservation Reserve Program • MBC, microbial biomass C • MEP, Mehlich III extractable P • MLRA, Major Land Resource Area • NRI, National Resource Inventory • PMC, potentially mineralizable C • PMN, potentially mineralizable N • TOC, total organic C • WSA, water stable aggregate

Abbreviations: *, ** Significant at the 0.05 and 0.01 levels of probability, respectively

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 
SOIL QUALITY has been defined as "the capacity of a soil to function within ecosystem and land use boundaries, to sustain biological productivity, maintain environmental quality, and promote plant and animal health" (Doran and Parkin, 1994). Soil functions that soil quality influences include the ability (i) to accept, hold, and release nutrients and other chemical constituents; (ii) to accept, hold, and release water to plants and surface and groundwater recharge; (iii) to promote and sustain root growth; (iv) to maintain suitable soil biotic habitat; and (v) to respond to management and resist degradation (Larson and Pierce, 1991). Maintaining or improving soil quality can provide economic benefits in the form of increased productivity, more efficient use of nutrients and pesticides, improvements in water and air quality, and amelioration of greenhouse gas emissions (USDA-Economic Research Service, 1997).

Because of its importance, a quantitative assessment of soil quality is needed to determine the sustainability of land management systems as related to agricultural production practices, and to assist government agencies in formulating and evaluating sustainable agricultural and land use policies (Doran and Parkin, 1996). However, soil quality cannot be measured directly, but must be inferred from soil quality indicators. Soil quality indicators are measurable soil attributes that influence the capacity of soil to perform crop production or environmental functions and are sensitive to change in land use, management, or conservation practices. However, many soil attributes are highly correlated (Larson and Pierce, 1991; Seybold et al., 1997). Correlated soil attributes do not change independently to changes in management, but respond as a group, integrating many complex interactions among biological, chemical, and physical soil processes. Single attribute indicators do not reflect interacting changes in soil quality that may occur with changes in management because of correlation among soil attributes. A more accurate assessment of soil quality may be achieved by evaluating several soil attributes simultaneously using statistical procedures that account for correlation among soil attributes.

Multivariate statistical analyses, such as factor analysis, provide techniques for studying the relationships among correlated variables (James and McCulloch, 1990; Johnson and Wichern, 1992). A regional-scale study of soil quality (Brejda et al., 2000) used factor analysis to statistically group 20 soil attributes on the basis of their intercorrelations into five factors for the Ascalon (fine-loamy, mixed, superactive, mesic Aridic Argiustoll) soil in the Central High Plains and six soil quality factors for the Amarillo (fine-loamy, mixed, thermic Aridic Paleustalf) soil in the Southern High Plains. Because each of these factors contributed to one or more soil functions, they were considered to represent soil quality factors and should not be confused with factors of soil formation proposed by Jenny (1980). The soil quality factors in each region were analyzed by analysis of variance and discriminant analysis to determine which were sensitive to differences in land use and could serve as potential indicators of soil quality at a regional scale. However, this analysis was done using only a single soil series within each region. Therefore, conclusions from analysis of the Ascalon and Amarillo soils are limited to these or similar soil series. Broader conclusions may be made concerning the composition of soil quality factors and their variation with different land uses or conservation practices if a large and diverse population of soil series from different regions were analyzed. However, the greater variation inherent in multiple soil series studies could mask our ability to identify soil quality factors or detect change in these factors with different land uses at the regional scale. Our objectives were (i) to identify soil quality factors at a regional scale for samples taken from a diverse population of soil series, (ii) to determine which factors vary significantly with land use, (iii) to select soil attributes within these factors that can be used as soil quality indicators with the NRI to assess effects of land use or soil conservation programs on soil quality, and (iv) compare these results with a similar study involving only a single soil series.

