SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 9 August 2007
Published in Soil Sci Soc Am J 71:1448-1454 (2007)
DOI: 10.2136/sssaj2006.0372
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
Related Collections
Right arrow Soil Methods/Instrumentation
Right arrow Structure and Properties
Right arrow Soil Analysis
Right arrow Soil Physics

SOIL PHYSICS

Noncontact Shrinkage Curve Determination for Soil Clods and Aggregates by Three-Dimensional Optical Scanning

Till Sander* and Horst H. Gerke

Institute of Soil Landscape Research, Leibniz-Centre for Agricultural Landscape Research (ZALF), Eberswalder Strasse 84, D-15374 Müncheberg, Germany

* Corresponding author (till.sander{at}zalf.de).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The classical saran resin method for determination of the soil shrinkage characteristic (SSC) curves based on volume determination of saran-coated clods is limited by water penetration during submersion, requires correction for the volume of coatings, and does not account for swelling. This study compared a novel noncontact method for SSC curve determination with the classical saran method. Natural, irregularly shaped clods of 18 to 45 g detached from four horizons of a clayey to loamy Chinese paddy soil were capillary saturated to a matric potential of –4 cm. During air drying, the changing clod volume was repeatedly measured by three-dimensional optical scanning and the water content determined gravimetrically. After resaturation of the same clods, measurements were repeated with the saran method. Optical scanning of clods from various positions resulted in sets of point clouds that required comprehensive data processing. The final three-dimensional volume objects were generated by adaptive triangulation. When assuming 0.3 to 0.8 g of water penetration for the saran method and a correction to account for a 20 and 18% loss of mass of air- and oven-dry saran coatings, respectively, the SSC curves of the two methods were relatively similar. Spearman's rank correlation coefficient was 0.88 at the 95% confidence level. Errors for the new three-dimensional method may occur during manual data processing. The new, three-dimensional optical method is suitable to determine the SSC curve of relatively small, irregularly shaped clods with low to medium shrinkage in a wide range of water contents.

Abbreviations: 3-D, three-dimensional • MB, megabyte • MEK, metyl ethyl ketone • SSC, soil shrinkage characteristic


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil shrinkage properties may influence mechanical and hydraulic soil properties directly by changing the pore space and pore geometry (Braudeau and Mohtar, 2004, 2006; Garnier et al., 1997; Giraldez and Sposito, 1985; Tuller and Or, 2003; Kim et al., 1999; Philip, 1969; Smiles, 2000; Chertkov, 2004) or indirectly by formation of shrinkage cracks (Beven and Germann, 1982; Chertkov and Ravina, 2002; Colina and Roux, 2000; Kirby and Ringrose-Voase, 2000; Sander and Gerke, 2007; Messing and Jarvis, 1990; Liu et al., 2003; Wopereis et al., 1994; Greco, 2002).

The soil shrinkage characteristic (SSC) curve quantitatively describes the change of soil volume vs. water content expressed by the moisture ratio (volume of water/volume of solid) and the void ratio (volume of voids/volume of solid; Bronswijk and Evers-Vermeer, 1990). Three shrinkage zones are generally distinguished in a shrinkage curve (Chertkov, 2000; McGarry and Malafant, 1987). In the structural shrinkage phase, the void ratio changes nonlinearly with moisture ratio because water leaves interaggregate pores (or macropores) between mostly saturated aggregates. In the normal shrinkage phase, the soil volume decreases linearly with the loss in water content due to a proportional deformation of the soil without entry of air into the clay pores. While for homogenized clay pastes, the change in soil volume may be equal to that of the water volume, the slope of the shrinkage curve in natural soils often is much smaller than unity (Braudeau et al., 1999). Residual shrinkage proceeds with entry of air also into the clay pores, thus loss of water can correspond to little or no volumetric change. Zero shrinkage may be considered an additional phase of the SSC curve (Bronswijk and Evers-Vermeer, 1990). Overlapping of two phases that correspond to two structural water pools results in curvilinear transitions and was considered an additional phase by Braudeau et al. (2004).

