| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Center for Agroforestry, School of Natural Resources, Univ. of Missouri, Columbia, MO 65211
b Dep. of Soil, Environmental and Atmospheric Sciences; School of Natural Resources, Univ. of Missouri, Columbia, MO 65211
* Corresponding author (UdawattaR{at}Missouri.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.01) greater number of pores, number of macropores, area for the largest pore, macroporosity, mesoporosity and significantly lower circularity than soil from the row crop treatment. Soil under trees, grass, and crop areas on average had 207, 87, and 44 CT-measured pores on a 3632 mm2 area, respectively. Soil under the trees had 2.5 and 3.6 times greater number of macropores than grass and crop areas, respectively. Computed tomography-measured number of macropores explained 64% of the variation for Ksat. Computed tomography-measured parameters that were correlated with saturated hydraulic conductivity included macroporosity, mesoporosity, area of the largest pore, macropore circularity, and number of pores. Results showed that CT-measured pore parameters can be used to predict saturated hydraulic conductivity as affected by land management practices. The study also showed that buffer practices improve soil pore parameters related to soil water infiltration.
Abbreviations: CT, computed tomography • Ksat, saturated hydraulic conductivity
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Macropores rapidly channel surplus flow, thereby reducing nutrient leaching through smaller pores (van Noordwijk et al., 1991a) and allowing movement of water and air into the soil. Many investigators have shown that macropore characteristics such as shape, size, orientation, and size distribution affect the rate, flow, and retention of water (Rasiah and Aylmore, 1998). In the absence of macropores, water movement occurs only through smaller pores, voids between aggregates and ped-faces, or voids between grains (Warner et al., 1989). In soils with high clay content and very low infiltration, the absence of macropores can lead to increased copious surface flow. Therefore, differences in porosity among soils need to be quantified to diagnose changes due to agricultural management practices and to develop best management practices (Pachepsky et al., 1996).
Previous studies often used soil bulk density, soil water characteristic curves, tension infiltrometers, and/or porosity measurements to explain differences in infiltration and water movement among different soils and treatments (Heard et al., 1988; Allaire-Leung et al., 2000; Kay and VandenBygaart, 2002). Porosity can be measured using traditional methods such as the soil core method (Anderson et al., 1990), Boyle's Law porosimetry (American Petroleum Institute, 1960), and thin section analysis (Van Golf-Recht, 1982). Porosity determined by traditional methods often lacks detailed information on pore characteristics and sometimes porosity is estimated by indirect procedures (Beven and Germann, 1982). These procedures do not provide information on the spatial distribution of pores (Gantzer and Anderson, 2002) and measurements may be based on observations in two-dimensions (Mooney, 2002). The measurement of temporal-spatial variability of hydraulic conductivity is time-consuming, expensive, and encounters many uncertainties (Suleiman and Ritchie, 2001). Models and equations that predict hydraulic properties and water movement require several constants or conversions to develop better relationships between the predicted and measured saturated hydraulic conductivity (Ksat; e.g., Lin et al., 1996; Lebron et al., 1999).
X-ray CT, first developed in early 1970s for medical imaging (Hounsfield, 1972; 1973), has received increasing attention in recent years in soil and earth sciences. In soil science, CT procedures have been used to examine solute movement (Anderson et al., 2003), porosity (Anderson et al., 1988; Rachman et al., 2005), pore continuity (Grevers and de Jong, 1994), fractal dimension of porosity (Rasiah and Aylmore, 1998; Gantzer and Anderson, 2002), and plant root development (Tollner et al., 1994). In a recent study, Rachman et al. (2005) used X-ray CT to show increased numbers of macropores under stiff-stemmed grass buffers compared with soil under row crop and no-till management. Their results were highly correlated with macropores estimated using water retention data.
