WATER RESOURCES RESEARCH, VOL. 39, NO. 10, 1295, doi:10.1029/2002WR001324, 2003
Dewatering and the hydraulic properties of soft, sulfidic, coastal clay soils Ian White,1 David E. Smiles,2 Silvana Santomartino,1,3 Pam van Oploo,4 Bennett C. T. Macdonald,1 and T. David Waite5 Received 20 March 2002; revised 23 December 2002; accepted 2 May 2003; published 22 October 2003.
 Dewatering and consolidation of saturated swelling soils are governed by pressure-
dependent soil hydraulic properties. Existing measurement techniques are difficult and slow. We illustrate a simple, rapid desorption technique, developed for industrial slurries, to measure hydraulic properties of a gel-like sulfidic, estuarine soil (70% water content). Measured hydraulic conductivities, K(y), were very small, 1010 m/s, giving a representative capillary fringe thickness of 7 m and characteristic gravity drainage times around 40 years. Capillarity therefore dominates flow in these soils. Estimated times for dewatering this soil under surface loading with closely spaced, vertical wick drains, are 2 to 70 years, consistent with experience. A Netherlands marine clay soil, saturated with seawater, is unexpectedly wetter than the brackish estuarine soil here at the same matric potential, y. However, K(y) for both soils overlap, suggesting the engineering approximation, K(y) / jyj1, for marine-deposited clays. The functional dependencies of hydraulic properties surprisingly are not consistent with similar-media or double-layer INDEX TERMS: 1829 Hydrology: Groundwater hydrology; 1866 Hydrology: Soil moisture; theories. 1878 Hydrology: Water/energy interactions; 1894 Hydrology: Instruments and techniques; KEYWORDS: swelling soils, marine clays, hydraulic properties, dewatering, consolidation, soptivity Citation: White, I., D. E. Smiles, S. Santomartino, P. van Oploo, B. C. T. Macdonald, and T. D. Waite, Dewatering and the hydraulic properties of soft, sulfidic, coastal clay soils, Water Resour. Res., 39(10), 1295, doi:10.1029/2002WR001324, 2003.
1. Introduction  The development of coastal lowlands is increasing [White et al., 1997] with associated expansion of major engineering works, particular roadways and embankments. Many of these lowlands have unconsolidated, saturated, estuarine- and marine-origin sulfidic, Holocene sediments with shallow groundwater. Large areas [108 ha] of these unoxidized (unripe), potential acid sulfate soils [Pons, 1973; Kittrick et al., 1982] were deposited in low-energy marine or estuarine environments throughout the world following the last major sea level rise [Dent, 1986]. These saturated soils have low bulk densities and volumetric water contents around 80%. They are frequently gel-like and pose considerable engineering and environmental problems because they shrink significantly on drying, deform and flow under surface loads, and export highly acidic groundwaters following drainage and subsequent sulfide oxidation [Willett et al., 1993; Tin and Wilander, 1995; Wilson et al., 1999]. Oxidation of sulfides and ensuing soil acidification produce 1 Centre for Resource and Environmental Studies, Institute of Advanced Studies, Australian National University, Canberra, ACT, Australia. 2 Land and Water, Commonwealth Scientific and Industrial Research Organisation, Canberra, ACT, Australia. 3 Presently at Department of Earth Sciences, La Trobe University, Bundoora, Victoria, Australia. 4 Biological, Earth and Environmental Science, University of New South Wales, Sydney, New South Wales, Australia. 5 School of Civil and Environmental Engineering, University of New South Wales, Sydney, New South Wales, Australia.
Copyright 2003 by the American Geophysical Union. 0043-1397/03/2002WR001324
high concentrations of dissolved multivalent aluminum and iron in pore waters that react with the soil’s exchange complex. These exchange reactions, together with drying, progressively alter the physical properties of the gel-like soil in a process known as ‘‘ripening’’ where soil structure develops.  Some constructions on soft, potential acid sulfate soils have subsided or failed [Smiles, 1973], with significant consequences, and large costs. During failure (see Figure 1), the surrounding subsoils heave and flow, moving sulfidic, subsoils from anoxic to oxic zones. This can increase oxidation of sulfides and the rate of export of acidic oxidation products into neighboring coastal streams [White and Melville, 1993]. A common construction strategy on soft soils is to consolidate the soil to sufficient depth by pre-loading the soil surface [Smiles and Poulos, 1969] and to install drainage systems, such as closely spaced, vertical, wick drains, to enhance dewatering of the loaded soil. The rate of consolidation of soft soils due to overburden and applied loads depends on the soil’s hydraulic properties [Terzaghi, 1923; Horn and Baumgartl, 1999]. The hydraulic properties of saturated, swelling soils differ markedly from those of rigid soils [Giraldez and Sposito, 1985; Baumgartl and Horn, 1999]. They are dependent on overburden pressure and not readily measured by conventional techniques [Smiles, 2000]. Richards [1979, 1992] and Gra¨sle et al.  have explored the impact of overburden and applied loads on the hydraulic properties on stiff and unsaturated swelling and aggregated soils.  The soils of interest here fall between the wet, model clay systems with moisture ratios [ratio of volume of water to volume of solid] around 30 [Smiles and Harvey,
12 - 1
12 - 2
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
soil here with those of Kim et al. [1992b] for a marinederived soil, in equilibrium with seawater. Finally we examine the functional dependencies of the hydraulic properties and compare them with those expected from similar media and double-layer theory.
Figure 1. Failure of a major highway embankment constructed on unripe potential acid sulfate soils. The site is less than 1 km from the soil sampling site for this work, Mcleods Creek, Tweed River, NSW, eastern Australia. 1973; Smiles, 1976] and stiff clay soils with moisture ratios around 1 [Baumgartl and Horn, 1999]. There are few measurements in the intermediate range. Kim et al. [1992a, 1992b] measured the hydraulic conductivity and moisture characteristic of shrinking, unripe marine clay samples, with initial moisture ratio around 4, using a modification of an evaporation method [Wind, 1966]. They compared this technique with the one-step outflow method in which an instantaneous suction is applied to the bottom of the sample and the soil moisture diffusivity is determined from the rate of change of outflow with soil water content remaining in the sample [Passioura, 1976]. Their preferred modified evaporation technique required the insertion of tensiometers in the consolidating soil, and although automated, ran for almost six weeks. That duration would preclude its use in routine testing during development projects. As well, evaporation concentrates pore water electrolytes, which could alter soil hydraulic properties. The hydraulic properties of clay slurries depend strongly on soil solution concentration [Smiles et al., 1985]. While the modified evaporation technique may be relevant to the ripening of marine sediments, it may not be applicable to dewatering during consolidation, where no increase in soil water electrolyte occurs.  In very wet, swelling materials such as bentonite and red mud slurries [a clay-iron oxide by-product of the processing of bauxite] measurement of the early stage of pressure-driven outflow take only a few hours to complete [Smiles and Harvey, 1973; Smiles, 1976]. If this is also the case for saturated soft soils, then the technique is attractive for predicting consolidation and dewatering. The technique has been thoroughly tested for the uniqueness of the resultant measured hydraulic properties against transient and steady state techniques for clay slurries [Smiles, 1976, 1978].  In this work we explore the application of that technique to soft, potential acid sulfate field soils, deposited in low energy shallow lakes in a brackish, estuarine environment and with initial moisture ratios about one tenth that of the slurries. The dependence of hydraulic properties on initial soil moisture ratio is considered. We use the results to estimate the impact of capillarity on flow and dewatering of the soil and compare hydraulic properties of the estuarine