	Materials and methods

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 
Two Major Land Resource Areas (MLRA), designated the Northern Mississippi Valley Loess Hills and the Palouse and Nez Perce Prairies, were selected for this study. The Northern Mississippi Valley Loess Hills covers 27090 km² and is located in southwestern Wisconsin, northeastern Iowa, and southeastern Minnesota (Fig. 1a) . Elevation ranges from 200 m on river valley floors to 400 m on the highest ridges. Average annual precipitation ranges from 750 to 900 mm with two-thirds or more falling during the growing season. Average annual temperature ranges from 7 to 10°C. About 40% of the MLRA is cultivated for corn (Zea mays L.), soybean (Glycine max L.), and small grain production; 20% is in permanent pasture and hay land; and the remainder is forested (USDA-SCS, 1981).

View larger version (21K): [in this window] [in a new window]   Fig. 1 Geographic distribution (shaded areas) of (a) the Northern Mississippi Valley Loess Hills, and (b) the Palouse and Nez Perce Prairies Major Land Resource Areas

Experimental Design
A statistically representative sample of 200 points were selected within each MLRA using the NRI sampling framework. Detailed descriptions of sample point selection within the NRI framework are presented elsewhere (Goebel and Baker, 1982; Nusser and Goebel, 1997; Nusser et al., 1998; Brejda et al., 2000). Some points were inaccessible, or fell on homesteads, urban areas, road pavement, or rock outcrops. These points were not sampled. As a result, only 186 points were actually sampled in the Northern Mississippi Valley Loess Hills, and 149 points were sampled in the Palouse and Nez Perce Prairies. Points were selected at random, without regard to soil series or land use. This resulted in sampling 75 different series in the Northern Mississippi Valley Loess Hills, and 58 different series in the Palouse and Nez Perce Prairies. The 186 soils sampled in the Northern Mississippi Valley Loess Hills were predominately Alfisols , but also included Mollisols , Entisols , Inceptisols , and Histosols . At two sampling points, the soil series were not classified. The 149 soils sampled in the Palouse and Nez Perce Prairies were predominately Mollisols , with a few Alfisols , Inceptisols , Andisols , and Entisols . At one sampling point, the soil series was not classified.

Soil Sampling and Analysis
At each sample point a soil pit was dug, the depth of the A horizon was measured, and A horizon hue, value, and chroma were determined using a Munsell color chart. If the soil had been recently cultivated duplicate 1000-cm³ soil samples were collected from the 0- to 10-cm depth. If the soil had not been cultivated, samples were taken from the 0- to 2.5- and 2.5- to 10-cm depth. For this analysis all data were analyzed for the 0- to 10-cm depth by taking a weighted average of samples taken at the 0- to 2.5- and 2.5- to 10-cm depths. One of the samples for each soil was used for biological analysis and was placed in a cooler with ice packs for transport to the lab. The other sample was used for physical and chemical analysis and was sent to the lab without refrigeration.

Samples collected for biological analysis were analyzed for MBC (Tate et al., 1988) using the correction factor of Sparling and West (1988), and potentially mineralizable C (PMC) and PMN on the <2-mm fraction using procedures outlined by Drinkwater et al. (1996) with the following modifications. Forty grams of soil were used in the analysis instead of 10 g, and samples were incubated for 35 d at 25°C instead of 30°C. Detailed descriptions of the methods used for biological analyses are given in the companion paper (Brejda et al., 2000).

Samples collected for physical and chemical analyses were analyzed for sand, silt, and clay content (pipette method), and WSA using screens with 4-, 2-, 1-, 0.5-, and 0.25-mm openings (Kemper and Rosenau, 1986). Aggregate weights were summed from each sieve and divided by the sample weight to calculate total WSA content. Samples were also analyzed for pH (1:1 soil/water), TOC, total N, cation-exchange capacity (CEC), exchangeable Ca, Mg, K, and Na, and acidity. Standard soil survey lab methods (USDA-NRCS, 1996) were used in these analyses. The samples were also analyzed for Mehlich III extractable P (MEP) (Mehlich, 1984) measured using inductively coupled plasma emission spectroscopy. All physical and chemical analyses were done on the <2-mm sieved fraction. Detailed descriptions of the methods used for chemical and physical analyses are given in Brejda et al. (2000).

