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


     

Published online 6 May 2005
Published in Soil Sci Soc Am J 69:807-815 (2005)
DOI: 10.2136/sssaj2004.0191
© 2005 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 Similar articles in ISI Web of Science
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 ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
Related Collections
Right arrow Soil Physics
Right arrow Dryland Soils
Right arrow Inverse Procedures/Parameter Estimation

Soil Physics

A Laboratory Method for Determining the Unsaturated Hydraulic Properties of Soil Peds

Darren G. Meadowsa,b,*, Michael H. Younga and Eric V. McDonaldc

a Desert Research Institute, Univ. and Community College System of Nevada, Las Vegas, NV 89119
b Hydrologic Sciences Program, Univ. of Nevada, Reno, NV 89532
c Desert Research Institute, Univ. and Community College System of Nevada, Reno, NV 89512

* Corresponding author (darren{at}dri.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The ability to estimate soil hydraulic properties of undisturbed soil peds is limited. Most laboratory methods repack samples into soil cores, thus destroying the structure. Most field methods average over multiple peds precluding measurements of individual peds. This article details a method for determining unsaturated soil hydraulic properties of individual highly structured soil peds. The procedure is based on the evaporation method. The peds were first coated in paraffin wax, with the top surface left open to the atmosphere, and then encased in expandable foam to provide extra support. After saturating, both mass (from digital balance) and soil water potential (from tensiometer) data were recorded every 5 min as the ped dried until the soil water potential reached approximately –70 kPa. The van Genuchten parameters ({alpha}, n) and saturated hydraulic conductivity (Ks) were estimated with inversion modeling, using the soil water potential data and final water content. Residual and saturated water contents, {theta}r and {theta}s, respectively, were measured independently. A mini-permeameter provided independent estimates of Ks. We compared the areally weighted average of our optimized hydraulic properties with results from a field infiltrometer test conducted at the exact location from where the peds were collected. The two methods compared well, though some disparity existed between the estimates of n and Ks, likely indicating the effects of interped cracks sampled with the tension infiltrometer. The method is potentially valuable for geostatistical and scaling studies because the smallest structural unit, the soil ped, can be sampled and analyzed.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
WATER FLOW THROUGH structured soils can substantially impact local water balances, contaminant transport, recharge, and plant-available water. This is typically due to the prevalence of macropores in the near surface soil horizons. Luxmoore (1981) defined these macropores as having an equivalent diameter of >1 mm. Macropores have been shown to contribute significantly to water flux (Parker and van Genuchten, 1984; White, 1985; Watson and Luxmoore, 1986; Lin et al., 1996). For instance, Watson and Luxmoore (1986) estimated that 96% of the water flux occurred through only 0.32% of the soil volume at their field site. Bouma and Dekker (1978) showed significant flux through roughly 2% of the total vertical interped regions. Beven and Germann (1982) give an excellent review of macroporosity and its role in water movement.

Significant flux can occur as bypass flow through the interped region (Nobles et al., 2004). This is because the saturated hydraulic conductivity of the cracks that separate individual peds can be orders of magnitude greater than that of the matrix itself, thus leading to highly accelerated rates of water movement and contaminant transport. However, for flow to occur in these cracks, the soil water potential must be at or near saturation. Flow within soil cracks, for example, can be initiated when rainfall intensity exceeds the soil's infiltrability.

Many laboratory methods exist for determining the water content–water potential [{theta}({psi})] and the hydraulic conductivity functions [K({psi})] (Dane and Topp, 2002). Popular techniques, such as outflow and evaporation methods, traditionally require that soil material (either disturbed or undisturbed) be contained in fixed diameter columns. Repacking unconsolidated granular soils may or may not provide representative hydraulic properties, depending on the original degree of soil structure and development. However, repacking highly structured soil peds into columns inevitably destroys ped structure, substantially altering the hydraulic properties from natural conditions. This alteration is most significant on the wet end where flow is largely structurally controlled (Jury et al., 1991). Error in estimates of hydraulic conductivity in the wet range would be particularly important because the majority of water flow occurs here. Our method allows hydraulic properties to be determined on intact, undisturbed soil peds, thus removing errors that could arise from repacking disturbed soil material.