Akin and Kovscek (2003) showed a close agreement (±1%) in porosity between CT and volumetrically derived estimates. Computed tomography procedures have been shown to be superior to traditional methods and provide a finer resolution on a millimeter- to micrometer-scale (Gantzer and Anderson, 2002; Akin and Kovscek, 2003; Carlson et al., 2003). By combining cross-sectional images, information can be available on a three-dimensional scale (Mooney, 2002; Carlson et al., 2003) with visualization of many pore phenomena that are otherwise undetectable in standard procedures (Akin and Kovscek, 2003). Two other important parameters, connectivity and tortuosity, cannot be estimated without three-dimensional images. The best-known advantage of CT is its ability to quickly and nondestructively image the interior of a three-dimensional object (Carlson et al., 2003) while retaining connectivity and spatial variation in pores (Al-Raoush, 2002).
The use of X-ray CT for observing differences in soil media as affected by land management practices is relatively new and there are gaps in our knowledge of how this information can be best integrated into decision making regarding best management practices. This relatively new technique can provide additional data on macropore parameters previously unavailable with traditional methods. We hypothesized that agroforestry and grass buffer practices improve porosity by changing the soil pore characteristics within soil. The objectives of this study were to: (i) evaluate differences in CT-measured macropore and coarse mesopore characteristics (number of pores, number of macropores, macroporosity, mesoporosity, area of the largest pore, and circularity) among treatments, and (ii) correlate CT-measured pore parameters with saturated hydraulic conductivity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Thirty-year mean annual precipitation in the region is 92 cm, of which over 66% falls from April through September (Owenby and Ezell, 1992). Mean annual air temperature is approximately 11.7°C with an average monthly low of –6.6°C in February and an average monthly high of 31.4°C in July (Owenby and Ezell, 1992). Snowfall averages about 51 cm per year, and snow can stay on the ground for extended periods.
Sample Collection and Preparation
Undisturbed soil cores from the surface 0- to 8-cm depth were collected on 23 June 2003 using a core sampler. Acrylic plastic sampling rings were 7.62 cm long x 7.62 cm diam., with a 3.2 mm thick wall. Treatments consisted of tree buffers, grass buffers, and row crop areas with six replicates. Samples for the tree and grass buffer treatments were collected from the second and third buffers counting from the southern edge of the watershed (Fig. 1). Samples for the tree buffer treatment were taken under the trees 15 cm from the trunk of six replicate pin oak trees, three in each buffer. Average height of trees in 2003 was 2.8 ± 0.5 m. Average tree diameter at 10-cm from the ground during the sampling year was 5.7 ± 1 cm. Taking a soil sample 15 cm from the base of the tree trunk avoided the disturbed soil adjacent to the tree that may have resulted from planting 6 yr ago. Samples for the grass buffer treatment were taken between two trees (1.5 m from trees) with three replicate locations per buffer. For the row crop treatment, three samples were extracted midway between the second and third buffers and three additional samples midway between the third and fourth buffers. The soil inside the cylinders was secured with two plastic caps at each end and with masking tape. The soil cylinders were labeled, placed in plastic bags, sealed, placed in individual cardboard containers and transferred to the laboratory in a wooden box. Soil samples were stored in a refrigerator at 4°C until analyses were conducted.
The bottom plastic cover was replaced with two layers of fine nylon mesh to secure soils within the cylinder. The top plastic cover was removed and the soil cores were placed in a 15-cm deep plastic tray. The soil cores were slowly saturated from the bottom with a solution containing calcium chloride (CaCl2; 6.24 g L–1) and magnesium chloride (MgCl2; 1.49 g L–1) using a Marriotte system. This concentration has been found to be similar to soil ionic concentrations in claypan soils (Palmer, 1979). After a 24-h saturation period, wet weights were recorded and samples were placed on a –3.5 kPa glass-bead tension table for 24-h for draining. This procedure removed water from macropores and coarse mesopores to allow better image contrast. Samples were weighed again, two plastic end caps were secured with masking tape, and the samples were prepared for transport to the CT scanner. Two phantoms, distilled water in an aluminum tube (outside and inside diam. 2.32 and 1.60 mm) and a solid copper wire (outside diam. 0.55 mm), were attached to the long axis of the Plexiglas cylinder for a standard comparison among scans. Copper has attenuation similar to Mn concretions present in claypan soils.