The SSC curve is determined by measuring the volumes of a soil sample at different water contents. For undisturbed samples with well-defined geometry (e.g., cores), the volume can be assessed by displacement transducers (Braudeau et al., 1999; Boivin et al., 2004, Crescimanno and Provenzano, 1999) or by the horizontal deformation from photographs of the cross-sectional top surface by image processing (Peng et al., 2006). These methods are not applicable for soil clods with complex shapes. A classic method for clods is based on Archimedes' principle. Before submerging the sample into water to measure its volume, the sample is covered with thin elastic, semipermeable skins, for instance, saran (Brasher et al., 1966; Bronswijk and Evers-Vermeer, 1990). However, the saran method has a number of limitations. Tunny (1970) found that five coats of saran resin, with 1:7 saran/metyl ethyl ketone (MEK) solution, considerably inhibited the swelling of natural clods when they were coated before initial saturation, while during shrinkage the resin coating did not properly contract with the clod. Schafer and Singer (1976) concluded that the coating (1:5 saran/MEK solution, four coats) did not pull away from the clod during shrinkage when saran was applied to the clod at a water content represented by 338 hPa of capillary suction. They suggested that the smaller water content of the clod (338 hPa in contrast to saturation) during coating probably avoided compaction by handling. No inhibition of swelling by saran was found above 338 hPa, which was explained by the strong swelling pressure of relatively dry clods. In contrast, the swelling pressure is weaker close to saturation, and inhibition of swelling by the coating may occur. The saran method was further developed by Bronswijk et al. (1997), who found a maximum error of about 6% for the void ratio and of about 7% for the moisture ratio. The method, however, is susceptible to errors of water penetration, the impact of the coating, and assumptions of saran properties to account for mass and volume of the coating. Moreover, after covering the sample with saran, the sample cannot further be used for soil physical or hydraulic investigations. Monnier et al. (1973) measured the volume of small aggregates (2–3 mm) by impregnation with kerosene and weighing under water without using coatings, although the influence of kerosene on swelling is unclear and larger macroaggregates might be destroyed during submersion in kerosene.

Advances in measurement technology provide new possibilities to develop new noncontact methods for volume determination. High-resolution computed tomography (Lehmann et al., 2006; Cnudde et al., 2006) is complex and expensive for measurements in short temporal intervals. Laser scanning to assess volumetric changes has been used for measurements of height and width of geometrically well-defined cylindrical samples (Braudeau et al., 2004); however, the hydraulic properties of naturally shaped soil clods can significantly differ from that of artificially shaped samples (Gerke and Köhne, 2002). Crescimanno and Provenzano (1999) found less shrinkage in confined cylindrical core samples than in resin-coated small natural aggregates.

The objective of this study was to test an optical noncontact method for assessing the shrinkage properties for complexly shaped, natural, relatively small (18–45 g) soil clods with small to medium shrinkage. The proposed new method was evaluated by comparing results with those obtained with the standard saran method (Bronswijk and Evers-Vermeer, 1990).


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Samples were taken from paddy rice fields at the Ecological Experimental Station of Red Soil of the Chinese Academy of Sciences, Nanjing, located in subtropical East China (Province Jiangxi). The soil is an anthric saturated Typic Epiaquept (Soil Survey Staff, 2006). The texture of the topsoil down to 42 cm is clay loam. For further information on the sampled sites, see Field 2 in Sander and Gerke (2007).

Small clods of 18 to 45 g oven-dry mass were manually detached from two soil monoliths of about 20-cm width by 25-cm length at the 5- to 23- and 25- to 42-cm depths. Five clods per horizon, in total 20 clods, were sampled from the cultivated A horizon (Sample A, 5–13-cm depth), the plow pan in the upper part of the cambic Bw1 horizon (Sample B, 13–23 cm), the bottom part of the compacted cambic horizon Bw2 (Sample C, 25–32 cm), and the subsoil (Sample D, 32–42 cm). The appearance of the clods from Samples A and D indicated that they were macroaggregates formed by assembles of several aggregates (less significant for Sample D than for A), while clods of Samples B and C probably were single aggregates.