The method uses the hydraulic property representation of van Genuchten (1980), which relates soil water content ({theta}) to soil water potential ({psi}):

[1]
where {Theta} equals relative volumetric water saturation, {theta}r and {theta}s are residual and saturated volumetric water contents, respectively, {alpha} and n are fitting parameters, h is the soil water potential, and m = (1 – 1/n), 0 < m < 1. Hydraulic conductivity is related to soil water potential [K({psi})] (Mualem, 1976; van Genuchten, 1980):

[2]
where Ks is the saturated hydraulic conductivity.

These equations are amenable to computer-aided optimization, where an objective function is defined in terms of the available experimental data, and parameters are optimized by minimizing the differences between the model and the experimental results. Optimization methods were first used to estimate flow parameters roughly 25 yr ago (Zachmann et al., 1981; Kool et al., 1985, Parker et al., 1985). Kool et al. (1987) review a number of inverse modeling techniques as applied to soil science. Eching and Hopmans (1993) presented an example of estimating the soil hydraulic properties from a transient, multistep outflow experiment. Simunek et al. (1998a) estimated soil hydraulic properties using both numerically generated data and experimental data from an evaporation experiment. Their results compared favorably with the simplified Wind method (Wendroth et al., 1993).

Previous investigators have developed methods to measure properties of intact soil peds, but the methods were limited to saturated hydraulic conductivity. McKenzie and Dexter (1996) described a technique for measuring the saturated hydraulic conductivity of a soil aggregate, which they adapted from Semmel et al. (1990). However, this method does not obtain unsaturated hydraulic properties. Leeds-Harrison and Youngs (1997) subsequently described a method to estimate the hydraulic conductivity of soil aggregates from field-determined sorption values measured with a miniature ring infiltrometer. However, their method also is not amenable to obtaining the {theta}({psi}) and K({psi}) functions.

The paucity of laboratory methods for determining the hydraulic properties of structured soils usually relegates one to field-based approaches, such as the tension infiltrometer (Ankeny et al., 1991; Reynolds and Elrick, 1991; Jarvis and Messing, 1995; Simunek and van Genuchten, 1996; Evett et al., 1999) or the ring infiltrometer (Youngs, 1987; Elrick et al., 1995). However, if large variations in hydraulic properties exist over small distances, then these methods may not capture the degree of variability, because the scale of measurement is larger than the inherent variability of the properties being characterized. Infiltrometers have circular footprints of widely varying sizes, ranging from a 4-mm (Leeds-Harrison and Youngs, 1997) to 20-m diam. (Youngs et al., 1996). In the case of larger diameter infiltrometers, a single measurement will include many individual peds, averaging those properties and masking small-scale variabilities. The intrinsic assumption in conducting infiltrometer experiments on structured surfaces is that the infiltrometer is sufficiently large to capture the interped variability in hydraulic properties. However, methods are not readily available to confirm this assumption. Furthermore, because extreme values of Ks are masked by the infiltrometer readings, potential flux into deeper soil layers could be underestimated.

The method we describe is largely based on laboratory evaporation experiments that have been used for soil hydraulic property determination for 35 yr. Various permutations of this approach exist (Wind, 1968; Tamari et al., 1993; Wendroth et al., 1993; Halbertsma and Veerman, 1994; Simunek et al., 1998a). Combining the laboratory evaporation experiment with an inverse procedure to optimize hydraulic parameters is an approach shown to be effective by a number of researchers (e.g., Ciollaro and Romano, 1995; Romano and Santini, 1999).