Statistical Analysis
Factor Analysis
Factor analysis was used to group the 20 soil attributes into statistical factors based on their correlation structure using PROC FACTOR in SAS (SAS Institute, 1989). Factor analysis was performed on standardized variables using the correlation matrix (Tables 1 and 4) to eliminate the effect of different measurement units on factor loadings (James and McCulloch, 1990; Johnson and Wichern, 1992). Factor loadings are the simple correlations between the soil attributes and each factor (Sharma, 1996). The 20 soil attributes analyzed were A horizon value, chroma, and depth; percentage sand, silt, and clay; WSA content; TOC; MBC; PMC; total N; PMN; MEP; pH; CEC; exchangeable Ca, Mg, K, Na; and acidity.

Communalities estimate the portion of variance in each soil attribute explained by the factors. A high communality for a soil attribute indicates a high proportion of its variance is explained by the factors. In contrast, a low communality for a soil attribute indicates much of that attribute's variance remains unexplained. Less importance should be ascribed to soil attributes with low communalities when interpreting variable associations represented by each factor.

The sample points used in this study are also sampled every 5 yr as part of the NRI. As a result, information on land use practices from 1989 through 1996 was available for each sample point. This information was used to place each point into one of four land use categories: (i) continuous crop land, (ii) Conservation Reserve Program (CRP) land, (iii) perennial forages comprised of native range or introduced grasses and legumes used for pasture and hay production, or (iv) forest and woodland. Factor scores from each observation were computed by SAS using the regression method (SAS Institute, 1989; Johnson and Wichern, 1992) and analyzed by analysis of variance using the GLM procedure with the four land use categories as the independent variable.

Discriminant Analysis
Discriminant analysis was used to select the statistical factor(s) that were most discriminating between the four land use categories. The analysis was done using PROC DISCRIM in SAS (SAS Institute, 1989). The covariance matrices for the land use groups were tested for equality at the significance level with the POOL = TEST option. The matrices were unequal in both regions, so the pooled within group covariance matrices and a quadratic discriminant function were used in the analysis (SAS Institute, 1989). Following selection of the most discriminating factor(s), the soil attributes that comprised these factors were also subjected to discriminant analysis to select soil quality indicators for each region. All soil attributes were tested for normality using the procedure of D'Agostino et al. (1990), and non-normally distributed soil attributes were log_e transformed prior to analysis (Brejda et al., 2000b).

	Results

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 
Northern Mississippi Valley Loess Hills
Significant correlation (P < 0.05) was present between 137 of 190 soil attribute pairs for samples from the Northern Mississippi Valley Loess Hills (Table 1). A horizon value, chroma, depth, and percentage sand were negatively correlated with most other soil attributes, the highest being with TOC . In contrast, percentage clay, WSA, TOC, MBC, total N, and PMN were positively correlated with most soil attributes other than those listed above. Cation-exchange capacity was strongly correlated with TOC content . Exchangeable Ca and Mg were strongly correlated with each other and with CEC, TOC, and percentage clay. The large amount of correlation present among the 20 soil attributes indicates they can be grouped into homogenous sets of variables based on their correlation patterns (Sharma, 1996).

Each of the first five factors had eigenvalues greater than one (Table 2) , and were retained for interpretation. These five factors explained >90% of the variance in percentage sand, TOC, total N, CEC, and exchangeable Ca, and 80% of the variance in percentage silt, MBC, PMC, PMN, MEP, pH, and exchangeable Mg, K, and acidity, as indicated by their communalities (Table 2). However, the first five factors explained <50% of the variance in A horizon depth, WSA content, and exchangeable Na. Less importance should be ascribed to A horizon depth, WSA content, and exchangeable Na when interpreting the factors.

The second factor had high positive loadings for percentage silt (0.88) and clay (0.76), and a high negative loading for percentage sand (-0.95) (Table 2), and was termed the soil texture factor. Grouping CEC with the organic matter factor rather than with soil textural properties resulted from the stronger correlation between CEC and TOC than between CEC and percentage clay (Table 1).