 The theory of one-dimensional flow in saturated swelling materials together with the fundamental differences between flow in swelling or consolidating and rigid systems are reviewed by Smiles [1986, 2000]. There are two critical differences between swelling and rigid soils. The first is in the energetics of soil water and the second is the coordinate system used to describe water flow.  In swelling and shrinking soils, the total potential of water (here all potential are expressed as work per unit weight of water, L) is the sum of the position in the gravitational field [L], the unloaded matric potential, y [L], and the overburden potential [L]. Overburden potential includes the load due to the weight of soil above the point of interest plus any imposed stresses [Croney and Coleman, 1961]. In swelling soils, water flow is described by Darcy’s law with flow relative to the soil particles, in a material or Lagrarian coordinate system, not relative to fixed spatial or Eulerian coordinates, as in rigid materials [Smiles and Rosenthal, 1968]. The water transport parameters that appear in Darcys law for swelling systems [Smiles and Rosenthal, 1968] are the material hydraulic conductivity, km(y) [LT1], and the material moisture diffusivity, Dm(J) [L2T1], with J the moisture ratio, [J = q/qS = q/(1 q) with q and qS the volume fractions of water and solid]. In geotechnical applications [Narasimham and Witherspoon, 1977], Dm(J), has long been known as the consolidation coefficient, a term that shall be used here, and is related to km and y by: Dm ðJÞ ¼ km ðJÞdy=dJ
In two phase systems, the material properties are related to the better-known, spatially defined hydraulic conductivity K(q) and moisture diffusivity D(q) by: K ðqÞ ¼ ð1 þ JÞkm ðJÞ
DðqÞ ¼ ð1 þ JÞ3 Dm ðJÞ
We use here the ‘‘early stages’’ of outflow during desorption of water from thin samples of saturated, consolidating soil under a range of imposed gas pressures to determine hydraulic properties [Smiles and Harvey, 1973]. The desorptivity, S [LT1=2 ], of the soil is determined from the rate of outflow of water from thin samples of soil, under a range of imposed gas pressures that are equivalent to unloaded matric potentials, y [L]. Thin samples with large area to height ratios are used so that friction between the sample and the wall of the apparatus is minimized [Horn and Baumgartl, 1999] and the effect of soil overburden is negligible relative to the imposed stress. By using thin samples we ensure that the matric potential at the end of outflow equals the imposed gas pressure. The equilibrium volumetric water content, q [m3 m3], or moisture ratio, at the end of each outflow measurement provides a point on
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
the moisture characteristic, J(y) of the soil. Each point on the moisture characteristic corresponds to a separate soil sample. This outflow technique has been thoroughly tested against transient and steady state techniques [Smiles, 1976, 1978]. Differentiation of the desorptivity data with respect to matric potential provides an estimate of the material hydraulic conductivity, km(y), while differentiation with respect to moisture ratio gives the consolidation coefficient, Dm(J) [L2T1].  Using the terminology of Philip , cumulative outflow, i [L3L2] in the early stages of outflow under an imposed gas pressure is: i ¼ S0 t 1=2
where S0 = S(y0, yn) is the sorptivity of the sample [Philip, 1957], defined as positive for adsorption and negative for desorption and yn is the initial unloaded matric potential of the sample and y0 is the surface matric potential or final soil matric potential in equilibrium with the imposed gas pressure. Sorptivity contains in an integral sense all the information about the changes in the pore system between the initial moisture ratio and the final equilibrium moisture ratio relevant to water flow [White and Sully, 1987].  Integral techniques for solving the flow equation subject to a step change of imposed pressure [Parlange, 1971; Philip, 1973; Philip and Knight, 1974] provide relations between sorptivity and soil water hydraulic properties [Smiles and Harvey, 1973; Parlange, 1975a; White and Perroux, 1989]: S02
Zy0 ¼ ðJ=bÞ
ZJ0 km ðyÞdy ¼ ðJ=bÞ
with J = J0 Jn, the difference in moisture ratio between final, J0, and initial, Jn, moisture ratios and b is: Zy0 b ¼ J
Zy0 km ðyÞdy=2
ðJ Jn Þkm ðyÞ dy F ðy; y0 ; yn Þ
In (6), F is the flux-concentration relation [Philip, 1973]. The factor b has relatively narrow limits and its value is determined by the ‘‘shape’’ of km(y)dy/dJ and by y0, yn [White and Sully, 1987]. This approach assumes that km and Dm are single valued functions of y and J, which is appropriate for saturated, consolidating soils. Differentiation of (5) with respect to the supply pressure y0 [White and Perroux, 1989] provides an estimate of km: km ðy0 Þ ¼ 2ðb=JÞS0 @S0 [email protected]
þ S02 @ ðb=JÞ[email protected]
Differentiation of (5) with respect to J0 [Smiles and Harvey, 1973] gives Dm: Dm ðJ0 Þ ¼ 2ðb=JÞS0 @S0 [email protected]
þ S02 @ ðb=JÞ[email protected]
The useful and quite general ‘‘optimal’’ case of Parlange [1975a], in which b = 1=2 and F = [(J Jn)/J]1/2, leads to: km ðy0 Þ ¼ ðS0 =JÞ½@S0 [email protected]
ðS0 =4JÞ@J0 [email protected]
Dm ðJ0 Þ ¼ ðS0 =JÞ½@S0 [email protected]
12 - 3
Table 1. Physical Properties at Sampling of the McLeods Creek Potential Acid Sulfate Soil at Depth 1.5 m Property
Clay content (wt %) Silt content (wt%) Fine sand content (wt %) Expandable lattice clay, Smectite (% of clay fraction) Kaolinite (% of clay fraction) Illite (% of clay fraction) Pyrite (wt %) Bulk density, r (t/m3) Particle density, rs (t/m3) Volumetric water content, q (m3/m3) Moisture ratio in field, J (m3/m3)
48 51 1 60 30 10 3.5 0.6 2.545 0.78 3.3
Parlange’s approximations (9) and (10) are sufficiently accurate for practical purposes in both swelling and rigid systems [Smiles, 1976; White and Perroux, 1989] and we use them here. To determine km and Dm from outflow experiments using (9) and (10), S0 and J0 are measured for different soil samples over a range of supply potentials, y0 by imposing a range of gas pressures p [p = y0] to the samples and determining S0 from the early stages of outflow through (4).
3. Methods 3.1. Soil Samples  Bulk samples of potential acid sulfate soil with a field moisture ratio of 3.3 were taken at depth of 1.5 m below the soil surface in a sugarcane farm at a site on McLeods Creek, Tweed River, northern NSW, in eastern Australia. The Holocene soils (Thionic Fluvisols) at this depth have the field texture of a light clay. The road site in Figure 1 is within 1 km of the soil sample site. Oxidized, actual acid sulfate soils at the site occur to a depth of about 0.8 m. Below this depth, unconsolidated, Holocene potential acid soils at the site extend to a depth of about 9 m, beneath which they are underlain by highly consolidated Pleistocene clays.  About 30 kg monoliths of the structureless gel soil were extracted from the profile and sealed in double plastic bags to exclude air. These were transported under refrigeration and stored at 4C to minimize oxidation. The physical properties of the soil are given in Table 1 and the composition of the clay exchange complex is given in Table 2, which reveals that magnesium dominates the exchange complex. Table 3 shows the concentration of major ions in the soil solution. The ratio of Na to Cl in the pore water solution [Na/Cl = 0.7 with concentrations in mg/L] shows the marine origin of these waters (seawater Na/Cl = 0.55) but the Cl concentration is only about 4% of that of seawater, reflecting perhaps the estuarine, back-swamp origin of these soils. 3.2. Sorptivity and the Moisture Characteristic  Approximately 10 mL of the potential acid sulfate soil were transferred into a pressure filtration cell [Smiles et al., 1985] which was filled to a depth of about 9 mm. The cell had a cross-sectional area of 1.14 103 m2 and was fitted with a 0.45 mm membrane at its outflow surface to retain clay. Air was excluded from the cell and a selected constant gas pressure, p, in excess of atmospheric pressure was imposed on the sample at time t = 0. This pressure was
12 - 4
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
Table 2. Composition of the Exchange Complex of the McLeods Creek Potential Acid Sulfate Soil at Depth 1.5 ma Exchangeable Cations
Abundance on Exchange Complex, %
Mg2+ Ca2+ Na2+ K+ Al3+ Fe3+ Mn2+
55 21.5 13 6.5 3.0 0.5 0.5
Total exchange capacity 59 mM(+)/kg.
