Increases in organic carbon concentration and stock after clay addition to sands: validation of sampling methodology and effects of modification method.
Sands are defined as profiles with sand to loamy sand topsoil texture (clay content <5%) with or without the presence of a clay B horizon; such profiles fall within the orders Chromosol, Sodosol, Kurosol, Tenosol, Calcarosol, Ferrosol and Kandosol (Australian Soil Classification; Isbell 2002). Sands cover ~80% of the Australian continent (Hatton et al. 2011) and 86% of Australia's cereal-cropping area (McKenzie et al. 2002). However, crop growth on sands is limited by poor water-and nutrient-retention capacity and, in some cases, water repellence. One strategy to improve crop production on sands is the addition of clay (clay modification), which increases the retention of water and nutrients. Modification methods include clay spreading, delving and/or spading (Davenport et al. 2011). These methods result in different distributions of the clay in the profile. Selection of the appropriate method is determined by the depth to clay-rich subsoil. In South Australia, the area of deep sand and sand over clay soils under agricultural production is ~2.7 Mha, with an estimated 2 Mha deemed suitable for clay modification (J. Hall, pers. obs.).
Clay spreading has been extensively used since the early 1990s in the south-east of South Australia (Cann 2000) and Western Australia (Carter and Hetherington 1994). Clay spreading is the only available option for deep sands where clay-rich subsoil is deeper than 70 cm from the surface. Clay delving, developed in the early 1990s, is used where clay-rich subsoil is present within 70cm depth (Desbiolles et al. 1997). For delving, specially designed tines raise the subsoil into the topsoil sand where it can be distributed and mixed in the topsoil. Spading has been used since the late 2000s as a modification method where clay-rich subsoil is within 30 cm of the surface. The spader can be used to both raise and incorporate the clay, resulting in a relatively even mix of subsoil clay and topsoil sand.
Clay modification was developed initially to overcome water repellence in sands (Ma'shum et al. 1989; Ward and Oades 1993; Cann 2000; Harper et al. 2000; McKissock et al. 2000; Betti et al. 2015). Additional benefits include yield increases of 20-130% (Hall et al. 2010; Davenport et al. 2011), increased nutrient availability (Hall et al. 2010; Bailey and Hughes 2012), reduction in frost damage (Rebbeck et al. 2007) and increased root growth (Hall et al. 1994; Bailey et al. 2010). Hall et al. (2010) reported that, 5 years after claying, organic carbon (OC) concentration increased by 0.2% in the top 10 cm. However, they did not report OC at depths greater than 10 cm.
Soils are the largest terrestrial sink for OC (Sanderman et al. 2010), and agricultural practices can lead to a loss of 50-75% of native soil OC (Lai 2007). On the other hand, improved management of cropland (improved rotations, adoption of no-till or stubble retention) can sequester OC at rates of 0.2-0.3 t C [ha.sup.-1] [year.sup.-1] compared with conventional management (Sanderman et al. 2010). However, the low nutrient-and water-holding capacity of sands make it difficult to increase their OC content by using those management practices (Hall et al. 2010). Clay modification may be an option to increase the OC concentration and storage capacity of sands, considering that in natural soils there is a positive correlation between clay and OC concentration (Dalai and Mayer 1986; Baldock and Skjemstad 1999). Sands have low OC concentration due to a combination of low OC input from plant residues and low OC protective capacity caused by low clay concentration (Jenkinson 1988; Baldock and Skjemstad 2000). Clay soils, on the other hand, have high OC storage capacity because decomposers have reduced access to organic matter. Accessibility is reduced by the binding of organic matter to the large specific surface areas of clay particles and by occlusion in aggregates formed by clay (Tisdall and Oades 1982; Skjemstad et al. 1993). Therefore, addition of clay to sands could increase OC concentrations by addressing constraints to plant production and by reducing OC decomposition through protection of organic matter from microbes.
The binding potential of clays depends on clay mineralogy, sesquioxide and carbonate concentration, and formation of stable micro-aggregates (Denef et al. 2001a, 2001b; Six et al. 2004; Fernandez-Ugalde et al. 2011 ; Saidy et al. 2012). In clay-modified sands, additional factors influencing OC retention capacity are clay application rate and distribution (including both depth of incorporation and size of clay clods) and time since modification. Determination of OC content in clay-modified soils requires reliable data on OC concentrations and distribution. However, the highly variable nature of clay distribution in these soils may necessitate the development of new sampling procedures.