The third factor had positive loadings for pH (0.89) and exchangeable Mg (0.71), a negative loading for exchangeable acidity (-0.72) (Table 2), and was termed the soil acidity factor. Grouping these three soil attributes together resulted from strong correlations between soil pH and exchangeable Mg , and acidity (Table 1).

The fourth factor had moderate positive loadings for A horizon value and chroma, and was termed the soil color factor. The fifth factor had high positive loadings for exchangeable K and MEP (Table 2), and was termed the fertility management factor.

Factor scores for all five factors varied significantly with land use (Table 3) . Average scores for the organic matter factor were negative for crop land and positive for land in perennial forages and forest and woodlands (Table 3). Organic matter factor scores were also negative for land in CRP, but the magnitude of the scores were not as large as for crop land. This pattern is consistent with the effects of management on soil organic matter quality (Gregorich et al., 1994).

Soil acidity factor scores were positive under crop land, near zero with land in CRP and perennial forages, and negative under forest and woodland (Table 3). Positive acidity factor scores for crop land resulted from higher soil pH and exchangeable Mg levels, and lower exchangeable acidity levels, probably resulting from lime applications as part of crop production practices with this land use.

Soil color factor scores were negative under forest and woodland and positive under the other three land uses. Forest and woodland had the lowest A horizon value and chroma, indicating darker soil colors, resulting in the lowest color factor scores. The highest soil color factor scores were under land in CRP and perennial forages, indicating lighter soil color. Crop land had intermediate soil color factor scores.

Fertility management factor scores were highest under crop land and land in perennial forages, probably resulting from the application of K and P as part of crop production practices (Table 3). Forest and woodland had the lowest fertility management factor scores, probably because this land use rarely receives K and P applications.

Discriminant analysis of the five factors indicated that the soil organic matter factor was the most powerful in discriminating between the four land use categories, based on the magnitude of their discriminant coefficients (Eq. [1]).

(1)

However, no single factor dominated the discriminant function in this MLRA. This is in contrast to results from the Central High Plains in which the discriminant coefficient for the soil organic matter factor was fourfold larger than the coefficients for soil texture, acidity, and color factors, and more than tenfold larger than the coefficient for the soil P factor (Brejda et al., 2000a).

Discriminant analysis of soil attributes that comprise the soil organic matter factor indicated that PMC, MBC, WSA, and TOC were the most powerful soil attributes in discriminating between the different land uses (Eq. [2]).

(2)

No one or two soil attributes clearly stood out as dominant indicators for detecting changes in land use in this MLRA. This is in contrast to results from the Central and Southern High Plains where two soil attributes were identified for each region as potential indicators because of their sensitivity to change with land use (Brejda et al., 2000a).

Palouse and Nez Perce Prairies
Significant correlation was present between 114 of 190 soil attribute pairs in the Palouse and Nez Perce Prairies (Table 4). As with the Northern Mississippi Valley Loess Hills, A horizon value and chroma were negatively correlated with most soil attributes, whereas WSA, TOC, MBC, total N, and PMN were positively correlated with most soil attributes other than A horizon value and chroma. Cation-exchange capacity was strongly correlated with both TOC and percentage clay .

The first six factors had eigenvalues greater than one (Table 5) and were retained for interpretation. The six factors explained >90% of the variance in percentage sand and silt, TOC, and CEC, and 80% of the variance in A horizon depth, percentage clay, total N, pH, exchangeable Ca, Mg, and acidity (Table 5). However, the six factors explained <60% of the variance in PMC, PMN, and MEP, and <50% of the variance in MBC.

The second factor had positive loadings (0.65–0.80) for percentage clay, CEC, and exchangeable Ca and Mg (Table 5) and was termed the exchangeable bases factor. Exchangeable Ca and Mg were significantly correlated with CEC and percentage clay .

