The hydrology of Vertosols used for cotton production: I. Hydraulic, structural and fundamental soil properties.
Water-use efficiency of cotton production and minimising the impact of the cotton industry on the environment have emerged as issues of great importance in recent times. Resolution of these issues is strongly reliant on a sound knowledge of the hydraulic properties of the cotton-growing soil types. The specific soil type used most extensively in the cotton industry is the heavy clay soil of the order Vertosol (Isbell 1996). Although some research has been carried out to quantify the different components of the hydrological cycle of such soils (Montgomery et al. 1999), this work has been limited to a few sites. Measuring the hydraulic properties of these soils is difficult; they typically shrink and swell, which means that the bulk density is a function of the water content (Berndt and Coughlan 1976), and their plastic nature leads to overburden pressures on the subsoil by the weight of overlying layers (Smiles 2000). Additionally, large cracks form at the surface through which large amounts of water can quickly infiltrate to the subsoil (Collis-George 1977). As reviewed by Smiles (2000), extensive theoretical work on water flow in cracking clay soils has been carried out in Australia (e.g. Smiles and Rosenthal 1968; Philip 1969a, 1969b; Philip and Smiles 1969; Collis-George and Bridge 1973), but considering the areal extent of these soils and the intensive irrigated crop production upon them, there are few hydraulic data available. This is most likely due to the expense and difficulty of obtaining such data from these non-rigid soils.
The development of pedotransfer functions (PTFs) may offer a means of obtaining accurate soil hydraulic information, which is otherwise laborious and expensive to collect. Although PTFs, which predict difficult-to-measure properties (hydraulic properties) from fundamental soil properties (soil texture, bulk density), have been extensively developed in Europe and the United States (Rawls et al. 1982; Wosten and van Genuchten 1988), there remains a dearth of such functions for Vertosols under Australian conditions. Many field measurements have remained unpublished, or have been published only in internal organisational reports and have therefore not had extensive exposure (e.g. the dataset used by Shaw and Thorburn 1985). The main published datasets (Prebble 1970; Forrest et al. 1985; Geeves et al. 1995) used to develop Australian PTFs (Cresswell and Paydar 1996; Minasny and McBratney 2000a) contain only a limited number of Vertosol samples. No dataset specifically includes all necessary hydraulic properties in combination with a large number of fundamental soil properties for a large number of Vertosols.
Because of the complex shrink--swell nature of Vertosols, a full description of the hydraulic properties important for irrigation management should include: saturated and near-saturated hydraulic conductivity, the water retention curve, and the shrink--swell relationships. Each of these measurements is complex in these soils, and standard methods cannot always be applied (Coughlan et al. 1991). Double-ring infiltrometers have been used to measure the final infiltration rate over several days (Talsma and van der Lelij 1976; McIntyre et al. 1982), but the effects of swelling complicate calculation of the hydraulic conductivity (Gerard-Marchant et al. 1997). Well permeameters have been used to measure the subsoil conductivity of Vertosols (Bird et al. 1996), but smearing can easily occur in wet soils. Disc permeameters offer a rapid surface infiltration method, which can estimate several unsaturated infiltration rates in a matter of hours (Clothier 2000) and allow for the calculation of an unsaturated hydraulic conductivity curve (Clothier 2000). Measurement of the water retention curve is complicated due to the lack of applied overburden pressure on samples from the subsoil during the analysis (Tariq and Durnford 1997). Placing weights on top of the cores to mimic the overburden pressure has been tried (e.g. Bronswijk 1990), but decision-making on the correct weight is arbitrary (Klute 1986). Shrink--swell relationships have been measured using several different techniques: by rewetting and reshaping loose soil (Tariq and Durnford 1993; McKenzie et al. 1994), but such techniques do not allow development of a full shrink--swell relationship; indirect measurements on clods or cores based on the displacement of a fluid using Archimedes' Principle to calculate the volume (Brasher et al. 1966); and measuring the dimensions of the sample directly, such as by using lasers (Braudeau et al. 1999) or by using in-sire depth gauges (Coquet et al. 1998). Costs and technical difficulties associated with the last 2 types of analyses currently limit widespread use.
Soil structure strongly influences the flow of water and the movement of air through soil (Hamblin 1985; Connolly 1998), but its measurement remains problematic. In Vertosols, structural form is highly developed due to shrink--swell behaviour (McGarry 1996) and the tendency to self-mulch (Grant and Blackmore 1991). During the last decade, improvement in image analysis techniques has allowed a more thorough quantification of soil structural form through the development of measurable pore and solid attributes, such as the pore intercept length, solid intercept length, and pore/solid surface area (McBratney et al. 1992; Horgan 1998). Such quantifiable structural properties, which together indicate the size, shape, and connectedness of soil pores, should theoretically be strongly correlated with observed soil hydraulic behaviour and other soil physical properties; however, few studies have been conducted to verify this (Bouma et al. 1979; Decks et al. 1999; Vervoort and Cattle 2002). Correlating soil physical and structural properties will improve the ease of estimating various physical properties, as these properties can be changeable and notoriously difficult to quantify in non-rigid soils such as Vertosols. The development of quantifiable relationships would empower the use of recently emerged techniques such as image analysis for soil physical assessment. In addition, the quantification of the structural form could help explain some of the variability in hydraulic conductivity and give more confidence in the capability to predict hydraulic conductivities from pore distribution models (e.g. Burdine 1953; Mualem 1976).
This paper aims to describe the development of a database of hydraulic, structural and fundamental soil properties of a range of Vertosols in south-eastern Australia, and, particularly, to ascertain the interdependence of these properties across all sampled sites. A companion paper will describe the formulation of a series of PTFs for Vertosol hydraulic properties using this database.
Sites and sampling
Topsoil (0-0.2 m) layers were sampled from 18 sites located in 5 different cotton-growing areas of southeastern Australia (Table 1). At 2 of these sites, soil pits were excavated and soil was sampled from 3 and 2 subsurface layers, respectively, to give a total of 23 Vertosol layer (or horizon) samples. The subsurface layers were sampled from depths of 0.4, 0.85, and 1.3 m at the first pit site, and from 0.85 and 2.1 m at the second pit site. There was no deliberate selection of sites or subsurface layers, although sampling across layer boundaries was avoided. With the exception of one site in the Gwydir valley and one site in the Macquarie valley, all sampled sites were more than 25 km from another sampled site. The sampling strategy allowed the incorporation of a range of Vertosols across a large area, as the focus of this research was on describing the relationship between the hydraulic and physico-chemical properties across all sites and layers rather than on describing the between-site variability.
Most of the sites were the same as those described in Vervoort and Cattle (2002). The sampling sites were all located in irrigated cotton fields at least 100 m from a fence line. or head- or tail-drain. Plant rows were selected for most of the measurements to avoid complications due to slaking and wheel traffic in furrows. Multiple soil cores of 2 different sizes were sampled from all 23 layers: 5 cores (75 mm OD, 80 mm height) to determine soil water characteristics, taken along the row; and 2 cores (150 mm OD, 200 mm height) for impregnation with resin and subsequent image analysis. Additionally, 5 clods (1-3 x [10.sup.5] [mm.sup.3]) were sampled from depths of 100, 200, and 400 mm at each site for the measurement of shrink--swell characteristics. Soil samples for the determination of gravimetric wetness were taken with the clod samples. Where the sample site was a pit, clods were sampled from each subsurface layer under investigation. Bulk soil (2 kg) was collected from all layers to analyse fundamental soil properties. Care was taken to perform all sampling within a 1-1.5-m length of the plant row or bench in a pit, with the core, clod and bulk soil samples, and the hydraulic measurements, being as closely spaced as possible.