Soil sampling procedures have significant impact on the validity and usefulness of data collected. Many sampling procedures have been described for Australian soils and are designed for certain purposes (McKenzie 2008). For carbon accounting, McKenzie et al. (2000, 2002) recommend collection, across a grid 25 m by 25 m, of 25 samples that are combined into five samples for analysis. Separate analysis of the five samples provides an assessment of the degree of variability of OC concentration within the grid, but also reduces costs compared with analysis of all 25 samples. Wilson et al. (2010) evaluated a set of soil indicators (total OC, bulk density, pH and total nitrogen) and found that the sampling intensity needed to achieve defined levels of precision and confidence differed among land uses and soil properties in northern New South Wales. They concluded that 10 samples across a sampling area of 25 m by 25 m yielded adequate precision and confidence for OC. The National Soil Carbon Research Program (SCaRP) (DAFF 2009) was designed to examine variations in OC content and composition under different agricultural practices and soil types across Australia (Sanderman et al. 2011). The protocol involved collecting 10 samples from a grid 25 m x 25 m. These methodologies were all designed to deal with natural soil variability. Clay-modified soils are highly heterogeneous horizontally and vertically owing to varying amounts and sizes of subsoil clay clods mixed with sandy topsoil. This greater variability may require an adapted sampling methodology, especially for clay-delved soils that have the greatest heterogeneity (Betti et al. 2015).
This study was designed to determine (i) the number of soil samples required within a 25-m grid for accurate assessment of OC and BD in clay-modified soils; and (ii) OC concentration, bulk density (BD) and OC stocks in clay-modified compared with unmodified soil.
Materials and methods
Study area and site selection
Two agricultural properties were selected in regions of South Australia where clay modification is common, the South East (SE) and Eyre Peninsula (EP) with temperate climate and rainfall predominantly in winter. Current land use (since clay modification) for both locations is winter cropping with a rotation of cereal (barley, wheat), canola, beans or vetch. Land use before clay modification was annual pasture with limited stock-carrying capacity (due to the inherent limitations to plant production in sands) and cropping in favourable seasons. Both properties had texture-contrast soils, sand overlying clay at 30-70 cm depth (Table 1). The clay was dominated by kaolinite with small proportions of illite with interstratified clay minerals (illite and kaolinite). The SE site, near Bordertown (UTM zone 54H, 475469E, 5993943N), had an average annual rainfall of 405 mm. The unmodified control was characterised by a 0-10 cm of sand with 3% clay and 0.82% OC and a B2 horizon with 38% clay, 58% sand, 4% silt and 0.29% OC. The EP site, near Ungarra (UTM zone 53H, 595313E, 6229862N), had an average annual rainfall of 345 mm. The unmodified control was characterised by 0-10 cm of sand with 1% clay and 0.38% OC over a B2 horizon with 46% clay, 48% sand, 6% silt and 0.25% OC. On each property, three areas within a 1-km radius were selected that were previously modified by clay spreading, delving and spading, and an unmodified control area. The clay-modification treatments were not replicated. At both sites, the spaded treatments received additional organic matter at the time of modification, at the SE site it was cereal straw at 61 [ha.sup.-1], and at the EP site lucerne hay at 10 t [ha.sup.-1].
Soil sampling and processing
A 25-m grid was established in a representative area for each treatment in accordance with several Australian studies measuring soil OC (McKenzie et al. 2002; Wilson et al. 2010; Sanderman et al. 2011). Individual soil cores were taken from 20 randomly selected grid intersects by using an hydraulic, trailer-mounted corer (diameter 50 mm). Sampling depth was determined by horizon from cores collected in the field. Samples were collected in depth increments of 0-10, 10-20, 20-30 and 30-50 cm at the SE location and 0-10, 10-30 and 30-50 cm at the EP location. Soil samples were collected after harvest, in November 2011 for the EP site and in February 2012 for the SE site.
Soil samples were processed following the method described in Sanderman et al. (2011). Soil samples were weighed, dried at 40[degrees]C for 48 h, re-weighed and sieved through a 2-mm sieve. The >2-mm fraction was ground, and because no gravel was present, it was mixed into the <2-mm sample. The combined <2-mm soil samples were analysed for OC (Walkley-Black method) and particle size.