The purpose of this article is to present a laboratory method for determining the unsaturated soil hydraulic properties of individual soil peds. The objectives of this research are to (i) develop and test the laboratory method; (ii) use that method to study soils collected from a field site in the Mojave Desert; and (iii) compare the results to a larger scale infiltrometer experiment that estimated the properties of the same soil material.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field work was conducted on an approximately 100000-yr-old (kA) desert pavement surface within the Mojave National Preserve, CA, which is roughly 120 km south-southwest of Las Vegas, NV. The surface soil is characterized as a vesicular A horizon (Av) of approximately 5- to 10-cm thickness, and is known to exhibit platy and columnar structure. This soil is classified as a Calcic Paleargid (McDonald, 1994), though no series name is known to exist. The clay mineralogy of the Av horizon on this surface is not dominated by one type and is extremely varied (E.V. McDonald, unpublished data, 1994).

To begin the study, a tension infiltrometer (Soil Measurement Systems, Tucson, AZ) experiment was conducted, with data collected similar to that described by Casey and Derby (2002). The infiltrometer test was run at four different soil water potentials: –1.1, –0.7, –0.4, and 0.0 kPa. These soil water potentials were not chosen for any specific reason but meant only to provide a range of potential. The parameters describing the water retention and hydraulic conductivity functions were subsequently estimated from these outflow data using the parameter estimation routine found in HYDRUS-2D (Simunek and van Genuchten, 1996; Simunek et al., 1998b). This approach was shown to be effective by Young et al. (2004) on the same structured soils.

Following the infiltrometer experiment, the soil underneath the disc was carefully excavated so that the entire sampled area (approximately 30 cm diameter) was sampled. This area was comprised of 22 individual soil peds (Fig. 1), which were returned to the laboratory for further investigation. The effective diameter of the peds ranged from slightly <4 to 8 cm.



View larger version (196K):
[in this window]
[in a new window]
  		Fig. 1. Photograph of the excavated soil from underneath the tension infiltrometer. Individual soil peds are clearly delineated. In the field, these soil peds were connected and part of a contiguous desert pavement surface. Once removed from the support of surrounding peds, they slumped to the position seen in this figure. Ring is 20 cm in diameter.

 
The process to determine hydraulic properties on individual peds was begun by first drying them for 24 h at 105°C. Irregularly shaped portions of the peds were carefully shaved with a razor blade when needed to create a near cylindrical shape. A piece of paper towel was cut to the approximate shape of the upper ped surface, where it was tied to the ped with string. The paper towel allowed easy removal of the wax on the upper ped surface and minimized disturbance to the sample. Each ped was then submersed briefly in liquid paraffin wax until fully coated. The bulk density of each ped was determined (Blake and Hartge, 1986) and then used to convert gravimetric water content to a volumetric basis during the subsequent experimentation (Note that no visual changes to the ped shape were observed during the drying or waxing processes). The tops of the peds were sliced off exposing only the upper ped surface. To provide extra support and to prevent slumping during the experiment, the waxed peds were then encased in expandable foam, which molds itself to the waxed surface of the ped before solidifying, forming a rigid wall.

Peds were saturated by first placing them into a chamber held under vacuum for roughly 3 h to remove entrapped air. The vacuum chamber was then pressurized with N2 to further encourage the removal of remaining air trapped inside the ped. A small piston pump inside the chamber was used to slowly saturate the ped at a flow rate of approximately 2 to 4 mL min–1, depending on ped size. Water was applied at a single location on the ped surface allowing air to escape as the wetting front impinged. Once saturated, the ped was weighed again to determine its water content. Tap water was used for both field and laboratory experiments to maintain consistency between the hydraulic property measurements.