Prior to the destructive soil sampling in the same 1-1.5-m length of the plant row, 2 disc permeameter measurements at 5 different supply potentials (-150, -100, 70, -50, and -20 mm) were carried out. From these measurements, the hydraulic conductivities of these layers were calculated (Clothier 2000) as described in Vervoort and Cattle (2002). Because Vertosols are often 'self-mulching', forming a distinct crumb layer at the surface (Grant and Blackmore 1991), the surface physical characteristics may differ markedly from those of the underlying soil (Pillai-McGarry and Collis-George 1990). Consequently, for the 18 topsoil layers sampled, one of the disc permeameter measurements was taken at the surface, including the crumb layer, while the second measurement was carried out after removing the crumb layer. Again, the focus was not on measuring the variability in hydraulic conductivities between the sites, but on how hydraulic conductivity across all sites was related to soil properties. Measurements at a -20-mm supply potential were also carried out in the furrow at 8 of the surface soil sites. Soil samples for the determination of gravimetric wetness were taken before and after the measurements to allow calculation of hydraulic conductivity (Clothier 2000). If a sample consisted of a subsurface layer, no crumb layer or furrow measurements were taken, meaning that only 1 infiltration measurement was carried out. If no distinct crumb-layer existed at a site, 'below crumb-layer' measurements of hydraulic conductivity were carried out at a depth of 25 mm.
In the field, intact soil clod samples were dipped twice in a solution of Saran resin dissolved in acetone in a 1:7 ratio (Brasher et al. 1966), and stored in an air-tight container after 0.5 h of drying. The resin forms a thin plastic coating on the clods. Weight and volume (by displacement in water) were determined immediately on arrival in the laboratory to represent field bulk density. The resin coating on the clods was peeled back slightly and the clods were allowed to resaturate on a sand table at a supply tension of -0.5 kPa. The clods were again dipped twice in the resin solution and left to air-dry in the laboratory. Weight and volume (by displacement in water) were determined at different times during this drying process. Clods were finally oven-dried at 105 [degrees] C, and weight and volume were once more determined. This allowed calculation of the oven-dry bulk density, and subsequently, by back calculation, the determination of bulk densities and gravimetric water contents (wetness) at each of the drying times. The different bulk density and wetness combinations were plotted and the 3-straight-line model (TSLM) (McGarry and Malafant 1987) was fitted to the data using standard least-squares techniques. Other models have been suggested (Braudeau et al. 1999), but these models all include more parameters, or the parameters have no physical meaning. The TSLM is defined as follows:
(1) [gamma](U) = [alpha] + rU if U < [U.sub.a] [gamma](U) = [alpha] + r[U.sub.a] + n(U - [U.sub.a]) if [U.sub.a] [less than or equal to] U < [U.sub.b] [gamma](U) = [alpha] + r[U.sub.a] + n([U.sub.b] - U) + s(U - [U.sub.b]) if U > [U.sub.b]
where [gamma] is the specific volume (or bulk [density.sup.-1, [cm.sup.3]/g) and U is the wetness (g/g). The model recognises 3 shrinkage phases. The first phase, that of structural shrinkage, runs from saturation to the point [U.sub.b] (g/g) on the curve. During this phase the loss in volume is less than the loss of water, or [differential]U/[differential][gamma] < 1. The second phase, that of normal shrinkage, runs between [U.sub.b] and [U.sub.a] (g/g), also known as the air entry point. Theoretically, the loss of water in this phase should be accompanied by an equal loss in volume, or [differential]U/[differential][gamma] = 1, but this only occurs in structureless soil (Chart 1982; Coughlan et al. 1991). The third phase, that of residual shrinkage, runs from [U.sub.a] to [alpha], the specific volume ([cm.sup.3]/g) for the oven-dry soil. During this phase the loss of water is again larger than the loss of volume, or [differential]U/[differential][gamma] < 1. The slopes of the different segments (g/[cm.sup.3]), or [differential]U/[differential][gamma], are defined by: s, the slope of the structure1 shrinkage phase; n, the slope of the normal phase; and r, the slope of the residual phase. In practice, not all 3 shrinkage phases might occur and not all have to behave according to the theoretical curve.
Soil water characteristic
Soil water characteristic curves were measured in the laboratory using individual pressure cells on the 5 samples (75 mm OD, 80 mm height) from each site. The cores were equilibrated at -0.5 kPa (without a load) using a hanging water column and placed in individual cells, and a pressure of 1 kPa was applied. This means that the resulting soil water characteristics do not reflect field overburden pressures. Cells were weighed regularly and the pressure was increased when the cell weights changed <0.1 g over 24 h. The cells were stepped through 1, 2, 4, 8, 16, 32, and 64 kPa. Further water retention at 100 and 250 kPa on a subsample from the cells, and at 1500 kPa using resaturated <2-mm sieved soil, was determined using a pressure apparatus (Klute 1986). A van Genuchtan-type equation (van Genuchten 1980) was fitted to the data, using standard least-squares techniques. Since the TSLM for the shrinkage curve also contains an [alpha] and n parameter, we have chosen to name the parameters of the soil water characteristic [[alpha].sub.WC] and [n.sub.WC] to differentiate between the 2 different models. The equation was defined in terms of the moisture ratio [??], which is defined as the volume of water to the volume of solids and can be calculated as the wetness, U, multiplied by the density of solids, [[rho].sub.s]. Note that the moisture content, [theta], equals 9 * [[rho].sub.s]/[[rho].sub.b], where [[rho].sub.b] is the bulk density. The fitted equation was:
(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where [[??].sub.r], [[??].sub.s] are the residual and saturated moisture ratios, respectively; [[alpha].sub.WC] is a scaling factor; [n.sub.WC] is a factor controlling the shape of the curve; and m is a constant taken as m = 1 - 1/[n.sub.wc]. The parameters [[??].sub.s] and [[??].sub.r] are generally also treated as fitting parameters (van Genuchten 1980), and in this case [[??].sub.r] was found to have very little influence on the fitting process and was subsequently set to 0. The fitted moisture ratios can be calculated back to gravimetric water contents and conversion to volumetric water contents can be performed using the shrink--swell relationship (Eqn 1).
Goodness-of-fit for both the soil water characteristic curve and the shrink swell relationships was defined in terms of the root mean square error (RMSE), which is defined as:
(3) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where Nd is the number of datapoints, [[??].sub.i] and [[gamma].sub.i] are the fitted moisture ratio and specific volume, and [[??].sub.i] and [[gamma].sub.i] are the observed moisture ratio and specific volume.
Plant available water contents (PAWC) were calculated as the difference between the product of the bulk density and the wetness at 10 kPa and at 1500 kPa using the shrink--swell relationship and the soil moisture characteristic. This represents the PAWC for a fully swollen, and freely swelling profile.
Soil structural analysis
The methods for sampling, preparation of soil cores, and analysis of soil structural parameters have been described extensively in Vervoort and Cattle (2002). In summary, the method of McBratney et al. (1992) was used to impregnate large cores (150 mm diameter and 200 mm height) with resin containing a fluorescent dye. The top 100 mm of each core was subsequently horizontally sectioned (in 10- or 20-mm increments) and photographed under ultraviolet light, and the images were converted to black and white binary images. As described in Vervoort and Cattle (2002), these binary images were, using various pixel-counting routines, analysed using the program SOLICON v2.1 (Cattle et al. 2001) for the following soil structural parameters: macroporosity, surface area (SA), mean pore and solid intercept length (MPIL and MSIL), and pore and solid star length (PSL and SSL). Based on the resolution of the camera, 1 pixel represents 0.15 mm in this analysis (Vervoort and Cattle 2002).