Bulk density samples were collected at three random grid intersects for each depth by using the intact core method (Cresswell and Hamilton 2002) with fixed-volume cores (diameter 73 mm, height 50 mm). Samples were weighed, dried at 105[degrees]C, re-weighed and sieved through a 2-mm sieve. The >2-mm fraction was hand-ground and weighed, and the volume calculated by displacement of water. The BD of the fine-earth (<2-mm) fraction ([BD.sub.fe]) was calculated.
Organic C stock of the 0-30 cm depth was calculated by using the average OC concentration (%) and [BD.sub.fe] (g [cm.sup.-3]) for each site according to the equation (Ellert et al. 2001):
Organic C stock (t [ha.sup.-1]) = OC concentration x [BD.sub.fe] x soil depth
Because OC concentration changes with depth, small differences in sampling depth among cores can affect the calculated OC stock. To account for this, OC stock to 30 cm was adjusted to an equivalent soil mass (Ellert and Bettany 1995). The equivalent soil mass of 5000 Mg soil [ha.sup.-1] was used, which is the maximum soil mass of the <2-mm fraction across all samples. For cores with <5000 Mg [ha.sup.-1] soil mass, a fraction of the 30-50 cm depth was added to achieve equivalent soil mass and the OC stock was calculated as: [OC.sub.0-10] + [OC.sub.10-20] + [OC.sub.20-30] + fraction of [OC.sub.30-50].
Data on OC concentration and BD for each site were analysed by repeated-measures analysis of variance (ANOVA) with depth as repeated-measures. The depth x modification interaction was significant and was used for multiple comparison tests (Tukey P [less than or equal to] 0.05) in GENSTAT 15th edition (VSN International, Hemel Hempstead, UK). Organic C stock was not analysed statistically because it was based on average OC and BD values.
Determination of sample number for reliable estimation of OC content
Several Australian studies (e.g. Wilson et al. 2010; Sanderman et al. 2011) have shown 10 samples within a 25-m grid to be adequate for OC sampling in natural soils, but it is not known whether this is also the case for clay-modified soils. The data from the 20 soil samples for each site were used to find the sample number for reliable estimation of OC content in clay-modified soils via two methods.
The average OC concentration was calculated by progressively averaging the OC concentrations of sample 1, sample 1 + 2, sample 1 + 2 + 3, etc., until all 20 samples were accounted for. This process was repeated 20 times, randomising the order of the samples averaged. The variance and coefficient of variation were determined for each progression. Sample number was estimated by plotting the variance and coefficient of variation against increasing sample number from 1 to 20 (Fig. 1). The sample number required for reliable estimation of OC content is reached when variance = 0.001 and coefficient of variation = 5%.
The method uses Scheffe's formula, n = t[alpha]2 x s2/[(x x d).sup.2], where [t.sub.[alpha]] is Student's t-test at level of confidence [alpha], s2 is sample variance, x is sample mean and d is level of precision ([+ or -]d%) as discussed in Wilson et al. (2010). Sample number was determined by using [alpha] = 5% and 10% ([alpha] =10%, or 90% confidence has been considered adequate for soil samples; Schipper and Sparling 2000; Wilson et at. 2010) and at precision levels d = 5%, 10%, 15% and 20% (precision levels previously used were 10% by Schipper and Sparling (2000) and 15% by Wilson et al. (2010)). The sample number required for reliable estimation of OC concentration was taken as the highest sample number across all depths for the site.
The sample number required for reliable estimation of BD was also determined using Scheffe's formula where a and d are both 10%.
Sampling methodology for delved soils
Delved sites have distinct areas of clay modification with strips along the delve line spaced 0.8-1.5 m apart within which the soil is disturbed to ~60-75 cm (Davenport et al. 2011), whereas the soil between the strips (off delve line) was modified to ~10-15 cm by cultivation after delving. At the EP site, two sampling methods were compared in the 25 m by 25m grid:
(i) Twenty samples were collected based on the proportion of area covered by the delve lines (30%, six samples) and between delve lines (70%, 14 samples) (stratified sampling).
(ii) Forty samples were collected, 20 from the delve line and 20 from the off-delve areas (whole-area sampling)
Three BD samples were collected from each area (delve line and between delve line) for each depth. The appropriate sampling methodology was determined separately for both OC and BD by comparing the sample number required for reliable estimation of OC content for (i) stratified sampling and (ii) whole-area sampling.