The third factor had positive loadings for percentage sand (0.88) and WSA (0.66), a high negative factor loading (-0.95) for percentage silt, and a weak negative (-0.39) factor loading for percentage clay (Table 5). This grouping resulted from the significant negative correlation between WSA and percentage silt , and positive correlation between WSA and percentage sand (Table 4). This factor represented the soil texture factor for the Palouse and Nez Perce Prairies. It was similar to the soil texture factor for the Northern Mississippi Valley Loess Hills, except that in the Palouse and Nez Perce Prairies it contained WSA and the loading on percentage clay was low (Tables 2 and 5).

The fourth factor had moderate positive loadings (0.57–0.73) for pH, and exchangeable K and Na, a moderate negative loading for exchangeable acidity (-0.58), and represented the soil acidity factor in the Palouse and Nez Perce Prairies (Table 5). The soil acidity factor was similar between the two MLRA (Tables 2 and 5), except that for the Palouse and Nez Perce Prairies, which contained the monovalent bases (K and Na), rather than the divalent bases (Ca and Mg).

The fifth factor had high positive loadings for A horizon value and chroma, and was identical to the soil color factor observed for the Northern Mississippi Valley Loess Hills (Tables 2 and 5).

The sixth factor had a high positive factor loading (0.87) only on A horizon depth (Table 5), and was termed the A horizon depth factor. A horizon depth was not an important soil attribute in the Northern Mississippi Valley Loess Hills.

Factor scores for five of the six factors varied significantly with land use (Table 6) . Only the soil acidity factor did not vary significantly with land use. Organic matter factor scores were lowest under CRP followed by continuous crop land, and highest under land in perennial forages and forest and woodland. This pattern is similar to the pattern for organic matter factor scores in the Northern Mississippi Valley Loess Hills, and probably reflects the effects of management on soil organic matter.

Soil texture factor scores were negative for crop land because of a higher silt content and lower sand and WSA content compared with the other land uses (Table 6). Soil texture factor scores were highest for land in perennial forages and forest and woodland, primarily because WSA concentrations were highest in soil under these land uses (Table 6).

Soil color factor scores were negative for crop land and positive for land in CRP, perennial forages, and woodland (Table 6). Crop land had the lowest A horizon value and chroma (Table 6), indicating darker soil colors. This is opposite to the pattern observed in the Northern Mississippi Valley Loess Hills, where crop land tended to have higher A horizon value and chroma, indicating lighter soil colors (Table 3).

A horizon depth factor scores followed the same pattern as the soil attribute associated with it. A horizon depth was deepest with land in perennial forages, resulting in large, positive depth factor scores, and shallowest in forest and woodland, resulting in large, negative depth factor scores (Table 6). A horizon depth, and depth factor scores were intermediate for crop land and land in CRP.

Discriminant analysis of the six factors indicated the soil organic matter factor, followed by the texture and color factors were the most powerful in discriminating between the four land use categories, based upon the magnitude of their discriminant coefficients (Eq. [3]).

(3)

However, as with the Northern Mississippi Valley Loess Hills, no single factor dominated the discriminant function with the data for the Palouse and Nez Perce Prairies.

Discriminant analysis of soil attributes that comprise the soil organic matter factor indicated that total N and TOC were the most powerful soil attributes in discriminating between land uses (Eq. [4]).

(4)

The discriminant coefficients for total N and TOC were more than fourfold larger than the coefficient for MBC, and 20-fold larger than the coefficient for MEP (Eq. [4]). Both TOC and total N varied significantly with land use with values decreasing in the order: forest and woodland > perennial forages > crop land > CRP (Table 6). Because TOC and total N were highly correlated , they may be redundant as indicators. Because of this redundancy, TOC may be the better soil quality indicator because it influences a wide range of soil functions including infiltration, aeration, water retention, aggregate formation, bulk density, pH, buffer capacity, cation-exchange properties, mineralization, and the activity of soil organisms (Larson and Pierce, 1991; Seybold et al., 1997).