maintained and the outflow of water from the cell was weighed (to 105 kg) as a function of time. During outflow the soil sample consolidates and shrinks one dimensionally under saturated conditions. All measurements were at 20C.  Sorptivity was determined by plotting the measured outflow as a function of the square root of time and determining the slope of the initial period of desorption as in (1). Measurement was continued until no further soil water was expelled from the sample. The cell was then disassembled and the water content of the soil in equilibrium with the gas supply pressure, p, was determined by oven drying at 105C. The moisture ratio in equilibrium with the supply pressure was found from J0 = qg rS/rW, with qg the gravimetric moisture content of the sample and rW the density of the soil water solution. We emphasize that the soil remains saturated throughout these measurements and that each measurement requires a fresh soil sample.  It is important to examine the uniqueness of the hydraulic properties produced by this technique. One way to do this is to use samples with different initial moisture contents [Smiles, 1976]. If the hydraulic properties are unique they should be independent of initial moisture content. In order to investigate the effect of initial moisture content, a diluted soil sample (5 soil:1 water) was prepared by adding approximately 40 ml of distilled water to 200 ml of soil. The sample was mixed and allowed to equilibrate. The original sample had an initial moisture ratio of 3.13 and the diluted sample had Jn = 3.36. The hydraulic properties of the diluted sample were then measured in the same way as those of the original soil sample described above.
Figure 2. Typical results for the cumulative outflow from an estuarine clay soil sample showing the early stage square root time behavior expected from (1) and the approach to equilibrium. Here the imposed gas pressure was 5.44 m H2O. stages of outflow on the square root of time predicted in (1) and typical of the results here, is evident in Figure 2 as is the relatively rapid approach to equilibrium.  Figure 3 shows the measured sorptivities as a function of the imposed gas pressure, p [= y0]. As
4. Results 4.1. Outflow and Sorptivity  Figure 2 shows the measured outflow, for an imposed gas pressure of 5.44 m H2O. The linear dependence of early Table 3. Concentration of the Major Ions in the Soil Solution, McLeods Creek Potential Acid Sulfate Soila Ion +
Na K+ Mg2+ Ca2+ Mn2+ Fe2+ Al3+ Cl SO42 a
Electrical conductivity 4.1 dS/m; pH 6.7.
Concentration, mg/L 547 41 260 129 19 4 0.9 785 1480
Figure 3. Dependence of sorptivity during dewatering on the imposed gas pressure [ p = y0], for both the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols). The lines are the fit of the data to (11).
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
12 - 5
Table 4. Functional Dependencies of the Hydraulic Properties S(y0), J(y0), km(y), K(y), Dm(J) and D(q) of the Estuarine Soil Together With Values of the Square of the Correlation Coefficient r Compared With Those of a Marine Soila Equation ln(S0) = G + H ln jy0j ln(S0) = G + H ln jy0j J = A B ln jy0j J = A B ln jy0j ln(km) = M N ln jyj ln(km) = M N ln jyj ln(km) = M N ln jyj ln(k) = m m ln jyj ln(k) = m m ln jyj ln(K) = m m ln jyj ln(Dm) = P QJ ln(Dm) = P QJ
this work, this work, this work, Kim et al. this work, Kim et al. all data this work, Kim et al. all data this work, Kim et al.
Source of Data
original sample diluted sample combined samples [1992b], combined datab combined samples [1992b], combined datab
11.04 10.76 2.34 3.24 22.43 21.76 21.72 21.21 20.32 20.42 16.46 21.54
0.46 0.40 0.30 0.25 0.39 1.09 0.82 0.50 1.2 1.0 2.0 0.36
0.98 0.99 0.90 – 0.86 – 0.87 0.98 – 0.90 – –
combined samples [1992b], combined datab combined samples [1992b], combined datab
See Kim et al. [1992b]. Data from Kim et al. [1992b] estimated from the ‘‘standard’’ results in their Figures 7 and 10.
expected, the sorptivities of the diluted, initially wetter samples are larger than those of the original, drier soil samples.  Measured sorptivities of mud and clay slurries determined during desorption have been found to follow: ln½S ðy0 Þ ¼ G þ H lnjy0 j
with G and H constants [Smiles, 1976]. Figure 3 also shows the fitted relations (11) and values for the constants, H and G are listed in Table 4 together with the square of the correlation coefficient, r. The logarithmic relationship (11) appears to describe the data adequately. The measured exponents, H, in Table 4, for the two samples are 0.4 and 0.46. Red mud slurries have H close to 0.4 [Smiles, 1976] while for bentonite slurries H 0.2 [Smiles and Harvey, 1973]. 4.2. Moisture Characteristic  The moisture characteristics, J(y), for the original and diluted samples are plotted in Figure 4. Within the scatter of data, there is no significant difference between the original and diluted samples. This is consistent with the measurements on bentonite and red mud slurries [Smiles and Harvey, 1973; Smiles, 1976]. Also plotted in Figure 4 is the moisture characteristic measured by Kim et al. [1992b] for an unripe, marine clay soil saturated with seawater. We defer discussion of this comparison to section 5.5 below.  The theory of weakly interacting planar double layers [Collis-George and Bozeman, 1970; Smiles et al., 1985; Sposito, 1989], suggests that J(y) in clay systems should follow: J ¼ A B lnjyj
r2. The semilogarithmic relation (12) gives a reasonable fit to the data. 4.3. Hydraulic Conductivity  The calculated material hydraulic conductivities, km(y), for the two samples are shown in Figure 5. Again, both the original and diluted samples are identical within the scatter of results. Material hydraulic conductivities of slurries [Smiles and Harvey, 1973; Smiles, 1975, 1995], appear to follow km / jyjN or: ln½km ðyÞ ¼ M N lnjyj
with M and N empirical constants. Figure 5 shows the combined data fits the logarithmic relation (13) quite well.
with B a parameter that depends on surface charge density on the clay, soil water electrolyte composition and concentration and temperature. For clay slurries, however, the magnitude of the dependence of B on electrolyte concentration is not predicted by simple double-layer theory [Smiles et al., 1985] and A and B must be considered as empirical parameters. The fit of the combined J(y) data to (12) is also shown in Figure 4 and values of the constants A and B in (12) are listed in Table 4, together with the value of
Figure 4. Moisture characteristics, J(y), for the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols). Also shown are results estimated from Kim et al. [1992b] for a marine-origin clay soil. The lines show the fit of the data to (12).
12 - 6
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
Figure 5. Dependence of the material hydraulic conductivity, km(y), on unloaded matric potential for both the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols) of the potential acid sulfate soil. Also shown is the fit of the data to (13). Values for the parameters M and N in (14) and r2 are also listed in Table 4 where we see a value of N 0.4 for the estuarine soil. Red mud slurries have N 0.3 [Smiles, 1976] but N 1.1 appears more appropriate for bentonite slurries [Smiles, 1995].  Values of the more familiar hydraulic conductivity K(y), evaluated using (9), are plotted in Figure 6. The combined data from both sets of samples are described very well by: ln½K ðyÞ ¼ m n lnjyj
Figure 6. Dependence of the hydraulic conductivity, K(y), on unloaded matric potential for both the original estuarine clay soil (solid symbols) and the diluted sample (open symbols) of the potential acid sulfate soil. The line is the data fitted to (14). using (3), are plotted in Figures 7 and 8. Again, results for the original and diluted samples in both figures are identical within the scatter of data. The Dm(J) data in Figure 7 show the semi-logarithmic dependence on moisture ratio expected
Equation (14) is plotted in Figure 6 and values for the constants, m and n are in Table 4. We note that the magnitude of K(y) in Figure 6 is quite small, typically of order 0.01 mm/day, close to the range where materials are considered impermeable, despite the high water content of these soils. 4.4. Consolidation Coefficient  The consolidation coefficient [Narasimham and Witherspoon, 1977] follows from the definition (8) and the empirical relations (12) and (13): ln½Dm ðJÞ ¼ M ln B
N 1 ð A JÞ B
which can be written as: ln½Dm ðJÞ ¼ P QJ
Consolidation coefficients, Dm(J), estimated from (15) and the better-known moisture diffusivity D(q), calculated
Figure 7. Dependence of the consolidation coefficient on the moisture ratio for both the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols) of the potential acid sulfate soil. Also shown is the fit of the data to (15).