Fundamental soil properties
Samples from all 23 layers were oven-dried at 105[degrees]C to enable the calculation of field wetness at the time of sampling. Bulk soil from each sampled soil layer was air-dried, crushed, and passed through a 2-mm sieve. Particle size analysis of the <2 mm fraction was performed by pipette and sieving (Kilmer and Alexander 1949), to determine the fractions; sand (2 mm-50 [micro]m), very fine sand (50-20 [micro]m), silt (20-2 [micro]m), coarse clay (2-0.2 [micro]m), and fine clay (<0.2 [micro]m). Total carbon was determined by total combustion using a CHN elemental analyser (Leco Inc., St Joseph, MI, USA). The pH was measured in a ratio of 1:2 soil to 0.01 M Ca[Cl.sub.2], while electrical conductivity (EC) was determined for a 1:5 soil to water solution. Exchangeable Ca, Mg, K, and Na contents were determined by extraction with 1 M N[H.sub.4]Cl at pH 8.0 and subsequent atomic adsorption spectroscopy. Effective cation exchange capacity (ECEC) was calculated as the sum of the exchangeable Ca, Mg, K and Na contents. The exchangeable sodium percentage (ESP) was calculated from the ratio of exchangeable Na content and the ECEC. Inorganic carbon in the form of CaC[O.sub.3] was determined using dissolution by HCl (Loeppert and Suarez 1996). Extractable sulfate levels were determined by extraction using 0.01 M Ca[Cl.sub.2] and subsequent precipitation using Ba and analysis of the remaining Ba using atomic adsorption spectroscopy (Hue and Adams 1979). Iron oxides were extracted using dithionite-citrate-bicarbonate (Mehra and Jackson 1960) and subsequent atomic adsorption spectroscopy. Phosphorus contents were determined colourimetrically (Rayment and Higginson 1992) after extraction using 0.5 M NaHC[O.sub.3].
The clay mineral suite of the soil at both sites was determined by X-ray diffraction analysis of sedimented aggregates. A standard method of sample preparation was employed (Whittig and Allardice 1986) and all analyses were conducted using CuK[alpha] radiation, spanning the angle range of 2-15[degrees] 2[theta]. The major clay mineral types present in the bulk clay traction (<2 [micro]m), and in the fine clay fraction (<0.2 [micro]m), were determined.
Data were analysed using a 1-way analysis of variance (ANOVA) and separation of means was carried out using the multiple comparison Tukey-Kramer test with both P 0.05 and 0.1. To define groups for the statistical analyses, a classification analysis (k-means classification) in JMP (SAS Institute Inc., Cary, NC, USA) was carried out based on all available data, and another classification was carried out using just the data for surface layers (above 400 mm depth). The hydraulic conductivity and soil structural analysis data were log-transformed before statistical analysis, due to the lognormal distribution of this data.
The classification analysis of all available data indicated that the main class separation was between layers below 400 mm depth and samples above (surface layers), with the first principal component strongly related to depth of sampling. Because of this strong relationship with depth, we present all surface and subsurface data separately in the results section. The classification of the 18 surface layers grouped these samples primarily on clay content (first principal component), while the second principal component was mostly correlated with the hydraulic conductivity ([K.sub.s]). As a consequence, the Namoi and Gwydir valley layers tended to be grouped together and the Macquarie valley layers tended to be grouped together. The Darling Downs and Lachlan valley layers tended to be spread across 3 classes. The third class (characterised by the highest sand contents) contained only one sample each from the Lachlan and Macquarie valleys. Therefore, with a moderate separation of surface layer samples on a locational basis, we present all surface layer results separated into valleys in the results section, both for the benefit of simplicity and to enable a comparison of soil layers with different parent materials.
Overall, the measured hydraulic conductivities were quite large for fine-textured soil (Table 2). The measurements made immediately below the crumb layer yielded larger mean values of saturated hydraulic conductivity ([K.sub.s]) at the Gwydir and Namoi valley sites than the Macquarie valley sites. In those valleys where hydraulic conductivities were measured at the soil surface (including the crumb layer), the mean [K.sub.s] values of the crumb layer were significantly larger than the below-crumb values at only the Namoi valley and Darling Downs sites, while the Macquarie valley surface [K.sub.s] values were significantly smaller than those of the other valleys. The differences between the surface and below-crumb [K.sub.s] values in the different valleys might reflect the degree of self-mulching of these soils, with a greater degree of self-mulching expected to provide a highly connected and conductive pore phase, increasing the surface [K.sub.s]. Subsoil values of [K.sub.s] in the Namoi and Gwydir valley pits were significantly smaller than the [K.sub.s] values in the surface and below-crumb layers of those pits. The mean hydraulic conductivities measured in the furrow at -20 mm supply potential were not significantly different to the [K.sub.-20] values measured for the below-crumb layer on the ridge (Table 2). This indicates that although the furrow is used by traffic and has regular water flow over it, there appears to be, at least in this dataset, no detrimental effect to the infiltration rates under unsaturated conditions.
Shrink-swell properties and soil water characteristic
The shrink-swell analysis highlighted some technical difficulties with using Saran resin to coat large Vertosol clods, as has been pointed out earlier (McGarry and Daniells 1987). Clod samples from the 100-mm depth tended to be quite porous and the Saran resin solution did not coat the large pore openings but infiltrated into the clod. It is questionable whether this led to sufficient coating of the inside of the large pores. It meant that the bulk density for these surface clods might have been overestimated, because the measurements did not include the volumes of the large pores. It also meant that the shrink swell curves tended to be quite noisy, as is probably reflected by the greater RMSE for the samples above 400 mm depth compared to those below 400 mm depth (Table 3). This problem might be overcome by using a more dilute Saran solution than the recommended 1:7 Saran : acetone ratio and by applying more layers of the Saran solution.
There were no significant differences in field wetnesses between the different valleys as measured from the clod samples, nor was there any difference in field wetnesses if all data (subsamples taken with the infiltration measurements and subsamples taken with the clod samples) were taken into account. There were also no significant differences in field wetnesses at the time of sampling between surface and subsoil samples.
Field bulk density is defined as the bulk density of the clod measured immediately on return to the laboratory and before re-saturation. The Darling Downs samples had significantly greater field bulk densities than samples from the Gwydir, Namoi and Lachlan valleys (Table 3). The field bulk densities of the subsoil samples from the Gwydir and Namoi valleys were also significantly greater than those of samples from the overlying topsoils, possibly indicating the effect of overburden pressures.
The overall mean value of n, which represents the degree of shrinkage in the normal zone, was smaller than the theoretical value 1 for nearly all samples (Table 3), and there were no significant differences in n for samples from different valleys. However, a clear distinction in shrinkage phases did not always exist for these field samples (e.g. Fig. 1), making the estimation of n problematic. This could be due to a lack of data points in critical regions, and seems to suggest that a more continuous function, such as a logistic fit, might be more appropriate (Braudeau et al. 1999; Crescimanno and Provenzano 1999). A disadvantage of these types of models is a lack of physical interpretation of the parameters (McGarry and Malafant 1987) or an increase in the number of parameters, requiring a larger number of observations (Braudeau et al. 1999). Most of the Vertosol samples used here did exhibit an S-shaped curve, however, confirming the existence of a region of greater shrinkage, bounded by an area of less shrinkage at the wet end and at the dry end.
None of the other shrinkage parameters estimated ([U.sub.A], [U.sub.B], [alpha], r, s) was significantly different between samples of the different valleys, but the parameter [alpha] was found to be significantly smaller below 400 mm depth for the Gwydir and Namoi valley samples (Table 3). This parameter represents the oven-dry specific volume, with smaller values indicating greater bulk densities and a less favourable structure (Coughlan et al. 1991). Differences in field bulk density between samples were therefore not mirrored in the oven-dry bulk densities, indicating small differences in soil swelling capacity. In general, there was little difference between the shrink swell properties of the soil samples from the various valleys. Soil water characteristic parameters also differed little between the sampled valleys, but there was a distinct difference between the surface and lower layers in the profile (Table 4). The subsoil samples had significantly greater [[alpha].sub.wc] values than the topsoil, indicating a greater air entry value, and thus a smaller 'largest drainable' pore size. All of the values of [n.sub.WC] are small (<1.2), indicating a wide pore size distribution and a slow release of water with increasing dryness, which is typical of clay soils (van Genuchten 1980). Subsoils had a slower release of water than surface soil layers, as indicated by a slight but significantly smaller value of [n.sub.wc], and a smaller saturated moisture ratio (Table 4). However, PAWC values indicate only a slight decrease in available water with depth in the profile (Table 4). For all of the Vertosol profiles sampled, there appears to be ~13 -16% of the soil volume available for plant water extraction. Note, however, that this version of PAWC probably overestimates the in-situ PAWC, since field soils are never free to swell in all directions due to confinement and overburden pressures.