	Discussion

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 
Based on the soil attributes that comprised them, all of the factors identified using factor analysis contribute to one or more soil functions proposed by Larson and Pierce (1991), and therefore are considered to be soil quality factors. The soil organic matter and texture factors contribute to the ability of the soil to accept, hold, and release nutrients and other chemical constituents, accept, hold, and release water to plants and for surface and groundwater recharge, promote and sustain root growth, maintain suitable soil biotic habitat, and respond to management and resist degradation (Larson and Pierce, 1991). The soil acidity and exchangeable bases factors contributes to the ability of the soil to supply nutrients and promote and sustain root growth. The fertility management factor is important in supplying K and P to the plant and promoting root growth. The A horizon depth factor influences seedling growth and the ability of the soil to resist degradation, and the color factor influences soil temperature and thus mineralization rates.

The reader should be aware that the soil quality factors identified in this manuscript and in the previous study (Brejda et al., 2000a) are not unique. Different results may have occurred if a different set of soil attributes had been analyzed, or had we used the covariance matrix or a different rotation in factor analysis. Some potentially important soil quality indicators were not included in these studies. The soil attributes we evaluated were selected by the USDA-NRCS as the soil properties they would consider monitoring in an assessment of soil quality using the NRI. The reason NRCS did not include other potential indicators, such as infiltration, is that the time and labor costs required to measure many other potential indicators were too high, making it infeasible to do on a large number of samples or regional scale. Despite this limitation, the set of 20 soil attributes used in these studies includes most of the indicators recommended in minimum data sets proposed by Arshad and Coen (1992), Doran and Parkin (1994), Kennedy and Papendick (1995), and Larson and Pierce (1991, 1994).

With factor analysis using the covariance matrix, soil attributes with large variances can unduly influence the determination of factor loadings (Johnson and Wichern, 1992). We had no a priori reason to believe that soil attributes with large variances are potentially more important soil quality indicators. Rather, we agree with Schipper and Sparling (2000) that soil attributes with large variability may be poor soil quality indicators because they may be too imprecise for detecting changes in soil quality following changes in land use or soil conservation practices. By using the correlation matrix in factor analysis, in which each variable is standardized to have a variance of one, the unequal variance problem was eliminated.

The purpose of factor rotation is to achieve a simpler factor pattern that can be meaningfully interpreted (Sharma, 1996). There are many potential rotations that can be used, but no rules to follow in selection of a specific rotation. Rather, Sharma (1996) states, "the solution that gives a theoretically more plausible or acceptable interpretation of the resulting factors would be considered to be the `correct' solution." We used the varimax rotation because it results in a factor pattern in which each variable loads highly on only one factor, and because it provided a "theoretically plausible and acceptable interpretation of the resulting factors."

The validity of our results from factor analysis is supported by the consistency in the factor patterns observed in three of the four MLRAs studied (Table 7) . The five soil quality factors identified in the Northern Mississippi Valley Loess Hills, where 75 different soil series were sampled, were identical to the five soil quality factors identified in the Central High Plains where only the Ascalon soil series was sampled (Table 7) (Brejda et al., 2000). The soil attributes that comprised these factors were also similar (Table 7). For the Palouse and Nez Perce Prairies, where 58 different soil series were sampled, six soil quality factors were identified, four of which were similar to the soil quality factors identified in the Central High Plains and Northern Mississippi Valley Loess Hills (Table 7). This suggests that these soil quality factors are common to a wide range of soils and geographic regions.

Despite differences in parent material, climate, and sampling design, the two soil quality indicators selected for the Palouse and Nez Perce Prairies (TOC and total N) were identical to indicators selected for the Central High Plains. Similarly, the two soil quality indicators selected for the Southern High Plains (TOC and WSA) were part of the set selected for the Northern Mississippi Valley Loess Hills (PMN, MBC, WSA, and TOC). Only TOC was selected as a soil quality indicator in all four regions. This result supports our previous conclusions (Brejda et al., 2000a). There may be no universal optimum set of indicators for monitoring soil quality on a regional scale in all regions of the USA. However, if only one soil attribute were used to monitor soil quality with the NRI, TOC appears to offer the greatest potential of all of the attributes we evaluated.Economic Research Service 1997; Soil Conservation Service 1981