Adjusting the confidence levels from [alpha] = 5% to [alpha] = 10% resulted in small differences in the sample number required for reliable estimation of OC content calculated by using Scheffe's formula (data not shown). However, the number of samples required was slightly lower with [alpha] = 10%. Therefore, this value was used for the calculations of sample number as in previous studies (Schipper and Sparling 2000; Wilson et al. 2010).
With a confidence level of 10%, the calculated sample number (using Scheffe's formula, method 2) required for reliable estimation of OC concentration varied greatly depending on the degree of precision selected. Large differences in optimum sample number (1-230) were calculated for the precision percentages 5%, 10%, 15% and 20%, with the largest sample number (230) required in soil with the lowest OC concentration (Table 2); therefore, the following rules for precision were applied to determine the sample number required for reliable estimation of OC depending on OC concentration:
(i) >0.9% OC: precision 10% (> [+ or -] 0.09%)
(ii) 0.5-0.9% OC: precision 15% ([+ or -] 0.08-0.09%)
(iii) <0.5% OC: precision 20% (<[+ or -] 0.1 %)
For example, when using a 5% precision ([+ or -] 0.02%) in the 20-30 cm layer of the SE delved treatment (Table 2) with an OC concentration of 0.4%, 146 samples in a 25-m grid are required. By using the proposed rules, because the OC concentration is <0.5%, a precision of 20% ([+ or -] 0.08%) would be selected, therefore requiring only 10 samples.
The highest value for each depth increment was selected as the sample number required for reliable estimation of OC concentration for a given treatment (Table 3). This approach was validated by comparison of the sample number required for reliable estimation of OC concentration (Table 4) derived by using plots of variance (Fig. 1). It can be concluded that 10 samples within a 25-m grid for clay-modified soils and 15 samples for unmodified control soils are sufficient for reliable estimation of OC concentration.
By using the precision and confidence levels determined for OC sampling ([alpha] and d = 10%), the sample number required for reliable estimation of BD was calculated (Table 3). Higher sample numbers were required for the SE (two-five samples) than the EP site (one or two samples).
Sampling methodology for delved soils determined that the sample number for reliable estimation of OC was 10 for stratified sampling and totals of 31 for SE and 18 for EP sites when sampling delve-line and off-delve areas separately.
There was no significant difference in BD values (data not shown) between delve-line and off-delve areas. Two samples were required for reliable estimation of BD in both areas.
Effect of clay modification
In unmodified control soil, the highest OC concentration occurred in the 0-10cm layer (AI) and the lowest at 10-30cm (bleached A2), with intermediate concentrations at 30-50 cm (clay B2 horizon). In the clay-modified treatments, the 0-10 cm layer also had the highest OC concentration, but the decrease to the 10-30 cm layer was less pronounced than in the control at the SE site in all modification treatments, and at the EP site in the spaded treatment only (Tables 5, 6). The EP delved treatment had OC distribution similar to the unmodified control site.
Compared with the unmodified control, OC concentrations were up to 2-fold higher at 0-10 cm depth in all EP clay-modified treatments and 25% higher in the SE clay spread treatment. The largest increase in OC concentrations relative to the control occurred at 10-30 cm depth, with up to 3-fold higher concentrations in all SE clay-modified treatments and the EP spaded treatment. At 30-50 cm depth, compared with the control, OC concentrations were higher in the EP spaded treatment but significantly lower in the EP clay spread treatment (Fig. 2; Appendix 1, P<0.001 treatment x depth interaction).
The coefficient of variation for OC ranged between 15% and 78% at the SE site and 14% and 48% at the EP site. The greatest variability generally occurred in layers with the lowest OC concentration (Tables 5, 6).
There were no significant differences in BD among the treatments (Tables 5, 6). The BD values ranged from 1.36 to 1.75g [cm.sup.3] at the SE site and from 1.51 to 1.92g [cm.sup.3] at the EP site. The coefficient of variation for BD ranged between 1% and 19% at the SE site and 1% and 7% at the EP site.