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
12 - 7
dilution we have used has little impact on soil solution concentration. 5.2. Effect of Gravity and Capillarity  Swelling reduces the impact of gravity and extends the ‘‘early stages’’, capillarity-dominated period of transient flow [Smiles, 2000]. Philip  identified a time, tgrav when the influence of gravity compared with capillarity, starts to be significant during infiltration. White and Sully  suggested that a consistent measure of tgrav is: tgrav ¼ bfS0 =½ K ðy0 Þ K ðyn Þ g2
Figure 8. Soil moisture diffusivity as a function of volumetric soil water content for both the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols) of the potential acid sulfate soil. The line indicates the mean value.
from (15). Values for the parameters P and Q in (15) are listed in Table 4. The value of Q = 2.0 measured here for the estuarine soil is similar to that for red mud slurries, Q = 1.5 [Smiles, 1976] but is much larger than that estimated for bentonite slurries, Q = 0.06 [Smiles and Harvey, 1973]. Again diluted and original samples are indistinguishable.  No significant dependence of D on q is evident within the scatter of the data in Figure 8 over the small range of volumetric moisture content sampled here. This means that D(q) can be conveniently represented by a constant mean value of 3.8 108 m2/s. Bentonite slurries appear to have a similar, weak or nonmonotonic dependence of D on water content [Smiles and Harvey, 1973].
Figure 9 shows tgrav calculated using (16) with b = 1/2 and the sorptivities and hydraulic conductivities measured in this work. From our results here we estimate K(yn) = 1.1 109 m/s for the undiluted sample and K(yn) = 1.2 109 m/s for the diluted sample at their respective initial moisture contents. Values of tgrav range from about 12 years to 67 years, and are approximately linearly dependent on the matric potential, with, tgrav = 15.6 + 2.65 y0. In bentonite slurries, tgrav can be as large as 400 years [Smiles, 1986].  The macroscopic capillary length lc of a soil is a measure of the representative capillary fringe thickness above a water table and, consequently, the characteristic pore size of the soil [Myers and van Bavel, 1963; Bouwer, 1964; White and Sully, 1987]. It can be estimated from [White and Sully, 1987]: lc ¼ bS02 =fq½ K ðy0 Þ K ðyn Þ g
where q = q(y0) q(yn) and b is given by (9). Values for lc calculated from (17) with b = 1=2, and are plotted in
5. Discussion 5.1. Uniqueness of Hydraulic Properties  A concern with the desorption technique used here is that it may yield nonunique hydraulic properties [Parlange, 1975b]. If this were so then we expect that hydraulic properties should depend on the initial moisture content of the sample. Previous measurements using the desorption technique on clay slurries have shown that the moisture characteristic, J(y), the consolidation coefficient, Dm(J) the material hydraulic conductivity km(y) and actual hydraulic conductivity, K(y) are unique and independent of the initial moisture content of the sample [Smiles, 1976, 1978]. In a limited way, and within the scatter of the experimental data, we have also found that here for the more complex soft, potential acid sulfate soil. The hydraulic properties for the 20% diluted sample in Figures 4 to 8 are identical to those of the original sample. Exchange with the clay complex in the soil here means that the 20%
Figure 9. Values of the gravity time, estimated from (16), for the estuarine clay soil for both the original estuarine clay soil sample (solid symbols) and the diluted sample (open symbols).
12 - 8
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
Figure 10. Values of the representative capillary fringe thickness, lc, estimated from (17) for the combined samples of estuarine clay soil.
Figure 10 for the combined samples. The estimated representative capillary fringe thicknesses show considerable scatter but span a range from 3 to 12 m. These large values suggest that the influence of the water table in these soils, particularly on evaporation, will be propagated over considerable depths in the profile.  The ratio L/lc, where L is a characteristic length of a source or sink of water, determines whether the dimensionality of flow needs to be considered [Bouwer, 1964]. When L/lc 1, flows can be treated as one-dimensional. The large values of lc found here indicate that the dimensionality of the flow from surface sources or sinks will be important in the early stages of flow in these soils. Together, the results for lc and tgrav suggest that most flows in these materials can be treated as capillarity-driven (gravity free) but that the dimensionality of flow will need to be considered for source or sink characteristic lengths of order 10 m or less. 5.3. Dewatering Times for Vertically Drained Soils  The actual magnitudes of the material hydraulic conductivity, the hydraulic conductivity and the consolidation coefficient measured here for the estuarine soil are very small. They indicate that dewatering these soils will be a slow process. To accelerate dewatering of soft coastal sediments, closely spaced, vertical wick drains are inserted 10 to 20 m into the soil and surface loads are gradually applied. Because gravity effects are negligible, the time, t, to dewater these soils can be estimated from [Smiles, 1973]: J Jn S0 pﬃﬃﬃﬃﬃﬃ ¼ erfc pt J 2JLD
 Figure 11 shows the estimated time for dewatering as a function of the drain half spacing, for a total overburden load of 5 m (water potentials here are expressed in work per unit weight of water so that the unit of potential and hence overburden is meters of water). For a typical value of LD of 1.0 m (18) predicts that over 2 years are required to dewater these sediments using vertical wick drains with a 5 m applied surface load. A major highway close to the soil sample site without surface drainage has continued to settle by 4 to 6 m over the past 40 years. Because dewatering is a desorption process, outflow and consolidation will be rapid at first but will decline as the square root of time (as in (4)). The early, rapid dewatering may lead to unfounded expectations of quick consolidation.  There is the prospect that the long dewatering times estimated here could be reduced by one to two orders of magnitude. The injection or in situ generation of higher solution concentrations of salt or cations of higher valence such as calcium, iron or aluminum in wick drains prior to dewatering may increase the consolidation coefficient markedly, close to outflow surfaces. Rosenqvist  reported dramatic ‘‘stiffening’’ of quick clays in the field after the injection of 2% brine solution. 5.4. Relevance of Laboratory Measurements to Field Situations  The question of the relevance of these laboratory measurements of the hydraulic properties of small samples to field situations is important. The measurements here have revealed very small hydraulic conductivities which give rise to low dewatering times even with relatively closely spaced vertical wick drains. These back swamp gel soils were mainly deposited in shallow, low energy, barrier lakes where sedimentation of fine particles occurred slowly. They are remarkably uniform. Recent highway construction has
Here J = J0 Jn and LD is the half-spacing between vertical drains. The usual goal is to drain the sediment to a relative water content of (J0 Jn)/J = 0.5.
Figure 11. Predicted dependence of the dewatering time on the half-spacing between vertical wick drains from (18) for a total overburden load of 5 m.