Structural form data
There were significant differences between the average values of only a few of the topsoil structural form attributes between the valleys, even though trends were visible (Table 5). It should be noted that these trends need to be treated with caution due to the small number of samples in some of the valleys. The samples from the Namoi and Gwydir valleys exhibited the largest macroporosities and mean pore intercept lengths (MPIL), and relatively small mean solid intercept lengths (MSIL). These attributes indicate porous surface soils with larger, more closely spaced macropores (Fig. 2a and c) enabling more rapid water and air flow (Vervoort and Cattle 2002). In contrast, the samples from the Macquarie valley exhibited the smallest MPIL and considerably less macroporosity. This indicates a more dense soil with only small pores, although some large fissures were generally present (Fig. 2b). Due to the small number of samples taken from the Lachlan valley and Darling Downs, comparisons with the other valley samples are not as clear. However, in general, the Lachlan valley samples appear to be similar to the Macquarie valley samples, while the Darling Downs samples have structural features more similar to the Namoi and Gwydir valley samples. Across all of the sampled surface soils, there was little difference in pore star lengths (PSL) and solid star lengths (SSL). The subsoil samples from the Namoi and Gwydir valleys exhibited significantly smaller values for macroporosity, surface area, MPIL and PSL, and significantly greater values for MSIL and SSL, than the corresponding surface soil samples. These data indicate that the subsoils are generally more dense than the overlying surface soils, which is quite clearly shown in a comparison of Fig. 2c and d.
Fundamental soil property data
The particle size distribution data (Table 6) indicated that the Gwydir and Namoi valley samples had the greatest average clay contents, and that the subsoils of these sites contained less sand and more fine clay than the topsoil. Silt contents were greatest in the Macquarie valley samples, while the Darling Downs samples contained the most non-clay sized material. The chemical data indicated that the samples from the Gwydir valley had a greater exchangeable Ca content and ECEC than the samples from the Namoi and Macquarie valleys, and a greater exchangeable Mg content than the samples from the Macquarie valley (Table 6). These trends seem to be a reflection of the clay type and mineral suite at these sample sites (Table 7), with the Gwydir valley and Darling Downs samples having little mica and kaolinite in the total and fine clay fraction. In contrast, mica and kaolinite were only present in the coarse clay fraction of the Namoi valley, but were present in both fractions in the Macquarie valley and Lachlan valley samples. There was also a greater abundance of smectitic 2:1 clay minerals in the samples from the Gwydir and Namoi valleys compared with the other areas.
Mean exchangeable [Na.sup.+] and ESP values were largest in the Namoi and Macquarie valley samples, and the extractable Fe content was greatest in the Lachlan valley samples. In contrast, CaC[O.sub.3] content was the smallest in the Gwydir valley samples. None of the other topsoil chemical properties measured displayed any trends across the valleys. General trends in the data were more a reflection of the depth variable, with clay content, EC, ESP, exchangeable [Na.sup.+], and extractable S[O.sub.4.sup.2-] increasing with depth, and total C content and sand content decreasing. However, the increase in EC with depth (Table 6) was mainly due to the very large EC values in the subsoil samples from the Gwydir valley. In the Namoi valley, subsoil EC values did not increase as much, and were not significantly different from the surface EC values. The average ESP of the topsoil samples was 2.8%, with a range of 0-9%. Sodic topsoils, indicated by ESP values >6%, were found in 2 of the samples from the Macquarie valley. The ESP of the subsoil (defined as below 400 ram) ranged from 6 to 37%, with an average of 13%, indicating that there is usually a considerable increase in sodicity with depth.
The mean [K.sub.s] values in this study are quite large compared with other Vertosol hydraulic conductivity datasets (Shaw and Yule 1978; Forrest et al. 1985; Bird et al. 1996). In the Forrest et al. (1985) study, the mean field [K.sub.s] value was 3.32 mm/h for a Vertosol topsoil near Narrabri, and 0.89 mm/h for a Vertosol topsoil near Dalby. Bird et al. (1996) found subsoil hydraulic conductivities around 10 mm/day on Vertosols in the Macquarie valley, and final infiltration rates ranged from 1 to 6 mm/h on Vertosols in the Emerald region (Shaw and Yule 1978). In contrast, the mean surface [K.sub.s] value of this study was 193 mm/h, with a mean value of 237 mm/h for the Namoi valley layers (which contributes the greatest number of samples to the dataset). These differences could be due to differences in soil management techniques before sampling, or to differences in the measurement methods employed. The latter reason appears most likely, as this study measured most infiltrations on the plant row and used the disc permeameter to estimate hydraulic conductivities. The plant rows are considerably more permeable than the subsoil below, partly because in many fields a tyne is pulled through the soil before the ridging operation. In addition, the cumulative amount of infiltrated water using the disc permeameter is only small, and the values obtained are generally only representative of the 50 mm of soil below the instrument (Minasny and McBratney 2000b). The Forrest et al. (1985) study used the Talsma method (Talsma 1969) for the field measurements, which is a ponded infiltration method, generally run over a longer time period and thus representative of a greater depth of soil. The work by Bird et al. (1996) used a well permeameter (Elrick et al. 1989) to estimate subsoil conductivities, whereas the research in the Emerald region involved measuring infiltration rates over a 5-h period (Shaw and Yule 1978).
At the 2 sites where the subsoil was sampled, the hydraulic conductivities below 400 mm depth were 2 orders of magnitude smaller than those of the overlying surface soils, despite the soil wetnesses not being significantly different. This large change in [K.sub.s] with depth may have been due to management factors at the locations (e.g. Chan 1982; Daniells 1989), but more probably highlights the dual effects of overburden pressure and large ESP on the hydraulic conductivity of Vertosols. Other evidence of the ESP-[K.sub.s] correlation, which was identified by McIntyre (1979), is indicated by the Macquarie valley surface samples, which yielded both the largest ESPs and the smallest [K.sub.S] values of all the surface samples. An additional depth trend is the increase in slope of the hydraulic conductivity curve (data not shown), meaning that the subsoil values for K decrease more rapidly with decreasing supply potential. Since depth of sampling strongly determined the value of hydraulic conductivity (Table 2), it is more appropriate to compare the values from the earlier measured datasets with the subsoil hydraulic conductivities in this study, for which the overall mean was 8.5 mm/h and with a mean of 10.6 mm/h for the more densely sampled Namoi valley sites.
The implications of this observation in terms of irrigation management are that the infiltration rate of the surface soil is very rapid, but that the depth of this layer with a large infiltration capacity is limited to the top 200-400 mm. Infiltration rates therefore decrease sharply once the upper part of the soil is saturated (after ~2 h based on the average [K.sub.s] observed in this study). This can lead to increased runoff or waterlogging if the irrigation water supply is not decreased once the topsoil is saturated. It is therefore important to consider the infiltration rate (or hydraulic properties) at 400 mm depth or lower when designing an irrigation scheme or irrigation management protocol. Table 8 illustrates the strong negative correlation of ESP with hydraulic conductivity, along with the general influence of exchangeable cations, organic carbon and texture on water flow in these soil types.
Although the measured shrink-swell and water retention properties of samples from the different valleys were generally not significantly different, several of the shrinkage parameters (e.g. [alpha], n) are thought to be indicative of soil structural condition. A physical interpretation of a value of n less than unity, which was the case for most samples analysed here, is that smaller values represent less porous samples; Coughlan et al. (1991) reviewed several studies of Vertosol physical properties and noted that those soils with noticeably 'poorer structure' consistently yielded n values <1. The parameter n has also been interpreted as reflecting shrink swell capacity, with soils exhibiting less shrin-swell capacity tending to have n values <1 (Braudeau et al. 1999). A further index that has been related to soil structure is the difference between n and s (McKenzie et al. 1991), with larger differences indicating more 'favourable structure'. In this dataset, the samples from the Namoi valley had significantly greater values of n-s than the Macquarie and Gwydir valley samples, suggesting a more porous structural form in the Namoi samples. This comparison is only moderately supported by the image analysis data describing structural attributes of these samples. The shrink-swell parameters most strongly correlated with physicochemical properties were the field bulk density, [U.sub.A] and [alpha]. Exhangeable [Na.sup.+] content and ESP were strongly positively correlated with the field bulk density and [U.sub.A], and negatively correlated with the [alpha] value, while exchangeable [K.sup.+] was negatively correlated with field bulk density and [U.sub.A] and positively correlated with [alpha]. The strongest influence on water retention properties was the depth of measurement, which again reflects the changing physico-chemical and textural properties down the profile. The exchangeable [Ca.sup.2+], [Mg.sup.2+] and [Na.sup.+] contents were positively correlated with [[alpha].sub.WC], while as would be expected, the sand content was positively, and the silt fraction negatively, correlated with [n.sub.WC].