Scanning and Image Analysis
Computed tomographic image acquisition was conducted using the University of Missouri Hospital and Clinic's Siemens Somatom Plus 4 Volume Zoom X-ray CT scanner. The scan system parameters were set to 125 kV, 400 mAs, and 1.5-s scan time to provide detailed and low noise projections. The field of view, that is, the cross-section dimension, was 100 mm with 512 by 512 mm picture elements (pixels) giving a pixel size of 0.19 by 0.19 mm. The X-ray beam width or "slice" thickness was 0.5 mm producing a volume element (voxel) size of 0.018 mm3. Each soil core was placed horizontally within the scanner so that X-rays intersected the soil core perpendicular to its longitudinal axis. The first scanned image was taken 15-mm distance from the top of the soil core. Four additional scans were taken at 26, 37, 48, and 59 mm from the top of the soil core. The data were stored on a CD for subsequent image analysis.
Pore characteristics of scanned images were analyzed with public domain software Image-J version 1.27 (Rasband, 2002). A 68-mm diam. region was demarcated using Area Selection Tools as the "Region of Interest (ROI)" to exclude voids near the core walls and to minimize beam hardening interference. The region adjacent to the interior wall may have higher porosity due to discrepancy between the radii of the curvature between the soil particles and the acrylic plastic wall (Al-Raoush, 2002). The exterior area was deleted using Clear Outside Tools from the "Edit-pulldown" menu. The two populations that were important in this study to distinguish pores were air-filled pore areas and the other regions within a scan. Segmentation or the separation of the two populations was completed by converting the gray scale image by identifying two populations in the image based on their intensity values (Fig. 2 ). The Threshold Tool was used to distinguish air-filled pores from solids after converting the image into an 8-bit grayscale image. Any soil pore developed as a result of root decay, soil shrinkage, or biological or natural causes was differentiated from the soil matrix. The intensity value (relative attenuation value ranged from 0 to 255) from the water phantoms (38–42, mean = 40) was used as the threshold value to differentiate air-filled spaces and the other regions within a scan with no overlap. Values lower than the threshold value were identified as air-filled pores and values greater than the threshold value were identified as non-pore. The selected value is in between the values used by Rachman et al. (2005) and Gantzer and Anderson (2002). Morphological features and pore scale parameters such as number of pores, pore area, pore perimeter, circularity, and porosity were obtained from the image adjusted by the threshold. Statistics of the image were obtained from Analyze Particles Tool. Pore area was used to estimate pore diameter and to classify pores into macro- (diam. > 1000 µm) and mesopore (diam. 200–1000 µm) categories (Scott, 2000). Pore area and pore perimeter were used to estimate circularity.
|
Statistical Analysis
Differences in pore characteristics among scans along the soil core were statistically compared to evaluate depth and management influences. Statistical analysis of data was completed assuming a completely randomized design with scans at different depths within a core as a split-plot. The PROC GLM procedure in SAS was used to test differences in depth within and among treatments and to compare differences among treatments (SAS Institute, 1999). Means, standard deviations, and differences among means for the measured parameters were determined with PROC MEANS. The MIXED procedure was used to determine differences among treatments within the same depth.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
The decrease in pores with depth could be attributed to better organic matter accumulation and root activity in the surface horizons. Longevity of roots and longer active growing season may have promoted more pores in the soil under permanent vegetation as compared with the crop areas. Van Noordwijk et al. (1991b) found that tree roots change pore-size distribution in soil as roots decay and add more pore volume. During 2003, soybeans were planted on June 19 and samples were collected on June 23, 4 d after planting. Therefore, the soybeans in the row crop treatment did not have developed roots penetrating the surface soil. Similar to our results, Rachman et al. (2005) observed significantly larger number of pores in soils under grass as compared with crop areas. In a previous study on the same watershed, Udawatta et al. (2005b) reported greater root length densities under trees and grass as compared with row crop areas for the entire 1-m sampling depth. In their study, the surface 10 cm of soil under grass, trees, and crop areas contained 87-, 36-, and 28-cm root length per 100 cm3 soil, respectively. The greater number of pores found in this study can be attributed to greater root development and subsequent root decay, addition of soil organic matter, and improvement of soil physical properties due to permanent vegetation as compared with seasonal crop growth.