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
occurred along nearly 400 km of the eastern Australia coastal floodplains covering nearly 50 km of unconsolidated sulfidic back swamp areas with depths of up to 30 m. Wick drains, with half-spacing of less than 0.7 m, were required to provide adequate dewatering over periods of 1 to 2 years (D. Cramer, personal communication, February 2002). This is entirely consistent with predictions made here from the laboratory measured hydraulic properties. 5.5. Comparison With Previous Measurements of Unripe, Marine-Origin Soils  The hydraulic properties of the soil studied here falls between the properties of wet clay slurries with moisture ratios around 30 and stiff clay soils with moisture ratios less than 1 [Baumgartl and Horn, 1999]. There are few comprehensive studies on soil systems with moisture ratios between 1 and 10 with which to compare our results. Kim et al. [1992b] measured the hydraulic properties of two samples of an unripe, marine clay soil with initial moisture ratios around 4. This soil is saturated with seawater and deposited under a similar low energy environment to the soil examined here. There are some discrepancies between the results from different measurement techniques used in their work. We will compare our measurement here to their ‘‘standard technique.’’  Netherlands soil, had a bulk density of 490 kg/m3, with clay and silt contents of approximately 46% and 53% respectively [Kim et al., 1992a]. The clay minerals were mainly kaolinite and illite with a small amount of montmorillonite, reflecting the marine origin of the soil. The soil solution was in equilibrium with seawater. Our eastern Australian clay soil had similar amounts of clay and silt (Table 1) but has smectite as the dominant clay followed by kaolinite with a minor amount of illite, showing its estuarine origin. It has a higher bulk density, lower moisture ratio at sampling and a smaller soil solution electrolyte concentration (1/25 seawater, based on chloride concentrations) than the Netherlands soil. The Australian soil was deposited around 6000 BP. It is presumed that the Netherlands soil is much younger than this. 5.5.1. Moisture Characteristics  We have combined the moisture characteristic data of the two samples from Figure 7 of Kim et al. [1992b] for jyj > 0.1 m H2O and have fitted the data to (12). Moisture ratios were calculated using their plotted values of q. The resulting moisture characteristic is compared with that found here in Figure 4. Values of the A and B parameters, found by fitting their data to (12), are also listed in Table 4. We do not list a value for r2 since it would merely be a measure of our ability to extract information from their Figure 7.  The moisture ratios of Kim et al.’s marine clay in equilibrium with seawater (Figure 4) are significantly larger, at the same matric potential, than those of the estuarine clay studied here. The ambient soil solution concentration of our estuarine soil was approximately 1/25 of that of the marine clay. Double-layer theory for weakly interacting, dilute clay systems indicates that the unloaded matric potential at fixed moisture ratio should be a function of the ambient solution electrolyte concentration and temperature [see, e.g., Sposito, 1989]. If these two soils were similar materials, both theory and experience with bentonite slurries [Smiles et al., 1985; Smiles, 1995] suggest that the moisture ratios at the higher
12 - 9
salt concentration in the Netherlands’ marine sediment should be smaller than those of our estuarine soil at the same matric potential. The value of the slope parameter B = 0.25 in (12) listed in Table 4 for the marine sample of Kim et al. is more than 80% of that for our estuarine sample.  The soils from the Netherlands and eastern Australia differ in two important aspects, their clay chemistry and the age of the soils. Kaolinite and illite are the dominant clays in the Netherlands soil compared with smectites in the Australian soil. At the same ambient soil solution concentrations the Netherlands soil would be expected to swell less than the Australian soil. The evaporation method used by Kim et al. [1992b] increased the soil solution concentration above that of the initial seawater saturating the samples. Under these higher electrolyte concentrations we should also expect the Netherlands soil to be drier than the Australian soil at the same matric potential. The soils, however, differ in age. It is quite possible that the Australian soil has experienced some consolidation under its 1.5 m of soil overburden in the 6,000 years since its deposition, whereas the younger Netherlands soil (from an unspecified but possible shallow depth) has not. We also note here that both soils contain over 50% silt. The role of this silt fraction in determining the moisture characteristic may be important. 5.5.2. Hydraulic Conductivity Moisture Potential Relations  Smiles et al.  found large differences between the moisture characteristics and the km(J) relations of bentonite slurries when pore water electrolyte concentrations were varied. When, however, material conductivities were expressed as a function of matric potential, km(y), the measurements for different electrolyte concentrations unexpectedly collapsed to a single curve [Smiles, 1995] and the empirical equation (13) provided a single, apparently unique, relation for Km(y) for bentonite slurries over a wide range of pore water electrolyte concentrations. The theoretical principles underpinning this collapse have yet to be discovered. We examine here if such a collapse is also possible for both the Netherlands and Australian soils.  Kim et al. [1992b, Figure 10] report measured K(y) for two samples of the marine-origin sediment. Figure 12 compares K(y) for their marine-origin soil with results from this work. The two soils, whose solution concentrations differ 25-fold, were measured using different techniques mainly over different y ranges. The larger K(y) for the Netherlands marine samples were determined at generally smaller values of jyj. Where the jyj measurement ranges overlap, the hydraulic conductivities of both soils also overlap. The low values of hydraulic conductivity found for estuarine soil here are therefore consistent with those found for the Netherlands marine-origin soils.  We have fitted the combined data of Kim et al. [1992b] to (14). The values found for the parameters m and n are listed in Table 4. The slope of n = 1.2 for the marine-origin soil is similar to that for bentonite slurries but differs from that found here for the estuarine clay, n = 0.5 [Table 4]. If both sets of K(y) data for the two soils are combined and fitted to (14), a plausible ‘‘engineering’’ approximation is suggested for marine-origin clays as shown in Figure 12 with K(y) 1.4 109jyj1.  We have also calculated the material conductivity, km(y) for the combined samples of Kim et al. [1992b].
12 - 10
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
Figure 12. Comparison of the hydraulic conductivity for the estuarine clay soil of this work (circles) with that of the marine-origin clay soil (crosses) of Kim et al. [1992b]. Also shown is the fit of the combined data for both soils to (14).