Organic C stocks were calculated based on equivalent soil mass of 5000 Mg [ha.sup.-1]. Stocks in unmodified controls were 20 and 12t [ha.sup.-1] at the SE and EP sites. The highest OC stocks were in the clay spread (33.5t [ha.sup.-1]) and delved (34t [ha.sup.-1]) treatments at the SE site and in the spaded treatment (34.5t [ha.sup.1]) at the EP site (Table 7). OC stocks in the clay-modified treatments were 7.5 - 14t [ha.sup.-1] greater at the SE site and 1-22.5t [ha.sup.-1] greater at the EP site than in the unmodified control, an annual increase at SE of 1.5 - 1.9t [ha.sup.-1] and at EP of 0.1 - 7.5t [ha.sup.-1] since clay modification (Table 7).
Relationship between organic carbon and clay concentrations
Clay concentrations of the unmodified control were 1-3% in the surface 30 cm, and 38% (SE) and 46% (EP) in the undisturbed clay horizon (B2). Clay concentration of the clay-modified treatments varied with depth, treatment and site; at 0-30 cm depth, the clay concentration ranges were 7-20% (SE) and 2-18% (EP) (Tables 5, 6).
When the whole profile (0-50 cm) was considered, there was no correlation ([R.sup.2] = 0.2) between OC and clay concentration. However, when the soil profile was divided into depth increments, OC concentration was positively correlated with clay concentration, with [R.sup.2] = 0.6 at 0-10 cm depth and 0.7 at 10-30cm depth (Fig. 3). There was no correlation ([R.sup.2] = 0.4) between OC and clay concentration at 30-50 cm, which was predominantly the B2 clay horizon.
Incorporation of subsoil clay into sand topsoil increased OC concentration, most likely as a result of increased OC input through improved plant growth, and greater protection from microbial breakdown of OC by binding to clay. Organic C stock increased by 7.5-14 and 1-22.5 t [ha.sup.-1], or an annual rate of 1.5-1.9 and 0.1-7.5 t C[ha.sup.-1] for the SE and EP sites. The annual increase of 7.5 t C[ha.sup.-1] is greater than would be expected from crop residue inputs alone. However, the spaded treatment from which this high value was obtained had 10 t [ha.sup.-1] of lucerne hay incorporated. Nevertheless, the other values were also greater than the 0.2-0.3 t C[ha.sup.-1] [year.sup.-1] presented by Sanderman et al. (2010) with improved management of croplands. The greatest increases in OC concentration over the unmodified control (generally ~2-fold) occurred in the 10-30 cm layer; additionally, OC was more evenly distributed down the profile with the clay-modification treatments. Root growth would be limited below 30 cm, which explains why clay modification had little effect on OC concentration in (he 30-50 cm layer. The positive correlation between clay and OC concentration at 10-30 cm depth is in agreement with other studies on unmodified soils (Dalai and Mayer 1986; McKenzie 2008) and confirms that binding of OC to clay can increase OC sequestration by reducing its accessibility (Baldock and Skjemstad 2000).
Effect of clay modification method, depth and site on clay and OC concentration
The effect of modification method on clay concentration differed between the SE and EP sites, which can be explained by depth of the clay-rich B2 horizon (Table 1) and depth of incorporation after clay spreading. The B2 horizon was within the 30-50 cm layer, explaining the high clay concentrations at this depth in all modification treatments execpt clay spread at the EP site where the B2 horizon was below 50 cm (Table 1). The incorporation of clay after spreading was deeper at the SE site (15 cm) than the EP site (8 cm), resulting in a more uniform clay distribution (Tables 5, 6). Clay concentrations in clay spread and delved treatments were generally higher at the SE site than the EP site although the clay concentration of the B2 horizon was higher at the EP site (46% v. 38%). This may be due to differences in machinery design, with the delving machinery used at the SE site being more efficient in bringing up clay, or clay-spreading machinery applying higher rates of clay to the surface.
Organic C concentrations and stocks were generally lower at the EP than the SE site, which reflects the difference in rainfall and fertility of the sites. At the SE site, all modifications resulted in similar OC stock increases over the unmodified control. This may be due to more uniform clay distribution with depth and higher clay concentrations than at the EP sites, where the increase in OC stocks varied among modification treatments. Clay spreading al the EP site did not increase OC stocks over the unmodified control, likely because of low clay concentration below 10 cm. By contrast, spading nearly tripled OC stocks and induced the greatest rate of annual increase. This can be explained by the high clay concentration at 0-30 cm depth and the additional input of OC with the deep incorporation of organic matter at the time of modification (10 t [ha.sup.-1] of lucerne hay).