The CT-measured number of macropores in the row crop treatment was significantly lower than the tree and grass buffer treatments (Fig. 3B, Table 2). There were also significant differences between the tree and grass buffer treatments (Table 2). On average, soil under trees, grass, and crops had 36, 14, and 10 macropores across all five depths. Management practices mostly affect the number and area of large elongated pores (Pachepsky et al., 1996). In Iowa, five times greater infiltration under multispecies riparian buffers compared with a crop site was attributed to a larger number of macropores (Bharati et al., 2002). They stated that the differences were due to greater root decay and fauna activity under buffers. Rousseva et al. (2002) observed that rainfall impact can cause the most variation in macropores due to loss of elongated pores. The slight reduction in macropores in the 15-mm depth relative to the 26-mm depth of the tree buffer treatment could be due in part to the effects of rainfall. At the second depth, the soil had the highest number of macropores for the tree buffer treatment. The number of macropores declined with depth below 26 mm in the tree and crop treatments. The grass treatment had the highest number of macropores at the 37-mm depth and then declined with depth.
Computed Tomography-Measured Macroporosity and Mesoporosity
Computed tomography-measured macroporosity and mesoporosity as influenced by the tree and grass buffers were significantly higher (p < 0.01) as compared with the row crop treatment (Fig. 4
; Table 2). The average CT-measured macroporosity for the tree and grass buffer treatments across all five depths was five times and two times higher, respectively, than the crop treatment (Table 2). The mesoporosity also showed similar differences among the treatments. Seobi et al. (2005) also observed similar macroporosity values by the soil water retention method for the tree buffer treatment. Their values for the crop and grass buffer treatments were higher than this study. The ability to more directly measure macroporosity is the major advantage of the method used in this study. We interpreted the lower macroporosity for the crop treatment in the current study to less root activity and its associated influence on biological activity and macropores.
|
In a recent study, Rasse et al. (2000) found that alfalfa (Medicago sativa) root systems increased total porosity by 1.7%, macroporosity by 1.8%, and saturated hydraulic conductivity by 57%. Their work suggests that dead roots increase connectivity of macropores rather than the macropore volume as the relative increase in hydraulic conductivity is not proportional to the increase in porosity. Studying soil hydraulic properties and pore distribution in soils of a previously forested site, van Noordwijk et al. (1991b) observed that tree roots change pore-size distribution as decaying roots increased macropore flow. In our study, we observed increased macroporosity in soils under tree and grass buffer areas after five growing seasons. The observed higher macroporosity could be attributed to increased rooting activity.
Computed Tomography-Measured Area of Largest Macropores
Since macropores have such a significant impact on water transport, the largest pores were isolated from each scan image and evaluated. Among the three treatments, the tree buffer treatment had the largest pore for soil depths evaluated followed by the grass buffer treatment (Fig. 5
; Table 2). The area of the largest pore at each depth was significantly different between the tree buffer and row crop treatments. However, soil macropores under trees and grass were not different at the two lower depths, but differences were significant for the upper three depths. In the row crop treatment, the average area of the largest macropore across the five depths was 5 mm2 whereas it ranged from 33 mm2 in the tree buffer treatment to 17 mm2 in the grass buffer treatment. The area of the largest macropore did not differ significantly across depth (Table 2). The largest pore area across the five depths for the three treatments varied between 16 and 22 mm2 (Table 2). For both the tree and grass buffer treatments, the largest pore occurred at the third depth while in the row crop treatment the largest pore occurred at the first depth. We speculate this difference could be due to surface desiccation with bigger roots of permanent vegetation tending to develop slightly below the surface.
|
Computed Tomography-Measured Macropore Circularity
Another important parameter used frequently to determine pore characteristics and water movement is pore shape. Circularity of pores is estimated by dividing the product of area of the pore and 4
by the pore perimeter squared (Tuller et al., 1999). Computed tomography-estimated macropore circularity values were significantly larger for the crop treatment compared with the other two treatments (Fig. 6
; Table 2). Larger circularity values imply more circular pores. The tree buffer treatment had the smallest range across all five depths (0.55–0.59). The values for the row crop treatment were the highest and ranged from 0.65 to 0.72. The circularity range across all five depths for the grass buffer treatment was 0.54 to 0.60 with the smallest mean circularity across all five depths. The differences in root densities in the surface horizon may have contributed to the observed differences in circularity among the treatments. A circularity value of one indicates a perfectly circular pore and smaller values are an indication of elongated and non-circular pores. The larger the pore, the higher the probability that the pore is elongated or planar and the lower the probability that it is round (Mermut et al., 1992; Pachepsky et al., 2000). Our data indicate that macropores in the crop area approached more circular shapes as compared with the pores in the other two treatments. Also, circularity values for the tree buffer treatment show that pores were more uniform compared with the other two treatments across the five depths.
|
Pore shape or form was highly correlated with vegetation management (Fig. 6). Tuller et al. (1999) and Lebron et al. (2002) stated that pore shape is important in determining soil hydraulic properties. Pore circularity can be related to water movement as less circular pores have more resistance to flow due to greater pore-wall surface area with respect to the cross-sectional area of the pore. Liquid displacement in cylindrical tubes during drainage is piston-like, leaving no liquid in the cross-sectional area whereas liquid remains in corners of angular and irregular pore shapes. Studying pore shapes on Comly silt loam, Pachepsky et al. (1996) stated that pore area and pore perimeter are related and pores with rugged outlines represented 15 to 30% of total pore number and provided 86 to 98% of the pore area.
Correlation of Pore Parameters and Saturated Hydraulic Conductivity
Saturated hydraulic conductivity varied among the three treatments. The highest values were measured for the tree buffer treatment, followed by the grass buffer and the row crop treatments (Table 1). The value for the row crop area was similar to that found by Seobi et al. (2005); however, the values from this study for the tree and grass buffer treatments were higher. The higher values observed in this study could be due to the differences in the Ksat estimation procedure. Seobi et al. (2005) used a bentonite slurry to plug visible macropores; thus, their hydraulic conductivity values were smaller relative to those from the current study.
Computed tomography-measured pore parameters were compared with measured saturated hydraulic conductivity. As a single variable, the number of macropores explained the largest degree of variability in saturated hydraulic conductivity (Table 3). Computed tomography-measured macroporosity and mesoporosity ranked second and third accounting for 55 and 48% of the variation, respectively. The area of the largest macropore, number of pores, number of macropores, macroporosity, mesoporosity, and macropore circularity in different combinations explained from 67 to 78% of the variation in saturated hydraulic conductivity in linear models with two and three variables. Both circularity and soil bulk density were negatively related with Ksat. Soil bulk density as a single variable explained 83% of the variation in Ksat.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 16, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Kumar, S.H. Anderson, L.G. Bricknell, R.P. Udawatta, and C.J. Gantzer Soil hydraulic properties influenced by agroforestry and grass buffers for grazed pasture systems Journal of Soil and Water Conservation, July 1, 2008; 63(4): 224 - 232. [Abstract] [PDF]
|
|
|
|
 |
|
 R. P. Udawatta, C. J. Gantzer, S. H. Anderson, and H. E. Garrett Agroforestry and Grass Buffer Effects on Pore Characteristics Measured by High-Resolution X-ray Computed Tomography Soil Sci. Soc. Am. J., January 25, 2008; 72(2): 295 - 304. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P. Udawatta, S. H. Anderson, C. J. Gantzer, and H. E. Garrett Influence of Prairie Restoration on CT-Measured Soil Pore Characteristics J. Environ. Qual., January 4, 2008; 37(1): 219 - 228. [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 | |||