Table 4 shows the corresponding parameters M and N when these results are fitted to (13). The results for km are similar to K(y). Where matric potentials are similar, the km for the Netherlands marine-origin soil and those of the Australian estuarine soil overlap. Again, combining both sets of data and fitting them to (13) results in an ‘‘engineering’’ approximation with parameters listed in Table 4. 5.5.3. Consolidation Coefficient  Finally, we can use the values of A, B, M and N for the marine soil of Kim et al. in Table 4 to calculate the consolidation coefficient, Dm(J), from their data using (15). Values for the parameters P and Q for their soil are compared in Table 4 with those found in this work. We note that the slope parameter, Q, for the marine-origin soil is of different sign to that of the soil investigated here and also to those for red mud and bentonite slurries. Results for bentonite [Smiles and Harvey, 1973], however, suggest that the functional dependence of Dm(J) may be nonmonotonic. 5.6. Functional Dependence of Hydraulic Properties for Saturated Clays  Philip  assumed in idealized swelling materials at equilibrium there was a balance between sedimentation and soil particle Brownian motion. For dilute, saturated, swelling soil pastes he derived the moisture characteristic Jþ1¼
kT 1 rW gv y
Here k is Boltzmann’s constant, T the absolute temperature, rW the specific gravity of the soil water solution, g the acceleration due to gravity, and v the mean particle volume. Equation (19) suggests that the value of B in (12) should be
close to 1.0. Here, we found B 0.3 for the estuarine soil and B 0.25 for the marine-origin soil of Kim et al. [1992b]. For red mud slurries B 0.2 [Smiles, 1976], while for bentonite slurries the value is much larger, with B ranging from approximately 3.2 to 5.6 dependent on soil solution electrolyte concentration [Smiles et al., 1985]. Equation (19) embodies the notions that a swelling suspension can expand without limit on the continued addition of water and soil particles are free to move under Brownian motion. Both appear unrealistic for field soils.  The saturated, swelling clay soils examined here are relatively simple, two-phase materials. It is reasonable to suspect that their hydraulic properties during dewatering, when normal volume change occurs [Croney and Coleman, 1961], may behave in a self-similar way during dewatering. If so, similar media concepts [Miller and Miller, 1956; Philip,1969] suggest that y / 1/l and S / l1/2, with l a characteristic pore size of the material. We would then expect S / y1/2. It must be recognized, however, that the dependence of S on y enters as the upper and lower limits of integration in (8) and (9). If we concentrate on the supply potential, y0, for desorption, where larger imposed jy0j mean that additional finer pores drain and more water is expressed, we might expect that S0 / y01/2 or more generally, S0 / jy0jH, as in (11). Here we have found H in the range 0.4 to 0.46, slightly larger than that for red mud slurries of 0.36.  Similar media theory also dictates that K / l2, so for a self-similar, dewatering clay soil we would expect K / 1/y2. If this is so, then (9) and (19) suggest that, for dewatering, km / 1/y. The estuarine soil here and slurries [Smiles and Harvey, 1973; Smiles, 1975, 1995] however, appear to follow more generally km / jyjN, as in (13). We have found N 0.4 for the estuarine soil, but the Netherlands soil had N 1.1, close to that for bentonite slurries, and to similar media expectations. Red mud slurries, however, have N 0.3, similar to the estuarine soil studied here.  Smiles  attempted to fit the Kozeny-Carman equation [Carman, 1939] to the material hydraulic conductivity of bentonite slurries. This equation can be expressed as: km ðJÞ / J3 =ðJ þ 1Þ2
If (19) is valid, then (20) also suggests that km is approximately proportional to 1/y. The functional dependence of km(J) predicted by (20), however, gives a poor fit to our measurements for the estuarine soil here and also for bentonite slurries.  The theory of weakly interacting planar double layers [Collis-George and Bozeman, 1970; Smiles et al., 1985; Sposito, 1989], suggests that J(y) in clay slurries should follow J / jyjB, as in (12). If we set A = aB in (12), then, at a fixed temperature, double-layer theory predicts for a single electrolyte species of ambient solution concentration, C: a ¼ ar þ lnðC=Cr Þ
and B ¼ Br
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Cr =C
where Br ¼ rS d=2
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
Here ar and Br are constants at reference concentration Cr at the same temperature, is the specific surface area of the solid particles and d the Debye length. The values of ar and Br also depend on the valence of the cations in the soil solution. The dependence of a on solution concentration in (21) is less than that of B. Both a and B should decrease with increasing soil solution concentration. We therefore expect that A in (12) should decrease by about the same relative amount as B with increasing solution concentration.  If it is assumed that the estuarine soil here and the marine soil of Kim et al. [1992b] behave similarly, doublelayer theory qualitatively predicts that the Netherlands soil, with its larger, ambient pore water salt concentrations should be drier and denser than our estuarine-origin soil. Rosenqvist  believed that gel nature of marinedeposited quick-clays only persists in situations where saline formation pore waters have been exchanged with fresh water. The clay fraction morphology in these soils is in itself complex [Sposito, 1989], their geologic history and the fine silt fraction may also play important roles.  If the two soils had behaved similarly then equation (21) would predict that the value of B for the marine samples of Kim et al. should be approximately 1/5 that of the estuarine samples here. Instead, B for the marine samples of Kim et al. is about 80% of the estuarine soils. In bentonite slurries, B decreased by a factor of only 1.7 for a 37-fold increase in pore water electrolyte concentration, much lower than expected from (18) [Smiles et al., 1985]. This posses a dilemma. It is well known that the model, parallel-plate, clay systems in the laboratory follow closely double-layer theory well [Israelachivili, 1985; Sposito, 1989; Kjellander et al., 1990; Quirk, 1994]. Perhaps in slurries and materials with ‘‘card-house’’’’ clay structures, the combination of edge and surface charges gives rise to different behavior.  The expanding lattice clays in of these soils, under suitable conditions, form expanded, gel-like structures held together by long-range attractive forces. In these expanded structures, clay platelets are arranged in a complex mixture of face-to-face, edge-to-face and edge-to-edge configurations [Van Olphen, 1964; Sposito, 1989; Quirk, 1994; de Krester et al., 1998]. Secondary minimum induced coagulation results in long distance associations with hydrated counter-ions taking up interlayer space [Sposito, 1989]. The stability of these structures depends critically on the electrolyte content of the pore water solution [Keren et al., 1988], on the confining pressure and on shearing forces. Increase in concentration of counter ions past a critical concentration enables primary minimum aggregation, where the strength of the face-face interactions increase and the complete gel structure collapses. Dramatic, ‘‘quick-clay’’like changes in the rheological and physical properties of these soils can occur when their pore water solution compositions are altered or when stresses are applied, causing the collapse of the gel-like structures [Rosenqvist, 1953, 1966; Lessard and Mitchell, 1985; Mitchell, 1986; Luckham and Rossi, 1999].  Within the scatter of results here, no abrupt changes in soil properties were observed with imposed stress. If, however, the rearrangement of clay platelets under increasing pressure does not follow a ‘‘self-similar’’ path, with a progression of simple, geometrically scaled pore dimen-
12 - 11
sions during dewatering [Miller and Miller, 1956], K(y) may not follow the 1/y2 behavior expected from scaling theory. It is clear that the nature of the K(y) relation for saturated swelling soil systems and the impact of soil solution electrolyte concentration on require investigation.
6. Conclusions  We have demonstrated that the outflow technique, developed for determining the hydraulic properties of industrial slurries with large moisture ratios during the early stages of dewatering [Smiles and Harvey, 1973; Smiles, 1976], is directly applicable to unripe, saturated marineor estuarine-origin soils that occur in coastal lowlands. Measurements of outflow using samples that are thin relative to typical overburden loads and imposed stresses, and have large area to height ratios, avoids complications caused by thick samples. It also allows dewatering to be completed in a few hours, even when the hydraulic conductivity is of order 1010 m/s, as here. The technique is rapid, as required in engineering applications, simple, and does not require the insertion of tensiometers or cause an increase in the pore water electrolyte concentration as in the slow, evaporation method of Kim et al. [1992b]. In addition, it has been thoroughly tested for slurries against other transient and steady state techniques. These rapid measurements yield all the necessary hydraulic properties, the moisture characteristic, material hydraulic conductivity and the consolidation coefficient, required for predicting flow and dewatering of consolidating materials using the macroscopic Darcy description of water flow.  Our measurements comparing 20% diluted samples with samples taken at field moisture content showed that the moisture characteristic, J (y), consolidation coefficient, Dm(J), the soil moisture diffusivity D(q), the material hydraulic conductivity km(y) and actual hydraulic conductivity, K(y) are, within experimental error, unique and independent of the initial moisture content of the sample.  The water transport properties of the gel soil measured here are small, close to being considered impermeable, despite their 70% water content. Associated with these very small transport properties, are large representative capillary fringe thicknesses, lc, between 3 and 12 m. Gravity times tgrav [Philip,1969] are corresponding long, between 12 and 67 years. We conclude that capillary forces will therefore dominate water flow in these soils and the dimensionality of flow from surface sources or to sinks will be important. As a consequence of this, predicted dewatering times for the soil, under surface loads and with vertical, wick drains are large, typically in the range 2 to 70 years, which is consistent with experience. Times required for consolidation during highway construction traversing long sections of the coastal floodplains in eastern Australia are consistent with the estimates based on the laboratory measured hydraulic properties here.  Previously used empirical or quasi-theoretical relationships proposed for the functional dependencies of hydraulic properties adequately describe the hydraulic properties measured here, within the scatter of the data. However the functional dependencies of the hydraulic properties during dewatering do not follow those expected from self-similar media. The simple relationship for
12 - 12
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS
hydraulic conductivity K(y) / yn describes our measurements very well. Hydraulic conductivities, K(y), determined here have an approximate 1/y1/2 dependence and not the 1/y2 dependence expected from ‘‘self-similar’’ media arguments [Miller and Miller, 1956; Philip, 1969]. The question of the functional dependence of K(y) for soft sediments and the impact of electrolyte concentrations on that dependence, we believe, warrant further attention.  Comparison of the measurements of hydraulic properties here with previous results for an unripe, marine-origin soil from the Netherlands is intriguing. The moisture content of the Netherlands soils is wetter than the estuarine soil from eastern Australia at the same matric potential. Simple double-layer theory would suggest that the Netherlands soil, with its 25 times larger pore water electrolyte concentration should be drier than the eastern Australian soil. Electrolyte effects on J(y) of bentonite slurries also do not appear predictable from double-layer theory [Smiles et al., 1985]. This is a puzzle since model, parallel-plate model clay systems obey theory well [Israelachivili, 1985]. We have speculated that this behavior, as well as the failure of slurries and the estuarine soil here to be self-similar during dewatering, may be due perhaps to the collapse of ‘‘card-house’’ clay structures, or to the different geologic histories and mineralogy of the soil, or to interactions with the silt fraction.  Where y ranges overlapped for the Netherlands and Australian soils, the K(y) for both soils also overlapped. This suggests that the low values found here are consistent with those measured for the similar clay and silt content soil from the Netherlands. Although the slopes of the K(y) relations differed for the two soils, the engineering approximation K(y) 1.4 109 jyj1 provides a first approximation for both soils, despite the 25-fold difference in pore water electrolyte concentrations. The collapse of K(y) to a single relation independent of soil solution concentration found by Smiles  for bentonite slurries and the overlap here of K(y) for the saturated, low-energy marine and brackish water-deposited clays, with markedly different pore water electrolyte concentrations, deserve further study. In addition, the potential to accelerate dewatering of these soils through injection or in situ generation of multivalent electrolytes requires additional research.