The structural form parameters for the surface and subsoil layers are a clear reflection of the measured hydraulic conductivities. For the surface layers, those sites with the largest values of [K.sub.s] also exhibited significantly greater macroporosities, while subsoil layers of the Gwydir and Namoi valley pits yielded significantly smaller [K.sub.s] and macroporosity values than the overlying surface soil. The other subsoil structural attributes of smaller surface area, MPIL and PSL, and larger MSIL and SSL, are indicative of a dense structure with small pores, which accounts for the smaller [K.sub.s] values obtained in those layers. The strongest correlations of fundamental and soil structural properties again involve the exchangeable cations and textural components (Table 9). Exchangeable [Na.sup.+] content and ESP are both strongly negatively correlated with macroporosity and pore surface area, but strongly positively correlated with the aggregate-related attributes of MSIL and SSL. Exchangeable [K.sup.+] exhibits opposite correlations. The silt content is negatively correlated with MPIL, suggesting that pore continuity may be reduced by closer packing of soil solids where silt-sized particles are more prevalent. Interestingly, the extractable sulfate content was strongly negatively correlated with all the pore-related attributes, and strongly positively correlated with all the solid-related attributes. This may reflect the destruction of gypsum and the concomitant precipitation of pedogenic calcium carbonate during dry periods; such a sequence of reactions would liberate considerable sulfate, yet consume much exchangeable [Ca.sup.2+], leaving the clay exchange sites relatively saturated by exchangeable [Na.sup.+]. This series of reactions has recently been documented as occurring in Vertosols of semi-arid India during very dry periods (Srivastava et al. 2002).
In broad terms, the data indicate a north-south trend, with the northern sampled areas (Gwydir and Namoi valleys) exhibiting a more developed, porous soil structure than the samples from the 2 more southern regions (Lachlan and Macquarie valleys). This is corroborated by greater values of pore-related structural parameters, such as macroporosity and MPIL, in the surface samples from the northern areas. The physico-chemical properties (Table 6) indicate greater exchangeable [Ca.sup.2+] and [Mg.sup.2+] contents, ECEC values and clay contents for the samples from the Gwydir and Namoi valleys compared with the other locations. Earlier work in the Namoi valley has found more favourable surface structure to be related to greater clay, exchangeable [Ca.sup.2+], and exchangeable [K.sup.+] contents, while less favourable structure was related to greater silt, exchangeable [Mg.sup.2+], and organic carbon contents (Little et al. 1992). In this study, it appears that the exchangeable [Mg.sup.2+] had a less deleterious effect on Vertosol structural properties.
These variations in physico-chemical and structural properties may be attributable to broad differences in the composition of alluvium leading to the formation of these Vertosols. As Barzegar et al. (1995) noted, the presence of greater amounts of 2:1 clay minerals in a soil tends to increase the rate of structure reformation under wetting and drying cycles. Not only do the samples from the Darling Downs and Gwydir and Namoi valleys have the largest clay contents, their clay fractions are dominated by smectitic 2:1 clay minerals (Table 7). This dominance of smectite is consistent with alluvial parent material sourced primarily from basic igneous rocks; all 3 northern sampling areas have significant areas of Tertiary basalt lying directly up-catchment. In contrast, the clay mineral suites of the samples from the southern valleys are characterised by appreciable amounts of mica and kaolinite, which do not have the same propensity for shrink-swell behaviour as smectitic clays. Correspondingly, the up-catchment lithology of the Macquarie and Lachlan valleys include many felsic rock types of the Lachlan Fold Belt, along with comparatively smaller areas of Tertiary basalt. Such a combination of geological units would account for the more mixed clay mineral suite of the Vertosols formed down-catchment.
The database developed in this research contains the hydraulic, soil structural, and physicochemical charateristics of 23 sampled Vertosol horizons in south-eastern Australia. An overview of these characteristics highlights the specific nature and complexity of Vertosols. Measured hydraulic conductivities were large compared with earlier reported research on these soils, while overburden pressures in these soils are believed to cause a sharp increase in bulk density with depth below the surface. This is magnified by an increased sodicity in the subsoil at the locations sampled in this research. The combination of these two factors caused a decrease of 2 orders of magnitude in the hydraulic conductivity between the soil just below the surface crumb layer and at 400 mm depth at 2 sites. The surface crumb layer exhibited a much greater conductivity than the underlying soil. The development of the crumb layer appears to depend on several fundamental soil properties and appeared to reflect a general north-south trend, with Vertosols from northern NSW and southern Queensland having a more developed, porous structure than the Vertosols sampled in southern NSW. This trend reflects differences in clay content and type, with the northernmost areas having greater clay contents and proportions of swelling clay minerals, and a relative absence of mica and kaolinite in both the total and fine clay fractions. Measured shrink-swell and water retention properties of the surface soils did not vary greatly across the valleys, but changes in these properties were evident with depth.