A tensiometer of 3-mm diam. (Soil Measurement Systems, Tucson, AZ) was inserted into the center of the saturated ped (longitudinally), and equipped with a pressure transducer (model PX170, Omega Engineering, Inc., Stamford, CT) to measure soil water potential. Visible cracking or splitting of the saturated peds was not observed in any samples. The entire assembly was then placed onto a digital balance (Explorer, Ohaus Corp., Florham Park, NJ) (Fig. 2). The ped was allowed to evaporate until the soil water potential reached approximately –70 kPa, usually within 4 to 8 h. Soil water potential data were collected at 5-min intervals using a data logger (model 23X, Campbell Scientific, Inc., Logan, UT). Data from the balance were collected using a computer running WinWedge 32 (version 3.0, TAL Technologies, Philadelphia, PA).



View larger version (132K):
[in this window]
[in a new window]
  		Fig. 2. Photograph of the experimental setup.

 
Because previous work suggests that continuous outflow experiments do not provide realistic estimates of Ks (Durner et al., 1999), we independently estimated Ks using a mini-permeameter (model M12, Decagon, Inc., Pullman, WA). These small diameter (approximately 3 cm) devices provide data on the cumulative infiltration into soils as a function of time for an instrument-specific soil water potential of –0.2 kPa. The benefit of these permeameters is that they are small enough to be placed on the surface of individual soil peds. To do so, a thin layer of moist sand was applied to the ped surface to improve hydraulic contact. A ring stand was used to keep the permeameter upright. The permeameter tests were run on air-dry peds, and cumulative outflow was recorded as a function of time by means of a pressure transducer (model PX170, Omega Engineering, Inc.). Experiments were run for approximately 1 h. Some of the larger peds were run slightly longer. HYDRUS-2D (Simunek et al., 1999) was used in an axisymmetric geometry to invert the outflow data for estimating Ks, the only parameter optimized in this inversion (all other parameters were fixed using the optimized values from the inversion of the evaporation data).

Numerical Experiment
Upon completion of the evaporation experiments, HYDRUS-2D was used for estimating the van Genuchten (1980) parameters described above in Eq. [1] and [2]. In most cases, the peds were modeled as 1-D columns due to their semi-cylindrical shape. However, HYDRUS-2D was occasionally used so that irregular shapes (sloping upper boundary or rougher vertical faces) could also be examined. The initial condition was defined in terms of soil water potential. A linear distribution with depth was assumed, using the initial tensiometer reading as a calibration point and extrapolating above and below it. No-flow boundary conditions were imposed on all sides except the top, where a constant flux was used (Simunek et al., 1998a). The flux was known from the change in ped mass from the balance. The objective function of the inverse problem was defined in terms of soil water potential, recorded with the tensiometer, and the final water content of the ped, obtained using the known mass of the wetted ped minus the water that evaporated during the experiment.

The model domain was adjusted for the characteristics of each ped. Ped height was measured to the nearest 0.1 mm, as was the location of the tensiometer. The area of the exposed ped surface was calculated in two ways. First, the ped volume was calculated from the bulk density measurement. It was then possible to calculate the ped surface area by dividing the volume by its height. The area was also calculated based on physical measurements of the exposed surface. Both methods typically gave similar results (data not shown).

Parameters optimized in the inversion (shown in Eq. [1]) included {alpha}, n, and Ks. The parameter {theta}s was known from the experimental setup, but was allowed to vary within a narrow range to account for uncertainty in our estimates from the laboratory setup. Values of {theta}r were approximated by taking an average value of the gravimetric water content of the surface soil after one month with no precipitation, and multiplying that by the individual bulk densities of the peds.

Once the hydraulic parameters were determined for each ped, we then visually reconstructed the soil surface by graphically placing the individual peds into their original positions. Some peds were moved slightly to account for dislocation during excavation and transportation to the laboratory to better reflect their positions in the field. The experimentally derived Ks (or other parameters) were then assigned to each ped. This enabled us to calculate an areally weighted mean for all hydraulic parameters and to compare that value with the bulk values estimated from the field tension infiltrometer. Areally weighted averages were calculated using arithmetic means for all parameters except for Ks, which was computed using a geometric mean.