N slope parameter in material hydraulic conductivity relation (13). n slope parameter in material hydraulic conductivity relation (14). p imposed gas pressure [ML1T2]. P constant in consolidation coefficient relation (15). Q slope parameter in consolidation coefficient relation (15). r correlation coefficient. S sorptvity [LT1/2]. t time [T]. tgrav gravity time defined in (16) [T]. T absolute temperature [K]. v mean particle volume [L3]. a = A/B in (21). d Debye length [L]. total potential per unit weight of water [L]. g surface tension of soil solution [MT2]. J moisture ratio = volume of water/volume of solid [L3L3]. J = J0 Jn. k Boltzmann constant. lc representative capillary fringe thickness defined in (17) [L]. q volumetric soil water content [L3L3]. qS volume fraction of solid [L3L3]. r bulk specific gravity of soil [ML3]. rS specific gravity of soil solids [ML3]. rW specific gravity of soil water [ML3]. specific surface area of solid particles [L2M1]. y unloaded matric potential [L]. Subscripts 0 value at supply potential y0. m material value. n value at initial conditions. r value at reference conditions.
A B b C D Dm F g G H i K km L LD M
Baumgartl, T., and R. Horn, Influence of mechanical and hydraulic stresses on hydraulic properties of soil, in Characterisation and Measurement of the Hydraulic Properties of Unsaturated Porous Media, edited by M. van Genuchten, F. J. Leij, and L. Wu, pp. 449 – 457, Univ. of Calif., Riverside, 1999. Bouwer, H., Unsaturated flow in ground-water hydraulics, J. Hydraul. Div. Am. Soc. Civil Eng., 90(HY5), 121 – 144, 1964. Carman, P. C., Permeability of saturated sands, soils and clays, J. Agric. Sci., 29, 262 – 273, 1939. Collis-George, N., and J. M. Bozeman, A double layer theory for mixed ion systems as applied to the moisture content of clays under restraint, Aust. J. Soil Res., 8, 239 – 258, 1970. Croney, D., and J. D. Coleman, Pore pressure and suction in soil, in Pore Pressure and Suction in Soil, pp. 31 – 37, Butterworths, London, 1961. de Krester, R., P. J. Scales, and V. D. Boger, Surface chemistry-rheology interrelationships in clay suspensions, Colloids Surf. A, 137, 307 – 318, 1998. Dent, D. L., Acid Sulphate Soils: A Baseline for Research and Development, Publ. 39, Int. Inst. for Land Reclam. and Impr., Wageningen, Netherlands, 1986. Giraldez, J. V., and G. Sposito, Infiltration in swelling soil, Water Resour. Res., 21, 33 – 44, 1985.
constant in moisture characteristic (12). slope parameter in moisture characteristic (12). parameter for sorptivity relation (3). electrolyte concentration in soil water [ML3]. soil moisture diffusivity [L2T1]. consolidation coefficient [L2T1]. flux-concentration relation. gravitational acceleration [LT2]. constant in sorptivity relation (11). slope parameter in sorptivity relation (11). cumulative desorption [L3L2]. hydraulic conductivity [LT1]. material hydraulic conductivity [LT1]. characteristic source or sink dimension [L]. half spacing between vertical drains [L]. constant in material hydraulic conductivity relation (13). m constant in hydraulic conductivity relation (14).
 Acknowledgments. We thank Mike Melville of the University of NSW for helpful discussions and cane farmers Robert and Allan Quirk and Robert Hawken for support throughout this work. Support from the Water Research Foundation of Australia, the NSW ASSPRO and the Australian Research Council, under ARC Large Grant A39917105, ARC Linkage Grant LP0219426 and Discovery Grant DP0345145 and are gratefully acknowledged.
WHITE ET AL.: HYDRAULIC PROPERTIES OF COASTAL CLAY SOILS Gra¨sle, W., B. G. Richards, T. Baumgartl, and R. Horn, Interaction between soil mechanical properties of structured soils and hydraulic processes: Theoretical fundamentals of a model, in Unsaturated Soils, vol. II, edited by E. E. Alonso and P. Delage, pp. 719 – 725, A. A. Balkema, Brookfield, Vt., 1996. Horn, R., and T. Baumgartl, Dynamic properties of Soil, in Handbook of Soil Science, edited by M. E. Sumner, pp. A19 – A51, CRC Press, Boca Raton, Fla., 1999. Israelachivili, J. N., Intermolecular and Surface Forces, Academic, San Diego, Calif., 1985. Keren, R., I. Shainberg, and E. Klein, Settling and flocculation value of sodium-montmorillonite particles in aqueous media, Soil Sci. Soc Am. J., 52, 76 – 80, 1988. Kim, D. J., H. Vereecken, J. Feyen, D. Boels, and J. J. B. Bronswijk, On the characterisation of properties of an unripe marine clay soil, I. Shrinkage processes of an unripe marine clay soil in relation to physical ripening, Soil Sci., 153, 471 – 481, 1992a. Kim, D. J., H. Vereecken, J. Feyen, D. Boels, and J. J. B. Bronswijk, On the characterisation of properties of an unripe marine clay soil, II. A method on the determination of hydraulic properties, Soil Sci., 154, 59 – 71, 1992b. Kittrick, J. A., D. S. Fanning, and L. R. Hosner (Eds.), Acid Sulfate Weathering, SSSA Spec. Publ. 10, Soil Sci. Soc. of Am., Madison, Wis., 1982. Kjellander, R., S. Marcelja, R. M. Pashley, and J. P. Quirk, A theoretical and experimental study of forces between charged mica surfaces in aqueous CaCl2 solutions, J. Chem. Phys., 92, 4399 – 4407, 1990. Lessard, G., and J. K. Mitchell, The causes of aging in quick clays, Can. Geotech. J., 22, 335 – 346, 1985. Luckham, P., and S. Rossi, The colloidal and rheological properties of bentonite suspensions, Adv. Colloid Interface Sci., 8, 43 – 92, 1999. Miller, E. E., and R. D. Miller, Physical theory for capillary flow phenomena, J. Appl. Phys., 27, 324 – 332, 1956. Mitchell, J. K., Practical problems from surprising soil behaviour, J. Geotech. Eng., 112, 259 – 289, 1986. Myers, L. E., and C. H. van Bavel, Measurement and evaluation of water table elevations, paper presented at 5th Congress, Int. Comm. on Irrig. and Drainage, Tokyo, May 1963. Narasimham, T. N., and P. A. Witherspoon, Numerical model for saturatedunsaturated flow in deformable porous media: 1. Theory, Water Resour. Res., 13, 657 – 664, 1977. Parlange, J.-Y., Theory of water movement in soils. II One dimensional infiltration, Soil Sci., 111, 170 – 174, 1971. Parlange, J.-Y., Determination of soil water diffusivity by sorptivity measurements, Soil Sci. Am. Proc., 39, 1011 – 1012, 1975a. Parlange, J.