The clods were capillary saturated from the bottom for 35 h on a suction plate at –4-cm pressure head. A small layer of quartz sand was spread on the nylon-membrane-covered disk to enhance the hydraulic contact. Particle density of mixed samples from the soil horizons was determined with a pycnometer (Blake and Hartge, 1986). Every clod was held on a tripod constructed from nails and oven-hardening polymer clay (FIMO soft, Eberhard Faber GmbH, Neumarkt, Germany). A flexible wire held a clod from the top (Fig. 1 ) and kept it stable during measurements when the tripod was moved and inclined (Fig. 2 ).


Figure 1
View larger version (128K):
[in this window]
[in a new window]

  		Fig. 1. Undisturbed soil clod held on a tripod.

 

Figure 2
View larger version (18K):
[in this window]
[in a new window]

  		Fig. 2. Schematic of the three-dimensional optical scanning device. Rotation and inclination of the plate enables a full view of the clods. The tripods with clods are fixed on the plate by screw clamps.

 
The geometry of the 20 clods was assessed during the shrinkage process by three-dimensional (3-D) optical scanning at regular intervals and the mass was determined using a balance (MC1 Sartorius, resolution 0.01 g) for calculation of the moisture ratio. We used the 3-D optical scanning device (hardware) of the Gesellschaft zur Förderung angewandter Informatik e.V. (GFaI, www.gfai.de), Berlin, Germany. The device was composed of a black-and-white camera and a white-light projector (Fig. 2). The measurements basically combined the coded light approach for rough coding with a phase-shift approach for refined coding (Luhmann, 2000, p. 464–472). The coded light approach uses a sequence of switchable, numbered dark and light lines from a projector. Different locations on the measured object are exposed to different combinations of light related to a gray code. Coding is refined for locations with the same gray code by using the sinusoidal distributed brightness of shifting stripes and the wavelength of the light. The lateral (direction approximately parallel to the surface of the object) resolution is about 0.1 mm and distance resolution (direction approximately perpendicular to the surface) 0.04 mm for a section of about 10 by 10 cm.

The tripod with clod was fixed on a movable object table and was measured in up to 14 positions (Table 1) to obtain a complete view of the sample from all directions. For combining data of different views to obtain a single 3-D object, the coordinates were automatically transformed by angles of rotation and inclination into the same single coordinate system. In the following, the term data set defines one 3-D optical volume measurement and the corresponding gravimetric moisture determination that represent a single point on the soil shrinkage curve. The processing of data for a single data set required about 20 megabytes (MB) of memory.


View this table:
[in this window]
[in a new window]

  		Table 1. Position of object table during three-dimensional optical scanning with the optical device shown in Fig. 2.

 
During the drying process, which took 3 d in total, 13 data sets each for Samples C and D and 14 for Samples A and B were obtained for every clod. Time for scanning of a single object was 7 to 8 min, including replacement and weighing of the tripods with held clods, which resulted in temporal intervals of measurements of about 160 min for 20 samples. Small particles that were lost during the drying process were sampled and weighed. For nightly interruptions of the measurements, the objects were secured in a cooler. The cooler box was cooled by ice and kept moist by wet paper towels to reduce evaporation.

Data for the drying of three randomly chosen clods per soil horizon, in total 12 clods, were processed using the proprietary software Final Surface (GfaI, Berlin). Up to two positions per object showed incomplete point clouds due to hardware problems. Thus, nine data sets out of 13 (or 14) x 12 clods were removed because the erratic positions showed only stripes instead of complete point clouds and the remaining positions did not assure a complete view. Erratic points may occur if either the camera or the projector is shaded by the arm or legs of the tripod or by the complex surface of the object. Measured point clouds in the background, from the tripod, and from small roots that stuck out of the clods were identified by hue and arrangement of points and removed manually (Fig. 3 ). Dislocated points that obviously did not belong to the object were removed.