Received for publication April 21, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion REFERENCES

 

• Arshad M.A., Coen G.M. Characterization of soil quality: Physical and chemical criteria. Am. J. Altern. Agric. 1992;7:25-31.
• Brejda J.J., Moorman T.B., Karlen D.L., Dao T.H. Identification of soil quality factors and indicators: I. Central and Southern High Plains. Soil Sci. Soc. Am. J. 2000;64:2115-2124 a (this issue).[Abstract/Free Full Text]
• Brejda J.J., Moorman T.B., Smith J.L., Karlen D.L., Allan D.L., Dao T.H. Distribution and variability of surface soil properties at a regional scale. Soil Sci. Soc. Am. J. 2000;64:974-982 b.[Abstract/Free Full Text]
• D'Agostino R.B., Belanger A., D'Agostino R.B., Jr. A suggestion for using powerful and informative tests of normality. Am. Stat. 1990;44:316-321.
• Doran J.W., Parkin T.B. Defining and assessing soil quality. In: Doran J.W., et al. , ed. Defining soil quality for a sustainable environment. Madison, WI: SSSA, 1994:3-21 SSSA Spec. Publ. 35..
• Doran J.W., Parkin T.B. Quantitative indicators of soil quality: A minimum data set. In: Doran J.W., Jones A.J., eds. Methods for assessing soil quality. Madison, WI: SSSA, 1996:25-37 SSSA Spec. Publ. 49..
• Drinkwater L.E., Cambardella C.A., Reeder J.D., Rice C.W. Potentially mineralizable nitrogen as an indicator of biologically active soil nitrogen. In: Doran J.W., Jones A.J., eds. Methods for assessing soil quality. Madison, WI: SSSA, 1996:217-229 SSSA Spec. Publ. 49..
• Goebel J.J., Baker H.D. The 1982 National Resources Inventory sample design and estimation procedures. Iowa State University, Ames: Statistical Laboratory, 1982.
• Gregorich E.G., Carter M.R., Angers D.A., Monreal C.M., Ellert B.H. Towards a minimum data set to assess soil organic matter quality in agricultural soils. Can. J. Soil Sci. 1994;74:367-385.
• James F.C., McCulloch C.E. Multivariate analysis in ecology and systematics: Panacea or pandora's box?. Annu. Rev. Ecol. Syst. 1990;21:129-166.[ISI]
• Jenny H. The soil resource. New York: Springer-Verlag, 1980.
• Johnson R.A., Wichern D.W. Applied multivariate statistical analysis. Englewood Cliffs, NJ: Prentice-Hall, 1992.
• Kemper W.D., Rosenau R.C. Aggregate stability and size distribution. In: Klute A., ed. Methods of soil analysis. Part 1, 2nd ed Madison, WI: ASA and SSSA, 1986:425-442 Agron. Monogr. 9..
• Kennedy A.C., Papendick R.I. Microbial characteristics of soil quality. J. Soil Water Conserv. 1995;50:243-248.
• Larson, W.E., and F.J. Pierce. 1991. Conservation and enhancement of soil quality. p. 175–203. In Evaluation for sustainable land management in the developing world. Vol. 2. IBSRAM Proc. 12 (2). Bangkok, Thailand. Int. Board Soil Res. Manage., Bangkok, Thailand.
• Larson W.E., Pierce F.J. The dynamics of soil quality as a measure of sustainable management. In: Doran J.W., et al. , ed. Defining soil quality for a sustainable environment. Madison, WI: SSSA and ASA, 1994:37-51 SSSA Spec. Publ. 35..
• Lynch J.M., Bragg E. Microorganisms and soil aggregate stability. Adv. Soil Sci. 1985;2:133-171.
• Mehlich A. Mehlich 3 soil test extractant: A modification of Mehlich 2 extractant. Comm. Soil Sci. Plant Anal. 1984;15:1409-1416.
• Nusser S.M., Goebel J.J. The National Resource Inventory: A long-term multi-resource monitoring programme. Environ. Ecol. Stat. 1997;4:181-204.
• Nusser S.M., Breidt F.J., Fuller W.A. Design and estimation for investigating the dynamics of natural resources. Ecol. Appl. 1998;8:234-245.
• Schipper L.A., Sparling G.P. Performance of soil condition indicators across taxonomic groups and land uses. Soil Sci. Soc. Am. J. 2000;64:300-311.[Abstract/Free Full Text]
• SAS Institute. SAS/STAT user's guide. Version 6, 4th ed. Vol. 1. Cary, NC: SAS Inst, 1989.
• Seybold C.A., Mausbach M.J., Karlen D.L., Rogers H.H. Quantification of soil quality. In: Lal R., et al. , ed. Soil processes and the carbon cycle. Boca Raton, FL: CRC Press, 1997:387-404.
• Sharma S. Applied multivariate techniques. New York: John Wiley & Sons, 1996.
• Sparling G.P., West A.W. A direct extraction method to estimate soil microbial C: Calibration in situ using microbial respiration and ¹⁴C cells. Soil Biol. Biochem. 1988;20:337-343.
• Tate K.R., Ross D.J., Feltham C.W. A direct extraction method to estimate soil microbial C: Effects of some experimental variables and some different calibration procedures. Soil Biol. Biochem. 1988;20:329-335.
• Economic Research Service. Agricultural resources and environmental indicators, 1996-1997. Washington, DC: Agric. Handb. 712. U.S. Gov. Print. Office, 1997.
• USDA-NRCS. Soil survey laboratory methods manual. Soil Survey Investigations Rep. 42, Ver. 3.0. Lincoln, NE: National Soil Survey Center, 1996.
• Soil Conservation Service. Land Resource Regions and Major Land Resource Areas of the United States. Washington, DC: U.S. Gov. Print. Office, 1981 USDA-SCS, Agric. Handb. 296..

