| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, Wheat, Sorghum, and Forage Research Unit, 344 Keim Hall, Univ. of Nebraska, Lincoln, NE 68583
b USDA-NRCS, P.O. Box 2890, Washington, DC 20013
c Dep. Soil, Water, and Climate, 1991 Buford Circle, Univ. of Minnesota, St. Paul, MN 55108
d USDA-ARS, Bushland, TX 79012
e USDA-ARS, National Soil Tilth Lab., 2150 Pammel Drive, Ames, IA 50011
f USDA-ARS, 215 Johnson Hall, Washington State Univ., Pullman, WA 99164-6421
Corresponding author (jbrejda{at}unlserve.unl.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: CEC, cation-exchange capacity • CRP, Conservation Reserve Program • MLRA, Major Land Resource Area • NRI, National Resource Inventory • OC, organic C • PSU, primary sampling unit • SOC, soil organic C
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
However, with the NRI, no soil samples are collected. Rather, data on land use and conservation practices are collected at each point and an estimated rate of soil erosion is calculated using the Universal Soil Loss Equation to determine whether the nation's soils are improving, stable, or degrading (U.S. Congress, Office of Technology Assessment, 1995). Several current environmental issues have stimulated interest in collecting soil samples at NRI sample points to improve NRCS' ability to monitor changes in soil and agricultural resource condition. The soil samples could be analyzed for appropriate physical, chemical, and biological properties that serve as indicators of resource condition. Organic C (OC) strongly influences both soil productivity and quality (Bauer and Black, 1994; Larson and Pierce, 1991; Seybold et al., 1998). Loss of OC from soils as a result of agricultural practices has contributed to increased atmospheric CO2 concentrations and reduced soil quality (Houghton et al., 1983). Increasing OC levels in agricultural soils has been proposed as a mechanism to sequester CO2 from the atmosphere to reduce greenhouse gas concentrations and help improve soil quality and productivity (Bruce et al., 1999; Lal et al., 1999).
In order for Congress to develop legislation for conservation programs that promote practices that increases SOC levels or minimize SOC losses, information is needed on the effects of land-use practices on SOC levels at regional and national scales. To provide this information, there is urgent need to develop robust, statistically based, and efficient methods for monitoring and verifying spatial and temporal changes in SOC at the field, region, and national scales (Post et al., 1999). The NRI may be a tool for quantifying the effects of land-use and conservation practices on SOC levels at a regional and national scale, if it can provide accurate and precise estimates across a wide range of soils and topographic positions. However, it is unknown what sources of variation are important in estimating SOC levels at a regional scale and should be accounted for in NRI assessments.
Soil OC levels are the product of complex interactions between climate, topography, texture, and land-use practices (Parton et al., 1987; Burke et al., 1989; Pennock and van Kessel, 1997). Because of the strong influence of climate on SOC levels, greater precision may be achieved if monitoring is conducted within regions containing similar climatic conditions. Major Land Resource Areas are geographic units of several thousand hectares in extent that contain similar patterns of climate, soils, water resources, and land uses (USDA-SCS, 1981). Major Land Resource Areas are important in agricultural planning at the state, regional, and national levels (USDA-SCS, 1981). Thus, the MLRA offers an appropriate regional scale unit for estimating SOC levels.
Topography, in terms of hillslope position and slope aspect, can affect soil water balance and aeration (de Jong, 1981), significantly influencing SOC levels across the landscape. Honeycutt et al. (1990a)( 1990b) reported effective precipitation and SOC levels often increased in a downslope direction from the summit to lower hillslope positions. Land-use practices may interact with hillslope position, resulting in different SOC levels under the same land use, depending on the position in the landscape. In central Saskatchewan, cultivation and subsequent erosion resulted in substantial losses of SOC from shoulder and backslope positions, but significant gains through deposition at footslope and level depressional areas (Pennock and van Kessel, 1997). In southwestern North Dakota, Aguilar et al. (1988) also reported significant losses in SOC at upper hillslope positions as a result of cultivation-induced erosion and accumulation at lower hillslope positions through redeposition of the eroded materials for soils formed in sandstone, siltstone, and shale parent materials.