Following the infiltration and numerical experiments, the peds were destroyed and homogenized for particle-size analysis using the Laser Light Scattering technique (model Saturn DigiSizer 5200, Micromeritics Instruments, Norcross, GA).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Table 1 shows texture, bulk density, and other experimental information for each ped, and Table 2 shows the van Genuchten parameters from Eq. [1] and [2] fitted using the above method. Recalling from above, values of {alpha} and n were obtained by optimizing the results of the desorption experiments. Values of {theta}r and {theta}s were known with sufficient confidence that {theta}r was not optimized and {theta}s was varied over a narrow range; Ks was obtained by optimizing the results of the sorption experiment (potential hysteresis effects will be discussed below). Using the bulk density values in Table 1, and calculating the porosity (upper value for {theta}s) it was found that the observed value of {theta}s was an average of 0.088 cm3 cm–3 below the calculated porosity. A slight trend toward higher differences was seen but similarities between the mean and the median (0.088 versus 0.0827, respectively), and a near zero skewness indicate that the differences were normally distributed. The differences between the calculated and observed {theta}s are due in part to the saturation method, which was done from the surface, but also from the presence of air vesicles in the ped, a diagnostic feature in Av horizons. Therefore, it is possible that the presence of air vesicles will lead to lower than expected water contents in field situations, especially for well developed peds similar to those used in these experiments.


View this table:
[in this window]
[in a new window]
  		Table 1. Experimental information on soil peds.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Hydraulic properties of individual peds determined by inversion of laboratory data.

 
We noted that the tensiometers often indicated zero soil water potential at the start of the experiment, even though the ped was not fully saturated. This apparent inconsistency can be explained by the likelihood that the ceramic cup of the tensiometer was in contact with both saturated and partially saturated soil regions, and therefore equilibrated to conditions in the saturated regions.

Figure 3 shows the graphically reconstructed area underneath the field infiltrometer. A total of 22 individual soil peds were recovered, tested, and represented here using Ks data. The black circle in the figure represents the outline of the 20-cm diam. tension infiltrometer. The summed area of the individual peds within the circle equals 91% of the total area of the circle. The results indicate a substantial degree (greater than two orders of magnitude) of interped variability in Ks over a distance of only 20 cm. Distinct zones of high and low Ks are evident as well. The figure also shows the relative area associated with the conductivity values. A trend toward a dominantly lower hydraulic conductivity is apparent, though it is somewhat masked by the high ped-to-ped variability. That nearly as many peds exist in the narrower, low conductivity range as in the higher conductivity range, which is ten-fold wider, indicates the predominance of the low conductivity peds within the sampled area. We expect that moving the zone of investigation would ultimately yield different patterns of conductivity, according to the intrinsic heterogeneity of the soil surface.



View larger version (49K):
[in this window]
[in a new window]
  		Fig. 3. Digitally reconstructed image of area underneath tension disc infiltrometer. Figure delineates individual peds and displays the respective values of Ks as determined from permeameter inversion. Units are cm d–1. Figure also shows the location of the 20-cm diam. tension disc.

 
Figure 4 shows the reconstructed area sampled by the tension infiltrometer, showing values of {alpha} obtained from the evaporation experiments. Like Ks, there is a large amount of variability in {alpha} values. Visually comparing Fig. 4 and 5, and evaluating correlation data from Table 3, shows that Ks and {alpha} are significantly correlated (r = 0.71 at p < 0.001). Correlation between these parameters is expected given that {alpha} is related to the size of the largest, continuous pore present (i.e., approximately the inverse of the air-entry pressure). Others have observed this positive correlation as well (e.g., Wang and Narasimhan, 1992; Zhu and Mohanty, 2003; Young et al., 2004). Although the peds have substantial percentages of fine-grained material, the flow parameters appear to be dominated by the internal structure of the ped rather than the particle-size distribution. The platy internal structure of the individual peds, combined with the presence of abundant vesicles, seem to control flow in the well-developed peds that we sampled. This observation may explain the lack of correlation between flow parameters and particle-size distribution (Table 3).