-Y., A note on the moisture diffusivity of saturated swelling systems from desorption experiments, Soil Sci., 120, 157 – 158, 1975b. Passioura, J. B., Determining soil water diffusivity from one-step outflow experiments, Aust. J. Soil Res., 15, 1 – 8, 1976. Philip, J. R., The theory of infiltration: 4. Sorptivity and algebraic infiltration equations, Soil Sci., 84, 257 – 264, 1957. Philip, J. R., Theory of infiltration, Adv. Hydrosci., 5, 215 – 296, 1969. Philip, J. R., Hydrostatics in swelling soils and soil suspensions: Unification of concepts, Soil Sci., 109, 294 – 298, 1970. Philip, J. R., On solving the unsaturated flow equation: 1. The flux concentration relation, Soil Sci., 116, 328 – 335, 1973. Philip, J. R., and J. H. Knight, On solving the unsaturated flow equation: 3. New quasianalytic technique, Soil Sci., 117, 1 – 13, 1974. Pons, L. J., Outline of the genesis, characteristics, classification and improvement of acid sulphate soils, in Proceedings of the International Symposium on Acid Sulphate Soils 13 – 29 Aug. 1972 Wageningen, Publ. 18, vol. 1, edited by H. Dost, pp. 3 – 27, Int. Inst. Land Reclam. and Impr., Wageningen, Netherlands, 1973. Quirk, J. P., Interparticle forces: A basis for the interpretation of soil physical behaviour, Adv. Agron., 53, 121 – 183, 1994. Richards, B. G., The method of analysis of the effects of volume change in unsaturated expansive clays on engineering structures, Aust. Geomech. J., G9, 29 – 47, 1979. Richards, B. G., Modelling interactive load-deformation and flow processes in soils, including unsaturated and swelling soils, in paper presented at 6th Australia-New Zealand Conference on Geomechanics, Inst. of Eng. Aust., Christchurch, New Zealand, 1992. Rosenqvist, T., Consideration of the sensitivity of Norwegian quick-clays, Geotechnique, 3, 195 – 200, 1953.
12 - 13
Rosenqvist, T., Considerations on the sensitivity of Norwegian quick clays—A Review, Eng. Geol., 1, 445 – 450, 1966. Smiles, D. E., An examination of settlement data for an embankment on a light wet clay, Aust. Road Res., 5, 55 – 59, 1973. Smiles, D. E., Some aspects of liquid movement in phosphate slime, Separation Sci., 10, 767 – 776, 1975. Smiles, D. E., On the validity of the theory of flow in saturated swelling material, Aust. J. Soil Res., 14, 389 – 395, 1976. Smiles, D. E., Transient and steady-flow experiments testing theory of water flow in saturated bentonite, Soil Sci. Soc. Am J., 42, 11 – 14, 1978. Smiles, D. E., Principles of constant pressure filtration, in Encyclopedia of Fluid Mechanics, edited by N. P. Cheremisinoff, pp. 791 – 824, Gulf Publ., Houston, Tex., 1986. Smiles, D. E., Liquid flow in swelling soils, Soil Sci. Soc. Am. J., 59, 313 – 318, 1995. Smiles, D. E., Hydrology of swelling soils: A review, Aust. J. Soil Res., 38, 501 – 521, 2000. Smiles, D. E., and A. G. Harvey, Measurement of the moisture diffusivity of swelling systems, Soil Sci., 116, 391 – 399, 1973. Smiles, D. E., and H. G. Poulos, The one-dimensional consolidation of columns of soil of finite length, Aust. J. Soil Res., 7, 285 – 291, 1969. Smiles, D. E., and M. J. Rosenthal, The movement of water in swelling materials, Aust J. Soil Res., 6, 237 – 248, 1968. Smiles, D. E., C. J. Barnes, and W. R. Gardner, Water relations of saturated bentonite: Some effects of temperature and solution salt concentration, Soil Sci Soc. Am. J., 49, 66 – 69, 1985. Sposito, G., The Chemistry of Soils, Oxford Univ. Press, New York, 1989. Terzaghi, K., Die Berechnung der Durchla¨ssigkeitsziffer des Tones aus den Verlauf der Hydrodynamischen Spannungserscheinungen, Sitzungber. Akad. Wiss. Wein Math. Naturewiss. Kl., Abt. 2A, 132, (3 – 4), 125 – 138, 1923. Tin, N. T., and A. Wilander, Chemical conditions in acidic waters in the plain of reeds, Vietnam, Water Resour., 29, 1401 – 1408, 1995. Van Olphen, H., Internal mutual flocculation in clay suspensions, J. Colloids Interface Sci., 126, 278 – 291, 1964. White, I., and M. D. Melville, Treatment and containment of potential acid sulphate soils: Formation, properties and management of potential acid sulphate soils, Tech. Rep. T53, CSIRO Cent. for Environ. Mech., Canberra, Australia, 1993. White, I., and K. M. Perroux, Estimation of the unsaturated hydraulic conductivity from field sorptivity measurements, Soil Sci. Am. J., 53, 324 – 329, 1989. White, I., and M. J. Sully, Macroscopic and microscopic length scales from field infiltration, Water Resour. Res., 23, 1514 – 1522, 1987. White, I., M. D. Melville, B. P. Wilson, and J. Sammut, Reducing acid discharge from estuarine wetlands in eastern Australia, Wetlands Ecol. Manage., 5, 55 – 72, 1997. Willett, I. R., M. D. Melville, and I. White, Acid drainwaters from potential acid sulphate soils and their impact on estuarine ecosystems, in Selected Papers of the Ho Chi Minh City Symposium on Acid Sulphate Soils, Publ. 53, edited by D. L. Dent and M. E. F. van Mensvoort, pp. 419 – 425, Int. Inst. of Land Reclam., Wageningen, Netherlands, 1993. Wilson, B. P., I. White, and M. D. Melville, Floodplain hydrology, acid discharge and water quality associated with a drained acid sulfate soil, Mar. Freshwater Res., 50, 149 – 157, 1999. Wind, G. P., Capillary conductivity data estimated by a simple method, in Water in the Unsaturated Zone: Proceedings of the IASH/UNESCO Symposium, vol. II, edited by R. E. Ritjema and H. Wassink, pp. 127 – 141, United Nations Educ., Sci., and Cultural Organ., Paris, 1966.
B. C. T. Macdonald and I. White, Centre for Resource and Environmental Studies, Institute of Advance Studies, Australian National University, Canberra, ACT 0200, Australia. ([email protected]
) S. Santomartino, Department of Earth Sciences, La Trobe University, Bundoora, Victoria 3086, Australia. D. E. Smiles, Land and Water, Commonwealth Scientific and Industrial Research Organisation, Canberra, ACT 2601, Australia. P. van Oploo, Biological, Earth and Environmental Science, University of New South Wales, Sydney, New South Wales 2052, Australia. T. D. Waite, School of Civil and Environmental Engineering, University of New South Wales, Sydney, New South Wales 2052, Australia.