Table 1. Location of the sampling sites and number of soil samples taken Area Sites Surface Subsoil Depth layers of subsoil (0-200 mm) layers sampled (mm) Darling Downs, Qld 2 2 0 Gwydir valley, NSW 4 4 2 650, 2100 Namoi valley, NSW 6 6 3 450, 850, 1300 Macquarie valley, NSW 4 4 0 Lachlan valley, NSW 2 2 0 Table 2. Mean values of [K.sub.s] for the surface soil (including the crumb layer and below the crumb layer) and the subsoil under a plant row, and mean values of K at -20 mm supply potential ([K.sub.-20]) for below-crumb layer soil under a plant row and under a furrow Within columns, values followed by the same letter are not significantly different at P = 0.05 [K.sub.s] (mm/h) [K.sub.-20] (mm/h) Surface Surface Subsoil Plant Furrow crumb layer below row below below crumb crumb layer crumb layer layer Darling Downs 414.9a 90.8ab -- 36.8 -- Gwydir 575.3a 351.3a 6.3 24.9 74.9 Namoi 697.4a 296.4a 10.6 32.2 13.6 Macquarie 38.4b 38.9b -- 17.7 12.7 Lachlan -- 41.6ab -- 23.1 -- Table 3. Average field bulk density and parameters of the Three Straight Line Model (McGarry and Malafant 1957) for the surface samples in each valley, and samples above (surface) and below (subsoil) 400 mm depth from the Gwydir and Namoi valleys Within rows, values followed by the same letter are not significantly different at P = 0.05 By valley Darling Gwydir Downs Field bulk 1.49a 1.31b density (g/[cm.sup.3]) [U.sub.A] (x [10.sup.-2] g/g) 1.50 3.62 [U.sub.B] (g/g) 0.41 0.33 [alpha] ([cm.sup.3]/g) 0.60 0.62 r ([cm.sup.3]/g) 0.03 0.04 n ([cm.sup.3]/g) 0.59 0.78 s ([cm.sup.3]/g) 0.07 0.24 RMSE 6.8 x [10.sup.-3] 5.2 x [10.sup.-3] By valley Namoi Macquarie Field bulk 1.28b 1.40ab density (g/[cm.sup.3]) [U.sub.A] (x [10.sup.-2] g/g) 7.90 6.33 [U.sub.B] (g/g) 0.40 0.25 [alpha] ([cm.sup.3]/g) 0.61 0.60 r ([cm.sup.3]/g) 0.23 0.05 n ([cm.sup.3]/g) 1.00 0.70 s ([cm.sup.3]/g) 0.13 0.21 RMSE 3.9 x [10.sup.-3] 2.3 x [10.sup.-3] By valley By depth Lachlan Average surface Field bulk 1.30b 1.26a density (g/[cm.sup.3]) [U.sub.A] (x [10.sup.-2] g/g) 5.00 5.00 [U.sub.B] (g/g) 0.22 0.36 [alpha] ([cm.sup.3]/g) 0.73 0.62a r ([cm.sup.3]/g) 0.09 0.12 n ([cm.sup.3]/g) 0.92 0.87 s ([cm.sup.3]/g) 0.25 0.19 RMSE 4.4 x [10.sup.-3] 4.7 x [10.sup.-3]a By depth Average subsoil Field bulk 1.41b density (g/[cm.sup.3]) [U.sub.A] (x [10.sup.-2] g/g) 6.00 [U.sub.B] (g/g) 0.37 [alpha] ([cm.sup.3]/g) 0.56b r ([cm.sup.3]/g) 0.05 n ([cm.sup.3]/g) 0.81 s ([cm.sup.3]/g) 0.15 RMSE 2.5 x [10.sup.-3]b Table 4. Average soil water characteristic parameters of surface samples from each valley, and samples above (surface) and below (subsoil) 400 mm depth from the Gwydir and Namoi valleys Within rows, values followed by the same letter are not significantly different at P = 0.05 By valley Darling Gwydir Downs [[alpha].sub.WC] (kPa) 12.65 20.65 [n.sub.WC] 1.15 1.14 [[??].sub.s] 0.92 1.06 RMSE 5.8 x [10.sup.-2] 6.10 x [10.sup.-2] PAWC (%) 15.8 14.2 By valley Namoi Macquarie [[alpha].sub.WC] (kPa) 14.61 15.26 [n.sub.WC] 1.16 1.15 [[??].sub.s] 1.11 0.94 RMSE 5.7 x [10.sup.-2] 6.4 x [10.sup.-2] PAWC (%) 15.8 14.9 By valley By depth Lachlan Average surface [[alpha].sub.WC] (kPa) 4.35 17.03a [n.sub.WC] 1.16 1.15a [[??].sub.s] 0.98 1.09a RMSE 4.7 x [10.sup.-2] 5.8 x [10.sup.-2] PAWC (%) 12.9 14.8 By depth Average subsoil [[alpha].sub.WC] (kPa) 95.8b [n.sub.WC] 1.12b [[??].sub.s] 0.89b RMSE 6.3 x [10.sup.-2] PAWC (%) 13.2 Table 5. Average derived soil structural form parameters of surface samples from each valley, and samples above (surface) and below (subsoil) 400 mm depth from the Gwydir and Namoi valleys Within rows, values followed by the same letter are not significantly different at P = 0.05 By valley Darling Gwydir Namoi Macquarie Lachlan Downs Macroporosity 0.12ab 0.16a 0.13ab 0.08b 0.07ab ([mm.sup.3]/ [mm.sup.3]) SA ([mm.sup.2]/ 0.56 0.78 0.62 0.45 0.38 [mm.sup.3]) MPIL (mm) 0.89a 0.88a 0.85ab 0.70b 0.76ab MSIL (mm) 5.24 4.00 4.80 6.75 7.35 PSL (mm) 1.87 1.85 1.88 1.69 1.65 SSL (mm) 17.12 13.63 15.02 20.05 25.03 By depth Average Average surface subsoil Macroporosity 0.14a 1.9 x [10.sup.-2]b ([mm.sup.3]/ [mm.sup.3]) SA ([mm.sup.2]/ 0.67a 0.12b [mm.sup.3]) MPIL (mm) 0.86a 0.66b MSIL (mm) 4.49a 21.32b PSL (mm) 1.87a 1.27b SSL (mm) 14.49a 44.73b Table 6. Physico-chemical soil properties of surface samples from each valley, and samples above (surface) and below (subsoil) 400 mm depth from the Gwydir and Namoi valleys Within rows, values followed by the same letter (a,b) are not significantly different at P = 0.05 By valley Darling Gwydir Downs Sand > 50 [micro]m (%) 24.4 10.2 Very fine sand 7.1ab 4.6b (50-20 [micro]m) (%) Silt 18.0 23.9 (20-2 [micro]m) (%) Coarse clay 20.4b 34.1a (2-0.2 [micro]m) (%) Fine clay 30.1 27.2 (<0.2 [micro]m) (%) pH(1:2 in Ca[Cl.sub.2]) 6.7a 7.1ab EC(1:5 water) (dS/m) 161.2 68.3 Exch. Ca (cmol(+)/kg) 23.1ab 23.9ab Exch. Mg (cmol(+)/kg) 10.1ab 12.5a Exch. Na (cmol(+)/kg) 5.0 x [10.sup.-2]b 0.9ab Exch. K (cmol(+)/kg) 0.8 1.1 Eff. CEC (cmol(+)/kg) 34.1ab 38.4a ESP (%) 0.1b 2.6ab Extract. S[O.sub.4.sup.2-] 6.9 3.3 (mg/kg) Extract. P (mg/kg) 47.8 36.0 CaC[O.sub.3] (g/kg) 2.6 2.2 Total C (%) 1.0 0.8 Extract. Fe (%) 8.0 x [10.sup.-2]b 13 x [10.sup.-2]b By valley Namoi Macquarie Sand > 50 [micro]m (%) 18.0 18.6 Very fine sand 5.8b 13.7a (50-20 [micro]m) (%) Silt 18.5 17.3 (20-2 [micro]m) (%) Coarse clay 27.1ab 24.3b (2-0.2 [micro]m) (%) Fine clay 30.6 26.1 (<0.2 [micro]m) (%) pH(1:2 in Ca[Cl.sub.2]) 7.3ab 7.5b EC(1:5 water) (dS/m) 157.6 180.5 Exch. Ca (cmol(+)/kg) 23.9a 16.4b Exch. Mg (cmol(+)/kg) 11.6a 8.3b Exch. Na (cmol(+)/kg) 1.1a 0.9ab Exch. K (cmol(+)/kg) 1.5 1.4 Eff. CEC (cmol(+)/kg) 38.2a 26.9b ESP (%) 3.2ab 3.6a Extract. S[O.sub.4.sup.2-] 42.0 32.4 (mg/kg) Extract. P (mg/kg) 40.4 32.0 CaC[O.sub.3] (g/kg) 3.4 12.3 Total C (%) 0.8 0.8 Extract. Fe (%) 5.0 x [10.sup.-2]b 6.0 x [10.sup.-2]b By valley By depth Lachlan Average surface Sand > 50 [micro]m (%) 22.5 15.7a Very fine sand 9.6ab 5.5 (50-20 [micro]m) (%) Silt 13.7 20.0 (20-2 [micro]m) (%) Coarse clay 18.7b 29.1 (2-0.2 [micro]m) (%) Fine clay 35.5 29.6a (<0.2 [micro]m) (%) pH(1:2 in Ca[Cl.sub.2]) 7.7b 7.2a EC(1:5 water) (dS/m) 149.2 130.1a Exch. Ca (cmol(+)/kg) 21.1ab 23.9 Exch. Mg (cmol(+)/kg) 9.8ab 11.8 Exch. Na (cmol(+)/kg) 0.9ab 1.1a Exch. K (cmol(+)/kg) 1.5 1.4a Eff. CEC (cmol(+)/kg) 33.3ab 38.3 ESP (%) 2.6ab 2.9a Extract. S[O.sub.4.sup.2-] 35.3 24.5a (mg/kg) Extract. P (mg/kg) 33.3 38.6 CaC[O.sub.3] (g/kg) 3.3 3.0a Total C (%) 0.6 0.8a Extract. Fe (%) 44 x [10.sup.-2]a 8.0 x [10.sup.-2] By depth Average subsoil Sand > 50 [micro]m (%) 7.4b Very fine sand 4.9 (50-20 [micro]m) (%) Silt 19.6 (20-2 [micro]m) (%) Coarse clay 33.2 (2-0.2 [micro]m) (%) Fine clay 34.9b (<0.2 [micro]m) (%) pH(1:2 in Ca[Cl.sub.2]) 7.5b EC(1:5 water) (dS/m) 503.1b Exch. Ca (cmol(+)/kg) 19.1 Exch. Mg (cmol(+)/kg) 12.6 Exch. Na (cmol(+)/kg) 4.26 Exch. K (cmol(+)/kg) 1.0b Eff. CEC (cmol(+)/kg) 37.0 ESP (%) 12.9b Extract. S[O.sub.4.sup.2-] 67.0b (mg/kg) Extract. P (mg/kg) 44.7 CaC[O.sub.3] (g/kg) 6.46 Total C (%) 0.5b Extract. Fe (%) 6.0 x [10.sup.