View larger version (47K):
[in this window]
[in a new window]
  		Fig. 4. Digitally reconstructed image of area underneath tension disc infiltrometer and representation of van Genuchten's {alpha} parameter. Units are cm–1. Figure also shows the location of the 20-cm diam. tension disc.

 


View larger version (47K):
[in this window]
[in a new window]
  		Fig. 5. Digitally reconstructed image of area underneath tension disc infiltrometer and representation of van Genuchten's n parameter.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Correlation coefficients among grain size fractions, bulk density, and hydraulic parameters.

 
Figure 5 shows values of n obtained from the evaporation experiments. Results indicate that a statistically significant negative correlation (r = –0.67, p < 0.001) exists between n and loge Ks. In our case, the increased development of structure within the ped likely diminished the homogeneity of the pore-size distribution, but without affecting the potential presence of larger pores within the peds that could lead to higher Ks. The many vesicles or channels within the well-developed peds, some of which are continuous (note high positive correlation between {alpha} and Ks) may channel flow through the ped, thus producing a soil with higher saturated conductivity, but with a wider pore-size distribution and a lower n.

Figure 6 depicts water retention and unsaturated hydraulic conductivity curves using properties obtained from the field infiltrometer test versus those obtained from the weighted means based on the laboratory measurements. Weighted mean values were calculated as:

[3]
where is the average parameter value found in Fig. 6, Y is the value obtained from laboratory and numerical experiments on individual peds, a is the area associated for the corresponding ped, A is the sum of areas for all peds excluding the interped region, and j is the individual ped number ranging from 1 to the maximum number of peds (in this case 22 total). All parameters were averaged using an arithmetic mean, excluding Ks, which was averaged using a geometric mean. Shapiro–Wilk tests (Shapiro and Wilk, 1965) were conducted to test for normality of both untransformed and loge transformed data for all parameters. All of the parameters except for Ks were better represented as normal distributions; Ks was better represented as a log-normal distribution. The agreement between the laboratory and field methods is good. The only parameter that shows a large deviation is n, where the weighted mean value (n = 1.21) is more typical of the finer texture of the individual peds. The value obtained from the field tension infiltrometer (n = 2.56) is more reflective of the presence of soil cracks that filled suddenly as the infiltrometer test proceeded, increasing flow from the infiltrometer reservoir, and yielding sharper inflections on the retention curve. Because the infiltrometer test was conducted at soil water potentials greater than –1.2 kPa, large changes in water flow rates toward the wet end of the retention curve could lead to preferentially higher n values that would otherwise not be expected given the silty/clayey texture of the soil. Thus, Fig. 6 shows the dramatic impact that soil cracks can have on the hydraulic properties of highly structured soils.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 6. (A) Water retention curves and (B) hydraulic conductivity functions using parameter values found in C. Solid lines are produced from values obtained from the field infiltrometer test; dashed lines are produced from the areally weighted mean parameters of the individual peds. Note that the solid lines are from a sorption experiment, whereas the dashed lines are from a desorption experiment.

 
The results in Fig. 6 also show a difference in Ks values between the two methods. The differences are not large, but it supports the general supposition that water flow in highly structured soils is controlled by the presence of soil cracks or regions of the soil horizon found between otherwise intact peds under near-saturated conditions. In the structured Av soils discussed here, 91% of the surface of the soil horizon sampled by the field infiltrometer was populated by structured peds (aped), and the other 9% represents either interped regions or portions of intact peds that sloughed during excavation, transport to the laboratory, or subsampling of the individual peds (aunsampled).

To study how the 9% unsampled area affects overall hydraulic conductivity for the entire area, we calculated Ks for the unsampled regions by equating the Ks obtained from the field tension infiltrometer to the weighted geometric mean Ks of the ped and unsampled regions,

[4]
where At is the total area sampled by the infiltrometer. Solving for Ksunsampled yielded a value of 1300 cm d–1, or more than 100 times higher than the weighted mean ped Ks. Therefore differences between Ks obtained from the field tension infiltrometer measurement, and the weighted mean ped Ks from laboratory tests can be explained by the different soils being sampled. The tension infiltrometer sampled the entire area underneath the plate, including the interped regions, and the laboratory measurements focused only on the peds themselves.