Figure 3
View larger version (51K):
[in this window]
[in a new window]

  		Fig. 3. Data processing of point clouds from three-dimensional optical scanning of Clod A1. The figure shows (a) the original data with erratic points on the tripod, the background, and roots, (b) a processed point cloud after removal of erratic points, and (c) a three-dimensional object after triangulation.

 
Volumetric objects were generated from point clouds by using adaptative triangulation (Hambrecht, 2002). Holes that resulted from removing the legs of the tripods were closed by triangulation. The triangles that fitted to points at the clod surface next to the holes determined the shape of the interpolated surface in the holes by a competitive mesh adaptation algorithm (Hambrecht, 2002). The volume was calculated by summing up all tetrahedric volumes and the surface by summing up tetrahedric areas of the volume objects (Final Surface software).

Measurement of shrinkage of the same clods was repeated with the saran resin method according to Bronswijk et al. (1997) for comparison. A thin cotton thread was tied around the clods that were air dried during the optical measurements. The clods were then resaturated in the same way as for the first measurement but for 48 h instead of 35 h to facilitate swelling of the initially drier clods. The clods did not fully regain the same mass as during the first saturation, but differences were <0.7 g. The rewetted clods were immersed two times into a mixture of Saran F310 (Nordmann, Rassman Ltd., Hamburg, Germany) resin and methyl ethyl ketone (400 g MEK/100 g saran). Submersion was restricted to a few seconds to avoid penetration of water into the coating. The coated clods including the thread were weighed and the volumes were determined by measuring the mass of water displaced by the clods after 1 h, after 1, 2, 3, 5, 7, 10, 13, 16, 20, and 37 d, and finally after oven drying (24 h at 105°C). Volume and moisture ratios of the clods were calculated as described in Bronswijk et al. (1997) using the same equations and assumptions about properties of the saran coating. The mass and volume of saran coatings is needed for correction of measurements and was calculated from the mass of applied solution. Thereby, the mass after hardening was estimated at 0.6 times the mass of the applied saran solution and the mass after oven drying at 0.54 times the mass of the applied saran/MEK solution. The volume was calculated by assuming standard properties of the saran coating for the specific density before (1.5 g cm–3) and after oven drying (1.6 g cm–3). We modified the original correction assuming 0.20% of applied solution for the mass of the coating after hardening and 0.18% after oven drying. For comparison, we used both the original and the modified correction factors. The solid volume was calculated from the mass of saran-coated oven-dry clods and the measured values of the particle density (Table 2), including a correction of the saran coating.


View this table:
[in this window]
[in a new window]

  		Table 2. Mass of near-saturated clods after first wetting (three-dimensional method) and second wetting (saran method). The small differences between the mass of the uncoated clods (second wetting) and the coated ones (determined at given times after coating) indicated that most of the applied saran evaporated within <1 h (and possibly some water). Volumes and masses of coated aggregates are given for the first and last measurements with the saran method. Spearman's rank correlation coefficients (R) are given for the results of the two methods after correction for water penetration and saran coatings.

 
The void ratios of the saran method were linearly interpolated for moisture ratios corresponding to data obtained with the 3-D optical scanning method to enable numerical comparison of the two methods. The correlation of each set of SSC curves measured by the two methods was tested by Spearman's rank correlation coefficient within the range of moisture ratios where both the saran method and the 3-D method SSC curves overlapped. The test is suitable for non-normally distributed data (Zar, 1972).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
During measurements, the rate of water loss decreased from 11.1 to 2.5 g d–1 (arithmetic mean of the 12 clods) for the 3-D method and from 1.0 to 0.1 g d–1 for the saran method (Fig. 4 ). The smaller rates (compared with initial rates), which were observed after about 2.5 d for the 3-D method and after 37 d for the saran method, indicated that the drying process was mostly completed. Storage of clods in the cooler box successfully reduced drying when measurements with the 3-D method were interrupted. This reduction of evaporation enabled an increase in the number of data points for determining the shrinkage curves (Fig. 4).