-2] Table 7. Mineralogical properties of the surface samples from each valley Abundance of the individual clay minerals is indicated by the number of + signs Darling Downs Gwydir Namoi Total clay fraction (<2 [micro]m) Mica + Trace ++ Kaolinite + + ++ Smectite +++ Other 2:1 clay mineral (A) ++ + +++ Total clay fraction (<0.2 [micro]m) Mica Trace Trace Trace Kaolinite + + +/Trace Smectite + +++ +++ Other 2:1 clay mineral (A) ++ + ++ Macquarie Lachlan Total clay fraction (<2 [micro]m) Mica ++ ++1/2 Kaolinite ++ +1/2 Smectite Trace Other 2:1 clay mineral (A) ++ + Total clay fraction (<0.2 [micro]m) Mica +1/2 + Kaolinite ++ +1/2 Smectite ++ Other 2:1 clay mineral (A) ++ ++1/2 (A) Mineral not clearly identifiable from X-ray diffraction patterns, but most probably an interstratified mica-smectite or another Smectite intergrade. Table 8. Correlations of fundamental soil properties and the saturated hydraulic conductivity and the log(saturated hydraulic conductivity) Fundamental soil properties with non-significant correlations are not included [K.sub.s] (mm/h) log ([K.sub.s] mm/h) Exchangeable Ca 0.24 0.36 * Exchangeable Na -0.32 ([dagger]) -0.57 * Exchangeable K 0.22 0.45 * ESP -0.38 * -0.64 * Total C content 0.23 0.35 * Total sand content 0.06 0.38 * Fine clay content -0.26 -0.36 * Total clay content -0.06 -0.33 ([dagger]) ([dagger]) P < 0.1; * P < 0.05. Table 9. Correlations between fundamental soil properties and logarithms of the soil structural parameters Fundamental soil properties with non-significant correlations are not included Log(Macropor) Log(SA) pH(1:2 Ca[Cl.sub.2]) -0.14 0.00 Exchangeable Na -0.59 * -0.60 * Exchangeable K 0.29 ([dagger]) 0.38 * ESP -0.56 * -0.56 * S[O.sub.4.sup.2-] content -0.74 * -0.70 * Total C content 0.31 ([dagger]) 0.34 ([dagger]) Silt content -0.25 -0.12 Log(MPIL) Log(MSIL) Log(PSL) pH(1:2 Ca[Cl.sub.2]) -0.47 * 0.06 -0.36 * Exchangeable Na -0.18 0.58 * -0.24 Exchangeable K -0.14 -0.35 * -0.11 ESP -0.22 0.55 * -0.28 S[O.sub.4.sup.2-] content -0.43 * 0.73 * -0.44 * Total C content 0.05 -0.30 0.02 Silt content -0.49 * 0.18 -0.37 Log(SSL) pH(1:2 Ca[Cl.sub.2]) 0.10 Exchangeable Na 0.47 * Exchangeable K -0.33 ([dagger]) ESP 0.43 * S[O.sub.4.sup.2-] content 0.73 * Total C content -0.24 Silt content 0.30 ([dagger]) P < 0.1; * P < 0.05.
We would like to acknowledge the Australian Cotton Cooperative Research Centre for funding this research through CRC Project 1.2.4.
Barzegar AR, Rengasamy P, Oades JM (1995) Effects of clay type and rate of wetting on the mellowing of compacted soils. Geoderma 68, 39-49. doi:10.1016/0016-7061 (95)00022-G.
Berndt RD, Coughlau KJ (1976) The nature of changes in bulk density with crack content in a cracking clay. Australian Journal of Soil Research 15, 27-37.
Bird TL, Willis TM, Melville GJ (1996) Subsoil hydraulic conductivity estimates for the lower Macquarie Valley. Australian Journal of Soil Research 34, 213-228.
Bouma J, Jongerius A, Schoonderbeek D (1979) Calculation of saturated hydraulic conductivity of some pedal day soils using micromorphometric data. Soil Science Society of America Journal 43, 261-264.
Brasher BR, Franzmeier DE Valassis SV, Davidson SE (1966) Use of saran resin to coat natural soil clods for bulk density and water retention measurements. Soil Science 101, 108-108.
Braudeau E, Constantini JM, Bellier G, Collieuille H (1999) New device and method for soil shrinkage curve measurement and characterization. Soil Science Society, of America Journal 63, 525-535.
Bronswijk JJB (1990) Shrinkage geometry of a heavy clay soil at various stresses. Soil Science Society, of America Journal 54, 1500-1502.
Burdine NT (1953) Relative permeability calculations from pore size distribution data. Transactions o/the American Institute of Mining, Metallurgical, and Petroleum Engineers 198, 71-77.
Cattle SR, Farrell RA, McBramey AB, Moran CJ, Roesner EA, Koppi AJ (2001) '[c] Solicon--PC Version 1.0.' (The University of Sydney & Cotton Research and Development Corporation)
Chan KY (1982) Shrinkage characteristics of soil clods from a grey clay under intensive cultivation. Australian Journal of Soil Research 20, 65-68.
Clothier BE (2000) Infiltration. In 'Soil and environmental analysis'. 2nd edn (Eds KA Smith, CE Mullins) pp. 239-280. (Marcel Dekker, Inc.: New York)
Collis-George N (1977) Infiltration equations for simple soil systems. Water Resources Research 13, 395-403.
Collis-George N, Bridge BJ (1973) The effect of the height of sample and confinement on the moisture characteristic of an aggregated swelling clay soil. Australian Journal of Soil Research 11, 107-120.
Connolly RD (1998) Modelling effects of soil structure on the water balance of soil-crop systems: a review. Soil and Tillage Research 48, 1-19. doi:10.1016/S0167-1987(98)00128-7.
Coquet Y, Touma J, Boivin P (1998) Comparison of soil linear shrinkage curve from extracted cores and in situ. Australian Journal of Soil Research 36, 765-781.
Coughlan KJ, McGarry D, Loch RJ, Bridge B, Smith GD (1991) The measurement of soil structure some practical initiatives. Australian Journal of Soil Research 29, 869-889.
Crescimanno G, Provenzano G (1999) Soil shrinkage characteristic curve in clay soils. Soil Science Society of America Journal 63, 25-32.
Cresswell HP, Paydar Z (1996) Water retention in Australian soils. I. Description and prediction using parametric functions. Australian Journal of Soil Research 34, 195-212.
Daniells IG (1989) Degradation and restoration of soil structure in a cracking grey clay used for cotton production. Australian Journal of Soil Research 27, 455-469.
Deeks LK, Williams AG, Dowd JF, Scholefield D (1999) Quantification of pore size distribution and the movement of solutes through isolated soil blocks. Geoderma 90, 65-86. doi:10.1016/S0016-7061 (98)00092-5.
Elrick DE, Reynolds WD, Tan KA (1989) Hydraulic conductivity measurements in the unsaturated zone using improved well analysis. Ground Water Monitoring Review 9, 193 204.
Forrest JA, Beatty J, Hignett CT, Pickering JH, Williams RGP (1985) A survey of the physical properties of wheatland soils in Eastern Australia. Divisional Report No. 78, CSIRO Division of Soils, Australia.
Geeves GW, Cresswell HE Murphy BW, Gessler PE, Chartres CJ, Little IP, Bowman GM (1995) The physical, chemical and morphological properties of soils in the wheat-belt of Southern NSW and Northern Victoria. NSW Department of Conservation and Land Management/CSIRO Division of Soils Occasional Report.
van Genuchten MTh (1980) A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Science Society of America Journal 44, 892-898.