The ability to measure hydraulic properties on ped-scale samples also helps to examine the relative influence of macropore regions on flux. For example, using the approach of Watson and Luxmoore (1986), data from the infiltrometer test and Poisseuille's equation, we calculated the maximum possible number of macropores using,

[5]
where µ is the dynamic viscosity of water, Km is the macropore conductivity, {rho} is the density of water, g is gravitational acceleration, and R is the minimum pore radius of a given soil water potential (or pressure) range. The effective macroporosity was calculated using {theta}m = N{pi}R2. Using Eq. [5] we calculated that 24% of the flux occurred through 0.0006% of the soil volume (Table 4). Furthermore, our calculations show that 94% of the flux occurred through only 0.006% of the soil volume. It is clear that a very small portion of the soil contributed to most of the total flux. These results suggest, as did the results of Watson and Luxmoore (1986), that macropores are a dominant hydrologic feature at near-saturated conditions and need not involve a large proportion of the total soil volume to be important.


View this table:
[in this window]
[in a new window]
  		Table 4. Macroporosity estimates based on tension infiltrometer experiment and Poisseuille's equation.

 
We attempted to more fully investigate the presence of macropores based on the value of {alpha} obtained from the laboratory evaporation and field tension infiltrometer experiments (0.0537 and 0.0543 cm–1, respectively). If one assumes that {alpha} is equivalent to the inverse of the air-entry value, then an average {alpha} = 0.054 cm–1 implies an air-entry value of –18 cm, and no differences in outflow rate from the tension infiltrometer would have been observed at the range of soil water potentials used in the experiment. Our observation of higher flow as potentials approach zero underscores the conceptual disconnect between the physical flow process and the way in which the process is parameterized.

We noted large standard errors for the final Ks estimates produced by inversion modeling using data from the mini-permeameter. These large standard errors have been noted in the past, and are likely due to a flat objective function near the minimum (J. Simunek, personal communication, 2004). Because Ks was the only parameter floated in this inversion, the solution was particularly sensitive to changes in it. Determining the magnitude of Ks was thus relatively simple. We evaluated the sensitivity of the fitted Ks value by altering the initial guesses by at least ±10% and reinitiating the model. In each case the fitted Ks value settled onto the same value, providing some assurances that the value was unique.

In contrast to Wind (1968) who used multiple tensiometers to monitor soil water potential, the relatively small size of the peds forced us to use a single tensiometer. Our experimental design may have impacted the opportunity to identify accurate estimates of Ks, but we note that other experimentalists have observed similar difficulties with estimating Ks. For example, Simunek et al. (1998a) found that inversion results of Ks exhibited much more uncertainty than the other fitted parameters, and Durner et al. (1999) found that Ks cannot be reliably estimated from continuous outflow experiments.

Hysteresis Issues
A potential concern with this method is that Ks is estimated from a sorbing experiment (mini-permeameter), and the other hydraulic parameters are estimated using a desorbing experiment (evaporation). We recognize that hysteresis could be a factor when combining these parameters into a full soil description. The effect of hysteresis, though, is typically more prominent in coarse-textured soils at lower suctions (Hillel, 1998, p. 161), and less so in finer-textured soils. A diagnostic feature of Av soil peds is that texture is dominated by finer earth fractions (particle sizes <50 µm diam.), which facilitates structure and soil aggregation. The peds described in this study were very fine grained: the average clay plus silt fraction of the 22 peds was 79.7%, and the sand fraction was 20.3% (see data in Table 1). Moreover, the model fitted the sorption data (for Ks) very well by holding constant the other van Genuchten parameters ({alpha}, n, {theta}s) at the values estimated from the inversion of the evaporation data ({theta}r was also constant based on independent data as described previously). With these conditions in mind, we believe any error introduced by mixing the experimental directions will not be substantial. Unfortunately, no independent methods are available to verify the results of the laboratory experiments on the individual peds, and so our conclusion is accompanied with some degree of conjecture.