This article has been cited by other articles:

				F. Zvomuya, H. H. Janzen, F. J. Larney, and B. M. Olson A Long-Term Field Bioassay of Soil Quality Indicators in a Semiarid Environment Soil Sci. Soc. Am. J., May 1, 2008; 72(3): 683 - 692. [Abstract] [Full Text] [PDF]

				A. C. R. de Lima, W. Hoogmoed, and L. Brussaard Soil quality assessment in rice production systems: establishing a minimum data set. J. Environ. Qual., March 1, 2008; 37(2): 623 - 630. [Abstract] [Full Text] [PDF]

				W. K. Jung, N. R. Kitchen, K. A. Sudduth, R. J. Kremer, and P. P. Motavalli Relationship of Apparent Soil Electrical Conductivity to Claypan Soil Properties Soil Sci. Soc. Am. J., May 6, 2005; 69(3): 883 - 892. [Abstract] [Full Text] [PDF]

				S. S. Andrews, D. L. Karlen, and C. A. Cambardella The Soil Management Assessment Framework: A Quantitative Soil Quality Evaluation Method Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 1945 - 1962. [Abstract] [Full Text] [PDF]

				G. Sparling, R. L. Parfitt, A. E. Hewitt, and L. A. Schipper Three Approaches to Define Desired Soil Organic Matter Contents J. Environ. Qual., May 1, 2003; 32(3): 760 - 766. [Abstract] [Full Text] [PDF]

				G. P. Sparling and L. A. Schipper Soil Quality at a National Scale in New Zealand J. Environ. Qual., November 1, 2002; 31(6): 1848 - 1857. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (28) Citing Articles via Google Scholar Google Scholar Articles by Brejda, J. J. Articles by Allan, D. L. Search for Related Content PubMed Articles by Brejda, J. J. Articles by Allan, D. L. GeoRef GeoRef Citation Agricola Articles by Brejda, J. J. Articles by Allan, D. L.