Figure 4
View larger version (34K):
[in this window]
[in a new window]

  		Fig. 4. Example of drying Clod A1 for the three-dimensional (3D) method without coating and the saran (S) method with coating (without correction for saran resin). Note the different time scales. The two periods with smaller loss of mass in the three-dimensional method indicate that drying was successfully reduced during the night when three-dimensional measurements were interrupted.

 
For both methods, the calculated SSC curves had the typical S-shape (Fig. 5 ) for the residual shrinkage, the structural shrinkage, and a linear part in between. The shapes of curves from both methods were, with some exceptions, similar, but the absolute location of the curves showed more or less constant offsets. Increasing volumes despite loss of mass occurred in a few cases for both of the methods and indicated erratic values. Curves of Samples B, C, and D obtained with the 3-D method showed a tendency for convex behavior in the near-saturated structural shrinkage phase compared with a rather concave behavior of curves from the saran method. Such differences of curve shapes (convex or concave) can be caused by faster drying of clods for the 3-D method than the saran method. Relatively fast evaporation could lead to local heterogeneity of water content, with higher soil moisture remaining in the center than in regions near the surface of clods. The higher heterogeneity of soil moisture might result in different shrinkage phases of the center and the near-surface regions, which combined could produce a modified shrinkage curve for the clod. It was not tested if nonequilibrium conditions occurred for the 3-D method.


Figure 5
View larger version (33K):
[in this window]
[in a new window]

  		Fig. 5. Shrinkage curves of three clods per horizon for the four soil horizons (Samples A, B, C, and D) measured with the three-dimensional method (gray, solid lines) and the saran method (black, broken lines) and 1:1 saturation line. Curves for both methods are based on the solid volume calculated from the solid mass of the coated aggregate after oven drying. The solid mass was corrected by 0.54 times the mass of applied saran.

 
Notice that the different curve shapes are based on relatively small volume differences; i.e., the means and standard deviations of the differences between the first and second volumes were 0.7 and 0.6% of the near-saturated volume for the 3-D method and 0.9 and 0.4% for the saran method, respectively. Further investigations are needed to evaluate whether such small differences are indicative of different shrinkage behavior for the two methods or are simply below the accuracy of the methods.

A possible source of errors during the 3-D method was the manual data processing of the 2142 files (representing all positions in this study). Manual processing might explain discontinuities of the curves. This required deciding whether points clearly represented the clods or belonged to roots or the tripod. All roots had to be removed to assure similar processing of all data sets. The removal of fine objects that stuck out of the clods, such as roots, was important because they could lead to inaccuracies in triangulation and overestimation of the volume. During optical scanning, errors may occur due to reflections of light from wet clods and shadows in depressions on the surface of complexly shaped objects. This was especially true for clods of the A horizon that consisted of microaggregates, which formed relatively complex surfaces (Fig. 3). Visual inspection, however, suggested that the scanning technique detected the natural, structured clod surfaces. Mechanical stress that took place during rotation and inclination of the objects increased the risk of affecting the results by detachment and dropping of soil particles or crumbs. For this study, however, such losses (0.04–0.12 g) were not significant.

As described by Bronswijk et al. (1997), the accuracy of the saran method depends on assumptions about the saran properties, especially the loss of mass and volume during hardening and oven drying, the density of the soil particles, and the disregard of the mass and volume of the thread. Furthermore, a formation of bubbles beneath the saran coating and of saran drops during coating was recognized here (Fig. 6 ). During the short-time submersion of the clods in water, air bubbles sometimes persisted outside of the coated clods in root channels and depressions and might have caused erratic volumes. The change in soil color during submersion indicated that small amounts of water penetrated into or through the coating, especially when the clods were relatively dry. Since some unknown portions of the saran penetrated into the clod surfaces, the correction of measured volumes (of coated clods) to account for the saran coating volume is limited. Considering that the measurements here were thoroughly performed according to Bronswijk et al. (1997), these faults can be considered to be typical for SSC curve determination of carefully handled clods with complex surface geometries using the saran method.