Other Potential Issues
As with all laboratory procedures, the method presented here has potential limitations. First, to estimate Ks and thus the hydraulic conductivity function, the ped must be sufficiently large for the mini-permeameter to be placed entirely onto the soil surface (e.g., effective ped diameter larger than 3 cm); water movement into smaller peds is at a higher risk of encountering the boundaries. Although these boundaries are accounted for in the construction of the model domain, the flow becomes increasingly complicated and thus more difficult to simulate as the wetting front impinges on the potentially irregularly shaped boundaries. Thus, the larger the ped, the more likely that flow can be simulated without encountering issues of wall flow. This limitation does not affect the evaporation experiments, which are initially saturated and are assumed to dry vertically.

Though peds are carefully prepared at the start of the experiment, the evaporation method could be affected by the presence of irregular ped edges. When present, these edges could cause flow to depart from one-dimensionality. The increased complexity of the flow system may not always be accounted for by the simulation, even when using a two-dimensional model. The experimentalist thus needs to carefully prepare the soil peds to remove obvious irregularities.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
In this article, we described a method for determining the unsaturated soil hydraulic properties of individual, structured soil peds. This is a rapid method, and the required materials are typically present in most modern soil physics laboratories. The method is comprised of two parts: an evaporation and an infiltration experiment. Although results show that inversion of the evaporation data yields reliable estimates of the van Genuchten parameters (based on parameter values that are "typical" of the particle-size distribution and comparable with field based measurements), our experiments consistently overestimated Ks. It was thus necessary to estimate Ks using a mini-permeameter, which yielded more realistic values. The approach of using two procedures for analyzing a single ped added a small amount of complexity to the overall analysis. However, use of the mini-permeameter is straightforward and the test was completed in approximately 60 min; moreover, because the mini-permeameter test is used to obtain a single parameter (i.e., Ks), the objective function can be determined using outflow data only.

This advance is relevant for a variety of applications. For example, this method is ideal for verifying the applicability of certain field measurements that span multiple peds, like larger tension infiltrometer plates, and for quantifying small-scale variability in hydraulic properties that larger-scale measurements cannot quantify. Because the most basic structural unit of soil (i.e., the ped) can be analyzed, the results could be used to study upaling techniques, as the scale of measurement increases from the ped to the pedon level. Contaminant transport and ecological investigations may find this method useful to identify the potential presence of high conductivity peds that could provide rapid pathways into the subsurface. Moreover, because we were able to measure the same soil, including interped cracks and excluding them, we clearly illustrated the impact that the cracks have on the soil's hydraulic properties.


   ACKNOWLEDGMENTS
 
Funding for this research was provided by U.S. Geological Survey under the Grant 104b program and the Center for Arid Lands Environmental Management of the Desert Research Institute. We also acknowledge the comments and suggestions of the anonymous reviewers.

Received for publication June 9, 2004.


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




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. Mertens, R. Stenger, and G. F. Barkle
Multiobjective Inverse Modeling for Soil Parameter Estimation and Model Verification
Vadose Zone J., August 24, 2006; 5(3): 917 - 933.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
D. G. Meadows, M. H. Young, and E. V. McDonald
Estimating the Fine Soil Fraction of Desert Pavements Using Ground Penetrating Radar
Vadose Zone J., May 26, 2006; 5(2): 720 - 730.
[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 Similar articles in ISI Web of Science
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 ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Meadows, D. G.
Right arrow Articles by McDonald, E. V.
Related Collections
Right arrow Soil Physics
Right arrow Dryland Soils
Right arrow Inverse Procedures/Parameter Estimation

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality