| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Desert Research Institute, Univ. and Community College System of Nevada, Las Vegas, NV 89119
b Desert Research Institute, Univ. and Community College System of Nevada, Reno, NV 89512
c Currently, Hydrologic Sciences Program, Univ. of Nevada, Reno, NV 89532
d Currently, Dep. of Geosciences, Boise State Univ., Boise, ID 83725
* Corresponding author (michael.young{at}dri.edu)
1 The lower case letter v formally stands for the presence of plinthite (Soil Survey Staff, 1998); however the v has been extensively used in a substantial number of studies of desert soils that possess horizons with abundant vesicular pores. 
Received 2 November 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
w using Wooding's method, as the soils aged. No clear trends in Ks or w were detected in the underlying horizon, indicating that the controlling feature at these sites, in terms of water entry, was the Ks of the surface (Av) horizon. Soluble salt concentrations within the profile indicated reduced infiltration with increased pavement development. Results showed that surface age can be used as an excellent predictor of saturated hydraulic conductivity (r2 = 0.9254). Further, results suggest that Av horizon development represents a key process controlling water cycling, potentially influencing ecosystem function in arid lands. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Desert pavements consist of a surface layer of closely packed gravel that overlies a thin (3–10 cm), fine-grained, gravel-poor, vesicular A (Av) soil horizon.1 Desert pavements are prominent features in arid environments and can be found on a variety of landforms of significantly diverse ages ranging from Holocene to Tertiary (Bull, 1991; Cooke et al., 1993). Pavements have been used in subdividing and correlating Quaternary alluvial fans for studying neotectonics and Quaternary climate change (McFadden et al., 1998; Bull, 1991; McDonald et al., 2003). Other research has shown that age is an important consideration in development of desert pavements, especially in areas downwind of source zones for aeolian deposited material (McDonald, 1994; McDonald et al., 1996; McFadden et al., 1998). Pavements tend to be more prevalent and more strongly developed on older surfaces where aerosolic clay- and silt-sized particles are deposited on the surface and are subsequently translocated downward into the soil profile. Increasing accumulation of aerosolic fines with time enhances the development of a highly structured Av horizon consisting of distinct columnar-shaped peds, ranging in diameter from about 3 to 8 cm, that part to platy peds, ranging in thickness from 0.2 to 1 cm. The formation of desert pavement and the Av horizon is extremely slow and can take from 4000 to 10000 yr to become established.
Hydraulic properties of structured soils have been studied for some time, primarily in environments different from the arid Mojave Desert. Some studies have shown that pedological development of clayey soils can explain dynamic water flow through cracks and other macropores (Lin et al., 1998, 1999; Bouma and Wosten, 1979; Jarvis and Messing, 1995). Others have sought to understand the pore classes that contributed the majority of flow (e.g., Watson and Luxmoore, 1986), many of which were also in clayey soils. The study described here is unique with respect to the significantly drier climate where the soils developed, the general lack of swelling clays in these arid climates, the difference in the time needed to develop the structure, and the fragility of these surfaces. Their disruption can have a potentially dramatic impact on water balance.
The understanding of how soil hydrology affects soil development spans a number of practical environmental applications. Knowledge of soil dynamics and the consequent linkages to ecosystem development will facilitate improved designs of evapotranspiration disposal covers, ultimately leading to stable and maintenance-free features on the landscape (Shafer et al., 2004). Soil hydrology plays an extremely important role in near-surface water balances and the associated responses in water-limited ecosystems. McDonald et al. (1996) demonstrated that the interrelated control of soil development and climate variation on soil water flux explains temporal variations in the depth of soil carbonates. Hammerlynck et al. (2002) used soil water potential data and desert shrub responses in Mojave Desert soils of different age and parent material to show that plant responses were strongly affected by soil age. These two studies (conducted at the same sites used in the present study) concluded that the depth of water infiltration was partly responsible for soil and ecosystem effects; however, these studies did not comprehensively quantify the hydraulic properties of the Av horizon. Briones et al. (1998) showed how interspecies competition increased when supplemental irrigation was used on study plots in the Chihuahuan Desert and how available soil water was partitioned among existing dominant species, with no interspecies competition. Porporato et al. (2002) developed a probabilistic framework showing how statistical distributions of soil hydraulic conductivity can be used as a tool for estimating ecosystem responses to different soil water balances. These results provide insight into ecosystem evolution in desert environments and the strong linkage between soil development, water cycling, desert ecosystem function, and desert landscape management and restoration.
Ecosystems are highly dependent on soil recharge in water-limited desert environments (Eagleson, 2002). However, detailed knowledge of the hydrological behaviors of different soils and geomorphic settings is needed to advance our understanding of many ecological processes, ranging from local to landscape-level ecosystems. We are aware of limited investigations on infiltration into soils of variable pavement development (Brown and Dunkerley, 1996; McDonald, 1994; McDonald et al., 1996), and these studies primarily focused on measuring infiltration capacity. The research reported herein extends the work of McDonald (1994) and seeks to provide a strong foundation to better understand the role of pavements in the hydrologic cycle of arid environments. We quantify, for the first time, how hydraulic properties covary with time in arid environments where desert pavements form.
The objectives of this work are to (i) evaluate how pedological development in near-surface soil horizons in an arid alluvial fan complex can affect soil hydraulic characteristics and (ii) to compare the use of Wooding's equation and inverse modeling for evaluating hydraulic conductivity in highly layered soils. These objectives were approached through field tension infiltrometer studies, soil sampling, and laboratory analyses of soil texture, water content, and soluble salt concentrations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The tests described here were constrained on soils composed of mixed plutonic parent material. General descriptions are presented in Table 1. Detailed descriptions and interpretation of soils measured in this investigation are provided in McDonald (1994) and McDonald et al. (1996)(2003). The sites chosen for measurement and analyses differed primarily in age, thus constituting a chronosequence from recent to approximately 100 000 yr old (100 ka). Soil surfaces were classified according to the nomenclature QfX, where Q represents Quaternary Period, f represents fluvial or fan, and X represents the subclassification that ranges from 8 to 1, denoting relative age (8 is most recent and 1 is oldest). For this research, soil surfaces of Qf8, Qf7, Qf6, Qf5, and Qf3 were used. These designations correspond to soil surfaces with ages of 0.05, 0.50, 4, 10, and 100 ka, respectively, as determined by McDonald et al. (2003) using local radiometric dates and soil-stratigraphic correlations to date deposits in nearby areas. Figure 1 is a conceptual diagram showing how the soil surface evolves with time both structurally and texturally. The increase in the thickness and structural integrity of the Av horizon is an aggradational process, where fines are transported by wind from upwind sources and then deposited. Over many millennia, the aeolian processes build up the fines content (silt plus clay) while the clasts slowly move upward (McFadden et al., 1987, 1998).
|
|
Data from the infiltrometers (Soil Measurement Systems, Tucson, AZ) were collected from differential pressure transducers (Casey and Derby, 2002) at intervals between 15 s and 1 min, depending on the logger unit and soil surface. Manual readings were taken periodically to verify operation and to better identify when the intake rate was at or near steady state. At that time, the infiltrometer was reset to a higher (less negative) pressure level. Four to five pressure steps were used for each test, typically at levels set to approximately –1.2, –0.9, –0.6, –0.3, and 0 (saturation) kPa.
At the conclusion of each test, the infiltrometer and contact sand were removed and another sample was taken for bulk density and final water content. Samples were taken with a soil ring (5-cm diameter, 2.5-cm height) pressed into the soil near the center of the infiltrated area. The ring was removed using a hand trowel and leveled appropriately to better define the sample volume. A soil pit was also excavated immediately adjacent to the infiltrometer test locations. Samples were collected at 5- to 10-cm interval depths throughout the upper 50 cm of the profile. Samples were stored in sealable, plastic bags to prevent water loss and were returned to the laboratory for further analyses. Particle-size distribution was determined using a Laser Light Scattering technique (model Saturn DigiSizer 5200, Micromeritics Instruments, Norcross, GA). Bulk density was determined using the clod method of Blake and Hartge (1986). Soluble salt (Ca, Mg, K, Na, and SO4 as S) content was determined for samples from the soil pit profiles. One gram of dried soil was placed into a plastic bottle with 25 mL of distilled water and agitated for 24 h on an orbital shaker at 100 rpm. The liquid was then decanted, filtered (0.45 µm), and acidified to pH <2 using trace-grade HCl, before analyzing with an inductively coupled plasma spectrometer (Thermo Intrepid Radial ICP-EOS, Franklin, MA).
Data Analysis
Data were analyzed using a manual, semiempirical method (Wooding, 1968) and a numerical inverse method (
imunek et al., 1999). The different methods were used to evaluate whether better estimates of hydraulic conductivity would be worth the extra postprocessing efforts required in the numerical method and because inverse modeling provides estimates of soil water retention properties, while Wooding's analysis does not. Wooding's analysis relies mostly on the data collected toward the end of each pressure step when the cumulative flux rate is constant in time and provides estimates of the saturated hydraulic conductivity and conductivity–water potential function, K(h). Inverse modeling uses all the data and results in estimates of the conductivity and water content–water potential functions, 
(h).
Vertical infiltration is initially governed by capillarity or sorptivity of water into the soil matrix, containing both vertical and horizontal components. Later-time infiltration becomes gravity driven and linear with time as soil capillarity forces are reduced, indicating that infiltration is at steady state. Based on Gardner's exponential K(h) function (Gardner, 1958), a three-dimensional, analytical solution for steady-state infiltration from a circular source was derived by Wooding (1968):
 | [1] |
w is a parameter (cm–1) affected by the pore-size distribution, and r is the infiltrometer radius (cm). The first term inside the bracket represents vertical gravity flow and the second term inside the bracket accounts for lateral movement due to capillarity. The resulting equation contains two unknowns, Ksw and w. Using the infiltrometer through a range of pressures, paired values of flux {q(h)} and pressures (h) are obtained. These known values are input for a nonlinear least-squares regression routine to solve for the two unknowns by minimizing error through iterative solutions (Logsdon and Jaynes, 1993).
In addition to using Wooding's solution for obtaining Ksw and w, a complete set of hydraulic properties was obtained through numerical inversion of the cumulative infiltration data (mL s–1) and the final water content. Inverse modeling was done with the HYDRUS-2D program (imunek et al., 1999). HYDRUS-2D is an axisymmetric finite element code that iteratively optimizes parameters found in the van Genuchten (1980) form of the soil water retention curve, and in the hydraulic conductivity equation derived by Mualem (1976) as modified by van Genuchten (1980). The retention curve has the form
 | [2] |
 | [3] |
e is the effective volumetric water content, s is the saturated volumetric water content, r is the residual water content, n and m are empirical parameters where the expression m = 1 – 1/n, vg (cm–1) is an empirical fitting parameter similar, though not equal, to the inverse air entry value, and Ksvg is the saturated hydraulic conductivity (cm s–1).
The conceptual model consisted of a single-layered system for the younger soils and a multilayered system for older soils where the Av horizon had developed or was clearly present. In the former case, only one optimization was done to characterize the soil. In the latter case, two simulations were done. The first simulation optimized the parameters of the subsurface (B) horizon only, using data from that experiment. The second simulation fixed the parameters for the B horizon in a layered simulation where parameters for the surface (Av) horizon were optimized. This reduced the number of parameters to be optimized in any single simulation, thus increasing the potential for a unique set of parameters for each layer. This approach was similar to that suggested by imunek et al. (1998).
All conductivity data were normalized to 20°C. The correction factor was derived from known variation of water viscosity and density with temperature (Lide, 2001). An exponential relationship was obtained using TableCurve (Version 1.12, Jandel Scientific, San Rafeal, CA), and all saturated hydraulic conductivity data were adjusted accordingly. The correction factor, C, was
 | [4] |
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Figure 3
is a comparison of the two approaches, and the graph shows that both methods produced very similar conductivities. One outlier was identified using inverse modeling, but overall the data very closely grouped around the 1:1 line; r2 = 0.861 (n = 26) without the outlier and r2 = 0.784 (n = 27) with the outlier included. An analysis of variance of the two data sets indicated no reason to suspect that the data were sampled from different populations (F = 0.2126, = 0.05). No improvement in model results was found when sites with an Av surface horizon were represented as a two-layered system. This finding is supported by the post sampling, which showed that wetting fronts did not penetrate beyond the Av horizon in older soils. Soil development apparently led to a higher water holding capacity. Therefore, the single-soil modeling approach for young soils was valid because no significant soil horizonation had developed, and it was valid for older soils because the wetting front remained within a single layer, the Av horizon. The results indicate that Wooding's analysis provided excellent estimates of hydraulic conductivity functions in this environment, but if estimates of water retention curves are needed, then inverse modeling is required.
|
w parameter, used in Eq. [1], also shows a clear decrease with increasing age in the surface horizon (Fig. 4B); again, no trend was observed in the subsurface horizon. Higher values of w lead to faster declines in unsaturated hydraulic conductivity with decreasing soil water potential (or decreasing water content). Thus, the approximately threefold reduction of w from 0.297 to 0.085 cm–1 from Qf8 to Qf3, respectively, will lead to substantially different K(h) functions and potential for fluid-dominated flow. For example, if unsaturated hydraulic conductivity is calculated for the observed range of w and for uniform Ksw of 100 cm d–1, the difference of hydraulic conductivity at a soil water potential equal to –10 kPa would exceed nine orders of magnitude, from 1.29 x 10–11 to 0.019 cm d–1. Fluid flow at 10–11 cm d–1 would be considered very low to negligible in many practical applications. These results highlight the importance of quantifying the dependence of K(h) on soil water potential (or soil water content) in predicting hydraulic behavior of these soils.
|
w and the sand percentage (r = 0.80, n = 15), which reflects the faster reduction in conductivity due to enhanced soil drainability. Correlation between vg and the soil textural components also fits the conceptual model that younger soils have more sand, higher drainability, and higher hydraulic conductivity. Overall, the conceptual model of hydraulic behavior in variably saturated soil was represented well in these analyses.
|
|
This work demonstrates that pedogenic development leads to large differences in hydraulic properties of the surface soil. These differences will affect infiltration rates during precipitation events, and subsequently, the depth of percolation and evaporative losses. To evaluate deep percolation potential for these five profiles, water-soluble salt concentrations were determined for Ca, Mg, K, Na, and SO4 as S. Profiles for Na are shown in Fig. 6A (profiles for all of these ions exhibited similar trends). Sodium concentrations were low (<100 g m–3) across the entire profile for Qf8, Qf7, and Qf6, indicating that percolating waters have removed these salts from the profile. In contrast, the profiles from areas of well-developed pavements (Qf5 and Qf3) exhibited elevated Na concentrations. At the base of Qf5 profile, soluble Na concentrations reached 500 g m–3, suggesting the maximum depth of percolation approached the bottom of the profile. On the Qf3 profile, peak values of >2500 g m–3 were observed at a depth of approximately 22.5 cm.
|
Figure 6B shows water content profiles 4 d after a slow-moving frontal storm left 3.9 cm of precipitation. The water content profiles suggest that differences in the water holding capacity are also strongly influenced by pedological development. For example, the observed trend of lower water contents with decreasing age (i.e., Qf8–Qf6) can be attributed to higher downward drainage and/or upward evapotranspiration. These results suggest that the water holding capacity in near-surface soils increased with increasing soil age and degree of soil development. The clearly increasing progression of near-surface water content and the total soil water storage within that profile (Fig. 7) correspond well with the increasing fines content shown in Fig. 2 and Table 1.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Pedological development of these surfaces impacts the water cycle in two important ways. First, lower hydraulic conductivities of this chronosequence dramatically limit infiltration and, by inference, will likely increase runoff. This will reduce the amount of water that percolates to plant roots and potentially increase the flux of water to ephemeral washes. Second, the higher clay and silt contents of older soils with desert pavements retain water longer in the upper, most bioavailable portion of the soil profile. This would provide plants more time to transpire the water, perhaps allowing the soil–plant system to be less susceptible to drought. These two characteristics of desert pavements may have profound impacts on ecosystem function in these arid systems. Correspondingly, disruption of these fragile surfaces and the underlying soil structure may have a dramatic impact on ecosystem function. Further studies will help to elucidate the complex interactions of soil development, hydraulic properties, and ecosystem responses that depend on water entry into soil profiles.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
imunek, J., R. Angulo-Jaramillo, M. Schaap, J.-P. Vandervaere, and M.Th. van Genuchten. 1998. Using an inverse method to estimate the hydraulic properties of crusted soils from tension disc infiltrometer data. Geoderma 86(1–2):61–81.[ISI]imunek, J., M. ejna, and M.Th. van Genuchten. 1999. HYDRUS-2D/MESHGEN-2D software for simulating water flow and solute transport in two-dimensional variably saturated media. Version 2.0. IGWMC-TPS-53C. International Ground Water Modeling Center, Colorado School of Mines, Golden, CO.This article has been cited by other articles:
|
|
 |
|
  R. C. Graham, D. R. Hirmas, Y. A. Wood, and C. Amrhein Large near-surface nitrate pools in soils capped by desert pavement in the Mojave Desert, California Geology, March 1, 2008; 36(3): 259 - 262. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Pelletier Cantor set model of eolian dust deposits on desert alluvial fan terraces Geology, May 1, 2007; 35(5): 439 - 442. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Zhu, M. H. Young, and M. Th. van Genuchten Upscaling Schemes and Relationships for the Gardner and van Genuchten Hydraulic Functions for Heterogeneous Soils Vadose Zone J., February 27, 2007; 6(1): 186 - 195. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. B. Ramos, M. C. Goncalves, J. C. Martins, M. Th. van Genuchten, and F. P. Pires Estimation of Soil Hydraulic Properties from Numerical Inversion of Tension Disk Infiltrometer Data Vadose Zone J., May 26, 2006; 5(2): 684 - 696. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 D. G. Meadows, M. H. Young, and E. V. McDonald A Laboratory Method for Determining the Unsaturated Hydraulic Properties of Soil Peds Soil Sci. Soc. Am. J., May 6, 2005; 69(3): 807 - 815. [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 |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