Figure 6
View larger version (61K):
[in this window]
[in a new window]

  		Fig. 6. Soil clods coated with saran resin of the cultivated horizon (Clod A2) and subsoil (Clod D1). Arrows show examples of encapsulated air bubbles under the coating of Clod A2 and deep pores in Clod D1. Deep pores may eventually not be perfectly coated, which may allow water to enter the clod. Air bubbles were preferably retained in depressions at the bottom of the coated clods during submersion.

 
The mass of clods for the saran method had to be corrected by the mass of applied saran/MEK solution including a correction factor for mass loss of applied solution during hardening and oven drying. The solid volumes, which were used for calculation of SSC curves for both methods, were determined from the coated oven-dry clods and required a correction for the oven-dry saran coatings. Unsuitable values of the correction factor may lead to a horizontal offset between the curves of the two methods. The relatively small increase of the mass of clods with and without coating (Table 2) indicated that most of the applied saran/MEK solution (and maybe also some water) evaporated within the first hour, which suggests a smaller correction than the one recommended by Bronswijk et al. (1997). Assuming that all MEK quickly evaporated during air drying and about 10% of saran during oven drying, only the mass of applied saran (20% of applied solution) was considered for modified correction during drying and 90% of applied saran (18% of applied saran solution) after oven drying and the corresponding volumes of the saran coating. We note that the choice of the correction remains uncertain because of the unknown contribution of evaporated water during rapid mass loss of wet clods. Furthermore, drying of the saran coating may proceed during the entire measurement period. For comparison, we used both the original factors of Bronswijk et al. (1997) and the modified correction factors (Fig. 5 and 7 ). For the saran method, the modified correction led only to a slightly increased moisture ratio while for the 3-D method, the moisture ratio decreased more strongly. In total, these corrections shifted the SSC curves obtained with both methods to a corresponding range of moisture ratios, with the three shrinkage phases ranging between approximately the same moisture ratios for clods from the same soil horizons (Fig. 7). The average Spearman's rank correlation between the SSC curves obtained with the two methods was R = 0.88 (Table 2, mean values for the 12 pairs of SSC curves) at a confidence level of >95% (one exception for Clod C2). Values >0.5 indicate a strong correlation.


Figure 7
View larger version (33K):
[in this window]
[in a new window]

  		Fig. 7. Shrinkage characteristics of three clods per horizon for the four soil horizons (Samples A, B, C, and D) measured with the three-dimensional method (gray, solid lines) and the saran method (black, broken lines) and 1:1 saturation line. Curves for both methods are based on the solid volume calculated from the mass of the coated aggregate after oven drying. The solid mass was corrected by 0.2 times the mass of applied saran for hardening and 0.18 for oven drying. The modified correction shifted the curves to approximately the same ranges of moisture ratio. Vertical lines roughly outline residual (left), normal (middle), and structural (right) shrinkage phases.

 
The remaining differences in void ratios of the two methods could be explained by assuming water penetration during submersion of clods for the saran method. Penetration of small amounts of water per area could result in significant quantities, considering the relatively large surfaces of 36 to 57 cm2 (determined from 3-D objects). A correction of the displaced mass by adding an estimated 0.3 to 0.8 g (3–8% of corrected solid volume) of water resulted in an excellent agreement between all corresponding SSC curves obtained with the two methods (Fig. 8 ). When using the water penetration correction and the modified correction factors, the moisture ratios were shifted away from the 1:1 line. Offsets were 0.13 ± 0.08 (mean ± standard deviation) for the 3-D method and 0.13 ± 0.06 for the saran method, compared with 0.06 ± 0.07) for the 3-D method and 0.06 ± 0.04 for the saran method without modified correction factors and water penetration. For the SSC curves without modifications, the offsets (even 0 for Clod B3, see Fig. 5) appear to be too small to represent the actual air-filled volumes, considering that the clods were initially wetted at equilibrium with a –4-cm pressure head, and thus were not completely saturated. Offsets from the 1:1 line could additionally result from the effects of entrapped air in disconnected pores (i.e., pseudo-saturation, Braudeau et al., 1999), which supports our assumptions underlying the correction-caused decrease in the moisture ratios.


Figure 8
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 8. Shrinkage curves of three clods per horizon for the four soil horizons (Samples A, B, C, and D) measured with the three-dimensional method (gray, solid line) and saran method (black, broken line), when assuming penetration of 0.3 to 0.8 g of water during submersion of coated aggregates, and using the modified correction. Note the different scale for Sample B. Vertical lines roughly outline residual (left), normal (middle), and structural (right) shrinkage phases.

 
The time consumption for measurements during the optical scanning (about 8 min) is slightly greater than that for the classic method (about 2 min) per measurement of mass and volume. The drying process took <3 d for uncoated compared with about 40 d for coated clods (Fig. 4). The software for analyzing the scanning data was not yet adapted to the processing of several objects at a time, thus data processing took about 5 h for a single SSC curve and about 15 to 20 h of computing for triangulation on an AMD 2800+ with 1024 MB RAM.

In addition to the advantage of noncontact measurements, the optical method provides 3-D data for visual comparison and additional inspection of processed point clouds, for instance, to evaluate possible disturbances during the measurement, to study anisotropic shrinkage, and to calculate morphological parameters of the clods. In return, technical simplicity and 40 yr of experience (Brasher et al., 1966) are important advantages of the saran method while providing sufficient accuracy for strongly swelling soils.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The new optical 3-D method for volume determination succeeded in measuring the shrinkage characteristic of irregularly shaped soil clods with little to medium shrinkage at a wide range of water contents. For the same soil clods, the SSC curves obtained with both the 3-D optical scanning method and the standard saran method were mostly similar when assuming water penetration during the saran method and modifying the correction that accounted for the mass and volume of the saran coating. Small remaining discrepancies in curve shapes between the results of the two methods might be caused by faster evaporation during the 3-D method; however, these differences were based on values that are close to the limits of accuracy of the measurements. The 3-D method allowed determining the volumes of unconfined clods without disturbances such as water penetration or effects of the coating. Additionally, morphological parameters of geometry and surface roughness of the natural clods, which affect mechanical, chemical, and hydraulic properties of aggregated soil, can be derived from such 3-D data.

The 3-D method might be further improved by developing algorithms to eliminate subjectivity during data processing and to decrease time consumption. Fine objects like roots should be removed from the clods before measurements. In contrast to established standard methods for SSC curve determination, the optical method appears to be promising for future research, for instance, on repeated shrinkage and swelling, measurement of swelling during infiltration, evaluation of the influence of coatings on natural unconfined aggregates, or for visualization of anisotropic shrinkage.


   ACKNOWLEDGMENTS
 
This project was financially supported by the Deutsche Forschungsgemeinschaft, Bonn (GE 990/5-1 and 5-2). We thank Prof. Dr. Zhang Bin and Li Jiangtao (Chinese Academy of Sciences, Nanjing) for valuable support during soil sampling and Robert Bartsch (Leibniz-Centre for Agricultural Landscape Research [ZALF], Müncheberg) for his help in data processing. The support of Lothar Paul, Jan Hambrecht, and Maik Bohnet (Gesellschaft zur Förderung angewandter Informatik e.V., Berlin) for enabling the 3-D optical scanning is gratefully acknowledged.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication October 31, 2006.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
B. G. de Almeida, A. P. da Silva, S.R. Raine, and G.C. Figueiredo
Evaluation of Automotive Varnish as a Coating for Density Measurements of Soil Clods
Soil Sci. Soc. Am. J., March 1, 2009; 73(2): 449 - 452.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
A. M. Rossi, D. R. Hirmas, R. C. Graham, and P. D. Sternberg
Bulk Density Determination by Automated Three-Dimensional Laser Scanning
Soil Sci. Soc. Am. J., September 30, 2008; 72(6): 1591 - 1593.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Sander, T.
Right arrow Articles by Gerke, H. H.
Related Collections
Right arrow Soil Methods/Instrumentation
Right arrow Structure and Properties
Right arrow Soil Analysis
Right arrow Soil Physics

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