Soil texture, particularly sand and clay content, can have a significant effect on SOC accumulation and mineralization patterns (Schimel et al., 1994). Soils high in clay provide both physical and chemical mechanisms for protecting OC from microbial breakdown (Oades, 1988). Conversely, soils high in sand generally have lower OC levels and higher mineralization rates because these protective mechanisms are reduced or absent. However, soil texture may also influence land-use patterns, in which landowners preferentially choose to cultivate soils higher in clay, leaving sandier sites in native vegetation or planting them to perennial forage crops. Thus, in a regional-scale assessment where a wide diversity of soils are sampled, variations in texture both within and between different land uses may need to be adjusted for to accurately assess the effects of land use on SOC content. The objectives of this study were (i) to identify important sources of variation for SOC levels at a regional scale using the NRI and (ii) to determine the precision with which SOC levels can be estimated under different land uses in four different regions of the USA.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
In the Northern Mississippi Valley Loess Hills and Palouse and Nez Perce Prairies, a sample of 200 points were initially selected from the total population of NRI points established in each MLRA during previous surveys. However, 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 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 soil series in the Northern Mississippi Valley Loess Hills, and 58 different soil series in the Palouse and Nez Perce Prairies. The 186 soils sampled in the Northern Mississippi Valley Loess Hills were predominately Alfisols (n = 127), but also included Mollisols (n = 32), Entisols (n = 22), Inceptisols (n = 2), and Histosols (n = 1). The soil series at two sampling points were not named. The 149 soils sampled in the Palouse and Nez Perce Prairies were predominately Mollisols (n = 136), with a few Alfisols (n = 7), Inceptisols (n = 2), Andisols (n = 2), and Entisols (n = 1). The soil series at one sample point was not named.
In the Central and Southern High Plains a sample of 100 points was initially selected from the population of established NRI points, with the restriction that points were confined to a single soil series. A smaller sample size was used in the Central and Southern High Plains because it was assumed that confining sampling to a single series would reduce data variability. In the Central High Plains, sampling was confined to the Ascalon series (fine-loamy, mixed, superactive, mesic Aridic Argiustoll), and in the Southern High Plains, the Amarillo series (fine-loamy, mixed, thermic Aridic Paleustalf) was sampled. These series were chosen because they are benchmark soils and have a wide geographic distribution. If the soil present at the sampling site was not the designated series the point was not sampled. As a result, only 64 points were sampled in the Central High Plains and 47 points were sampled in the Southern High Plains.
Field Data Collection and Soil Sampling
Field crews located each sample point using aerial photographs taken for previous NRI surveys. The correct location of each point was verified in the field using global positioning system technology. At each sample point the land use and predominant hillslope position and slope aspect classes were determined using the model of Ruhe (1975). The land-use information collected at each point was combined with land-use information collected during previous NRI surveys to place each point into three to six land-use classes, depending on the MLRA. In the Northern Mississippi Valley Loess Hills five land-use classes were used: (i) continuous row crops, (ii) row crop–perennial forage rotation, (iii) Conservation Reserve Program (CRP) land, (iv) tame pasture and hayland, and (v) forest and woodland. In the Palouse and Nez Perce Prairies six land-use classes were used: (i) continuous small grains, (ii) wheat–fallow rotation, (iii) CRP, (iv) tame pasture and hayland, (v) native rangeland, and (vi) forest and woodland. In the Central High Plains the sample points were placed in five land-use classes: (i) wheat–fallow rotation, (ii) wheat–row crop rotation, (iii) CRP, (iv) tame pasture and hayland, and (v) native rangeland. In the Southern High Plains, the small sample size resulted in only three land-use classes: (i) continuous cropland, (ii) CRP, and (iii) native rangeland.
Three to five hillslope position classes were evaluated depending on the MLRA. In the Northern Mississippi Valley Loess Hills and Central High Plains the hillslope position classes consisted of: (i) shoulder, (ii) a relatively linear backslope, (iii) a concave footslope, (iv) toeslope, and (v) level terraces and benches. In the Palouse and Nez Perce Prairies, the same hillslope position classes were used, with the exception that the terrace and benches position was combined with the toeslope position, resulting in only four landscape positions in this region. In the Southern High Plains, because of the relatively level topography of the landscape, only three hillslope positions could be readily identified: (i) shoulder, (ii) backslope, and (iii) terrace and benches.
In the Northern Mississippi Valley Loess Hills, Palouse and Nez Perce Prairies, and Central High Plains, eight slope aspect classes were evaluated: (i) north, (ii) northeast, (iii) east, (iv) southeast, (v) south, (vi) southwest, (vii) west, and (viii) northwest. The slope aspect classes were based on the predominant direction the slope faced. However, with a few points in each MLRA a dominant slope aspect could not be determined. These sample points were excluded from the analysis of aspect on SOC levels. Because of the relatively level topography, aspect classes could not be readily identified in the Southern High Plains. Therefore, slope aspect classes were not identified in this MLRA.
At each sample point a 1000-cm3 soil sample was collected from the surface 10 cm. The soil sample was analyzed for sand, silt, and clay content using the pipette method, cation-exchange capacity (CEC) by the ammonium acetate method, pH (1:1 soil/water), and OC concentrations by dry combustion measured with a Leco SC-444 analyzer (Leco Corp., St. Joseph, MI). Soil analyses were done by the USDA-NRCS National Soil Survey Center in Lincoln, NE, using standard soil survey lab methods (USDA-NRCS, 1996). It was impractical to measure soil bulk density (
) in the field because some of the sample points in each MLRA had been recently cultivated. Therefore, a soil bulk density value for each sample point was calculated using the bulk density algorithm provided in the WEPP model (Alberts et al., 1995) where
 | (1) |
 | (2) |
The mass of SOC in each sample was calculated by multiplying the OC concentration by the calculated bulk density value and soil depth.
Statistical Analyses
Variation in SOC levels with land use, hillslope position, and slope aspect was evaluated separately for each MLRA by analysis of covariance using the GLM procedure and Type III sum of squares in SAS (SAS Institute, 1989). To adjust for differences in soil texture under the different land uses or hillslope positions either sand or clay content in each soil sample was used as a covariate, depending on which textural component accounted for the greater proportion of variance. The analysis was run twice. The first analysis was a factorial combination of hillslope position and land use, and the second was a factorial combination of aspect and land use. Hillslope position and aspect were analyzed separately because there were insufficient sample sizes to run a complete three-way factorial. Least square means for significant main effects were compared using the pdiff option in SAS (SAS Institute, 1989). The pdiff option calculates a separate probability value for each pair of means being compared. Means were considered significantly different at the 
= 0.05 probability level.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and neutral to slightly acid pH under all land uses (Table 2). Soils under forest and woodland and row crop–forage rotations had the highest sand content and lowest clay content.
|
|
|
|
and neutral to slightly acid pH under all land uses (Table 2). Soils under native rangeland and forest and woodland had the highest sand content, and soils planted to continuous small grains had the highest clay content. These textural differences support the assumption that landowners tend to leave sandier sites in native vegetation. Land use and clay content of the soil were significant sources of variation in SOC content at the regional scale in this MLRA (Table 3). Soil OC content did not vary significantly with hillslope position, slope aspect, or the interaction of these factors with land use. Soil OC content was greatest under forest and woodland and native rangeland, intermediate under tame pasture and hayland, and lowest under wheat–fallow rotations, continuous small grains, and CRP (Table 4).
The greatest precision in estimating TOC content was with land in wheat–fallow rotation (SE = 2.2 Mg ha-1). This land use also had the largest sample size. The lowest precision in estimating SOC content was with land in perennial herbaceous cover. These were CRP (SE = 5.0 Mg ha-1), tame pasture and hayland (SE = 4.3 Mg ha-1), and native rangeland (SE = 4.1 Mg ha-1). The precision in estimating SOC content in soil under forest and woodland (SE = 3.5 Mg ha-1) was intermediate between the two cropland uses and the three perennial herbaceous cover land uses.
Central High Plains
Soils sampled in the Central High Plains had a sandy loam texture 
and neutral pH under all land uses (Table 2). Soil under tame pasture and hayland had the highest clay content, but variation in clay content with the different land uses was small (range was 13.3 to 17.7%).
Land use, slope aspect, and sand content of the soil were significant sources of variation in SOC content at the regional scale in the Central High Plains (Table 3). The greater effectiveness of sand content than clay content as a covariate was probably a result of the low variability in clay content in these soils (Table 2). Hillslope position and the interactions of hillslope position x land use and aspect x land use were not significant sources of variation in SOC content at the regional scale in this MLRA.
Soil OC content was greatest in north, northeast, west, and northwest facing slopes (Fig. 3) . In contrast, SOC content was significantly lower in south, southeast, and southwest facing slopes (Fig. 3). This pattern may reflect the greater solar radiation received on south and west facing slopes than on north and east facing slopes, resulting in either greater soil temperatures or greater evaporation, both of which may reduce OC accumulation in the soil.
|
The greatest precision in estimating SOC content was achieved with land in wheat–row crop rotations (SE = 0.9 Mg ha-1), despite the fact that this land use had the smallest sample size (n = 7). The lowest precision in estimating SOC content occurred with native rangeland (SE = 3.1 Mg ha-1). Precision in estimating SOC content under wheat–fallow rotation, tame pasture and hayland, and CRP were intermediate between wheat–row crop rotations and native rangeland.
Southern High Plains
Soils sampled in the Southern High Plains had a sandy loam texture under land in continuous cropland and CRP, and loamy sand texture under native rangeland (Table 2). The samples had a neutral to slightly alkaline pH under all land uses (Table 2).
Land use and sand content of the soil were significant sources of variation in SOC content at the regional scale in the Southern High Plains (Table 3). Neither hillslope position nor the interaction between hillslope position and land use were significant sources of variation in SOC content at the regional scale in this MLRA. Soil OC content was greatest under native rangeland, intermediate under CRP, and lowest under continuous cropland (Table 4). However, SOC content was not significantly different in soil under continuous cropland compared with CRP, nor between CRP and native rangeland.
The greatest precision in estimating SOC content was achieved with soil under continuous cropland (SE = 0.5 Mg ha-1), which also had the largest sample size (n = 26). The lowest precision in estimating SOC content occurred with native rangeland (SE = 0.9 Mg ha-1; n = 13), despite the fact that this land use had a larger sample size than land in CRP (n = 8).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A more powerful use of the NRI, however, would be to monitor SOC sequestration or loss on an areal basis with time following changes in land use or adoption of soil conservation practices. Each NRI sample point has a weight associated with it that is constructed to represent a known surface area (Nusser and Goebel, 1997; Nusser et al., 1998). These weights are used to scale-up point estimates made at each PSU to areal estimates for each MLRA. By applying the weighting factors, the SOC estimates made using the NRI could be used to quantify the amount of SOC present in the surface soil on an areal basis within each MLRA. An initial inventory would establish baseline SOC levels and land-use histories at each PSU within a MLRA. Then, future NRI assessments made at the same sample points at 5-yr intervals could be used to verify the amount of SOC that has been sequestered following the adoption of conservation practices designed to increase SOC. The statistically valid and precise estimates of SOC sequestration that may be made using the NRI could be useful in helping the USA meet international treaty obligations. They could not be used, however, to justify C credit payments made to farmers because the location of NRI sample points are not released to the public.
An additional strength of the NRI is that it could be combined with geographic information systems and process-based models to estimate both spatial and temporal SOC sequestration or losses with changes in land use at the regional scale (Paustian et al., 1997; Post et al., 1999). Mitchell et al. (1997) combined information from the 1992 NRI with the EPIC (Erosion Productivity Impact Calculator) model to evaluate the impacts of alternative production systems and soil conservation policies on SOC levels. When used in combination with a process-based model, the measurement of SOC levels at a subset of NRI sample points could be used to both calibrate and continually validate the model results for accurate interpretation and extrapolation.
A potential criticism of this study is that only one soil sample was taken in each PSU, which covers an area of 64.8 ha. Although considerable variation in SOC can be expected in a field of this size, the objective of the study was to quantify SOC levels on a regional scale, not at the field scale. Increasing the number of soil samples collected in each PSU would improve estimation of PSU mean values, but would be relatively inefficient for reducing regional-scale variability and improving the precision with which regional SOC means are estimated. It is the PSU that is the replicate in this study. Improved precision in estimating SOC levels within a region would be best achieved by increasing the number of PSUs sampled in each MLRA, not the number of subsamples collected within a PSU.
A second potential criticism of this study is the shallow depth sampled (10 cm) (Davidson and Ackerman, 1993). This depth was chosen because changes in SOC levels will be most evident at the soil surface. In contrast, Post et al. (1999) recommended a sampling depth of 30 to 40 cm at 7.5- to 10-cm increments, for monitoring SOC sequestration on both cropland and rangeland. Now that important sources of variation have been identified, future NRI assessments could include sampling at a greater soil depth or by genetic horizons, should that be identified as a critical need.
Results from this study indicate that regional-scale estimates using the NRI to quantify SOC levels under different land uses may be improved by adjusting for differences in soil texture at each sample point. A significant amount of variation in all four MLRAs was accounted for by including sand or clay content as a covariate in the analysis. In addition, the confounding effect of different soil textures under different land uses was removed in estimating SOC levels. However, the textural fraction that was important in accounting for this variation differed between the four MLRAs. Thus, in a national NRI assessment, selection of the appropriate textural fraction to use in adjusting land-use means for textural differences may need to be determined separately for each MLRA.
The development and adoption of conservation and management practices to increase SOC levels, in combination with regional and national NRI assessments to evaluate the effectiveness of these practices both spatially and with time, could help improve the quality and productivity of the nation's soils and provide a statistically valid method to verify the amount of SOC sequestered using these practices.
Received for publication June 16, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Chen, D. E. Kissel, L. T. West, W. Adkins, D. Rickman, and J. C. Luvall Mapping Soil Organic Carbon Concentration for Multiple Fields with Image Similarity Analysis Soil Sci. Soc. Am. J., January 11, 2008; 72(1): 186 - 193. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Kucharik Impact of Prairie Age and Soil Order on Carbon and Nitrogen Sequestration Soil Sci. Soc. Am. J., March 12, 2007; 71(2): 430 - 441. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Guo, R. Amundson, P. Gong, and Q. Yu Quantity and Spatial Variability of Soil Carbon in the Conterminous United States Soil Sci. Soc. Am. J., February 27, 2006; 70(2): 590 - 600. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Tan, R. Lal, N. E. Smeck, F. G. Calhoun, B. K. Slater, B. Parkinson, and R. M. Gehring Taxonomic and Geographic Distribution of Soil Organic Carbon Pools in Ohio Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 1896 - 1904. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. S. Homann, S. M. Remillard, M. E. Harmon, and B. T. Bormann Carbon Storage in Coarse and Fine Fractions of Pacific Northwest Old-Growth Forest Soils Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 2023 - 2030. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
in four Major Land Resource Areas sampled for soil organic C content