Gerard-Marchant P, Angulo-Jaramillo R, Haverkamp R, Vauclin M, Groenevelt P, Elrick DE (1997) Estimating the hydraulic conductivity of slowly permeable and swelling materials from single-ring experiments. Water Resources Research 33, 1375-1382.
Grant CD, Blackmore AV (1991) Self-mulching behaviour in clay soils: its definition and measurement. Australian Journal of Soil Research 29, 155-173.
Hamblin AP (1985) The influence of soil structure on water movement, crop root growth, and water uptake. Advances in Agronomy 38, 95-158.
Horgan GW (1998) Mathematical morphology for analysing soil structure from images. Geoderma 49, 161-173. doi:10.1046/J.1365-2389.1998.00160.X.
Hue NV, Adams F (1979) Indirect determination of micrograms of sulfate by barium adsorption spectroscopy. Communications in Soil Science and Plant Analysis 10, 841-851.
Isbell RF (1996) 'The Australian Soil Classification.' (CSIRO Publishing: Collingwood, Vic.)
Kilmer VJ, Alexander LT (1949) Methods of making mechanical analysis of soils. Soil Science 68, 15-24.
Klute A (1986) Water retention: Laboratory methods. In 'Methods of soil analysis. Part 1. Physical and mineralogical methods'. (Ed. A Klute) pp. 635-662. (ASA/CSSA/SSSA: Madison, WI)
Little IP, Ringrose-Voase AJ, Ward WT (1992) Surface structure in grey clays of northwestern New South Wales in relation to micromorphology, cation suite and particle-size attributes. Australian Journal of Soil Research 30, 1-16.
Loeppert RH, Suarez DL (1996) Carbonate and gypsum. In 'Methods of soil analysis. Part 3 Chemical methods'. (Ed DL Sparks) pp. 437474. (Soil Science Society of America: Madison, WI)
McBratney AB, Moran CJ, Stewart JB, Cattle SR, Koppi AJ (1992) Modifications to a method of rapid assessment of soil macropore structure by image analysis. Geoderma 53, 255-274. doi: 10.1016/0016-7061(92)90058-F.
McGarry D (1996) The structure and grain size distribution of Vertisols. In 'Vertisols and technologies for their management'. (Eds N Abroad, A Mermut) Developments in Soil Science 24, 231-259.
McGarry D, Daniells IG (1987) Shrinkage curve indices to quantify cultivation effects on soil structure of a Vertisol. Soil Science Society of America Journal 51, 1575-1580.
McGarry D, Malafant KWJ (1987) The analysis of volume change in unconfined units of soil. Soil Science Society of America Journal 51, 290-297.
McIntyre DS (1979) Exchangeable sodium, subplasticity and hydraulic conductivity of some Australian soils. Australian Journal of Soil Research 17, 115-120.
McIntyre DS, Loveday J, Watson CL (1982) Field studies of water and salt movement in an irrigated swelling clay soil. I. Infiltration during ponding. Australian Journal of Soil Research 20, 81-90.
McKenzie DC, Abbott TS, Higginson FR (1991) The effect of irrigated crop production on the properties of a sodic Vertisol. Australian Journal of Soil Research 29, 543-548.
McKenzie NJ, Jacquier DJ, Ringrose-Voase AJ (1994) A rapid method for estimating soil shrinkage. Australian Journal of Soil Research 32, 931-938.
Mehra OP, Jackson ML (1960) Iron oxide removal from soils and clays by a dithionite-citrate system bufferd with sodium bicarbonate. Clays and Clay Minerals 7, 317-327.
Minasny B, McBratney AB (2000a) Hydraulic conductivity pedotransfer functions for Australian soil. Australian Journal of Soil Research 38, 905-926. doi: 10.1071/SR99110.
Minasny B, McBratney AB (2000b) Estimation of sorptivity from disc-permeameter measurements. Geoderma 95, 305-324. doi:10.1016/S0016-7061(99)00096-8.
Montgomery J, Faulkner R, MacLeod D (1999) Managing the hydrological cycle in relation to cotton. In 'Proceedings of the Cotton CRC Research Conference'. pp. 60-62. (Cooperative Research Centre for Sustainable Cotton Production: Narrabri, NSW)
Mualem Y (1976) A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resources Research 12, 513-522.
Philip JR (1969a) Moisture equilibrium in the vertical in swelling clay soils. I. Basic theory. Australian Journal of Soil Research 7, 99-120.
Philip JR (1969b) Hydrostatics and hydrodynamics in swelling soils. Water Resources Research 5, 1070-1077.
Philip JR, Smiles DE (1969) Kinetics of sorption and volume change in three component systems. Australian Journal of Soil Research 7, 1-19.
Pillai-McGarry U, Collis-George N (1990) Laboratory simulation of the surface morphology of self-mulching and non self-mulching Vertisols. II. Quantification of visual features. Australian Journal of Soil Research 28, 141-152.
Prebble RE (1970) Physical properties of 17 soils from Queensland. CSIRO Division of Soils Technical Memorandum 10/70.
Rayment GE, Higginson FR (1992) 'Australian laboratory handbook of soil and water chemical methods.' (Inkata Press: Sydney)
Rawls WJ, Brakensiek DL, Saxton KE (1982) Estimation of soil water properties. Transactions of the American Society of Agricultural Engineers 25, 1316-1320, 1328.
Shaw RJ, Thorburn PJ (1985) Prediction of leaching fraction from soil properties, irrigation water and rainfall. Irrigation Science 6, 73-83.
Shaw RJ, Yule DF (1978) The assesment of soils for irrigation Emerald, Queensland. Agricultural Chemistry Branch Technical Report No. 13. Queensland Department of Primary Industries, Brisbane.
Smiles DE (2000) Hydrology of swelling soils: a review. Australian Journal of Soil Research 38, 501-521.
Smiles DE, Rosenthal MJ (1968) The movement of water in swelling materials. Australian Journal of Soil Research 6, 237-248.
Srivastava P, Bhattacharyya T, Pal DK (2002) Significance of the formation of calcium carbonate minerals in the pedogenesis and management of cracking clay soils (Vertisols) of India. Clays and Clay Minerals 50, 111-126. doi:10.1346/000986002761002874.
Talsma T (1969) In situ measurement of sorptivity. Australian Journal of Soil Research 7, 269-276.
Talsma T, van der Lelij AJ (1976) Infiltration and water movement in an in situ swelling soil during prolonged ponding. Australian Journal of Soil Research 14, 337-349.
Tariq A, Durnford DS (1993) Soil volumetric shrinkage measurements: a simple method. Soil Science 155, 325-330.
Tariq A, Durnford DS (1997) Moisture retention of a swelling soil under capillary and overburden pressures. Journal of Hydrology 203, 119-126. doi:10.1016/S0022-1694(97)00089-9.
Vervoort RW, Cattle SR (2002) Linking hydraulic conductivity and tortuosity parameters to pore space geometry and pore-size distribution. Journal of Hydrology 272, 39-49.
Whittig LD, Allardice WR (1986) X-ray diffraction techniques. In 'Methods of soil analysis. Part 1. Physical and mineralogical methods'. (Ed. A Klute) pp. 331 362. (ASA/CSSA/SSSA: Madison, WI)
Wosten JHM, van Genuchten MTh (1988) Using texture and other soil properties to predict the unsaturated soil hydraulic functions. Soil Science Society of America Journal 62, 176-1770.
Manuscript received 18 December 2002, accepted 16 June 2003
R. W. Vervoort (A,B), S. R. Cattle (A), and B. Minasny (A)
(A) Faculty of Agriculture, Food and Natural Resources, The University of Sydney, NSW 2006, and The Australian Cotton Cooperative Research Centre, Locked Bag 59, Narrabri, NSW 2390, Australia. (B) Corresponding author; email: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Vervoort, R.W.; Cattle, S.R.; Minasny, B.|
|Publication:||Australian Journal of Soil Research|
|Date:||Dec 1, 2003|
|Previous Article:||The use of remote sensing data to extract information from agricultural land with emphasis on soil salinity.|
|Next Article:||Deep banding improves residual effectiveness of manganese fertiliser for bauxite residue revegetation.|