The effect of the modification method on OC stocks was sites-specific. More important for an increase in OC stocks than the modification method appear to be the final concentration of clay and its uniform incorporation throughout the top 30 cm of soil. Nevertheless, there was a maximum increase in stock of 141 ha 1 (SE) and 22.5 t [ha.sup.-1] (EP) over the unmodified controls.
The addition of subsoil clay to sand resulted in a heterogeneous environment with various-sized clay clods within the sand matrix. Variability in OC concentration as measured by the coefficient of variation was 15-78% for SE and 14-48% for EP, which is higher than the 30-40% reported by Singh et al. (2001) and the 15-35% by Sluikla et al. (2004) for unmodified soils. With the high coefficient of variation in clay-amended soils, sampling methodology must ensure collection of an adequate number of samples by the correct methodology to have confidence in the results.
However, soils with the lowest OC concentrations required the greatest number of samples for reliable estimation of OC concentration. By using different precision values developed in this study, a balance can be reached between cost of sampling and rigour. It would be adequate to collect 10 samples within a 25-m grid for clay-modified soils, but 15 samples would be required for unmodified sands because of their lower OC concentration. Collecting additional samples (e.g. 15 instead of 10) will increase the precision, but at these low OC concentrations (<0.5%), the extra time and analytical costs outweigh the small increase in precision gained. Therefore, we propose 10 samples in a 25-m grid to be sufficient to represent the variability of OC concentrations in clay-modified soils and unmodified controls, which is in accordance with previous recommendations for unmodified soils (Sanderman et al. 2011; Wilson et al. 2010).
The optimum number for BD samples was between one and five within a 25-m grid (Table 3). The higher sample number was due to high variance at the SE site (coefficient of variation 1-19%) compared with the EP site (1-7%). The higher variability at SE site may be caused by the lower water content at sampling (<4% compared with 5-15% at the EP site). Low water content reduces soil cohesion, therefore increasing the likelihood of loss of soil from cores during sampling, which leads to higher variability than with moister soil. Wetting the soil before sampling may reduce variability. When the SE treatments were re-sampled the following year, the dry soil was wetted before BD sampling and BD variation was considerably lower (data not shown). With this adjustment to BD sampling technique, the optimum sample number for bulk density is two or three within a 25-m grid.
We recommend stratified sampling of delved sites based on the proportion of area covered by the delve lines and between delve lines because this sampling method required the lowest number of samples to obtain reliable estimations of OC. The proportion of the area covered by the delve-line and off-delve areas will vary depending on the type of machinery used. On our sites, ~70% of the area was off-delve and 30% was on the delve line; therefore, for OC sampling, seven samples should be collected off-delve and three on the delve line, and for BD, two off-delve and one on the delve line.
This study demonstrated that clay modification could increase OC storage in sands. However, this increase varied with site and modification method, which can be partly explained by differences in clay concentration, because there was a positive correlation between OC and clay concentrations. Ideally, high clay concentrations should be achieved uniformly to a depth of at least 30 cm. It was determined that 10 samples for OC concentration and three samples for BD are required for reliable estimation in clay-modified and unmodified soils in a 25-m grid following the proposed rules of precision and practicality. Further research is required to optimise the effect of clay modification on plant growth and OC storage, for example determining the most appropriate method of clay modification for a given soil type, timing of clay modification, technique of incorporation, addition of organic matter, as well as improving equipment design.
This work was funded by the Department of Agriculture, Fisheries and Forestry, Australia, as part of the Caring For Country Project GMX-OC12-00240, the Agriculture Bureau of SA Inc., the Department of Environment, Water and Natural Resources and Rural Solutions SA. Landholders Terry Young and Roger Groocock are sincerely thanked for providing the locations, knowledge and background information for the selected sites. Drs Cameron Grant, Gordon Churchman and David Chittleborough are thanked for their supervision and guidance. Peter Ciganovic, Rural Solutions SA is thanked for processing of the soil samples.
Bailey G, Hughes B (2012) An observational study of clay delving and its impact on the A2 horizon in sand over clay soils. In 'Soil solutions for diverse landscapes. Proceedings 5th Joint SSA and NZSSS Soil Science Conference', 2-7 December 2012, Hobart, Tas. (F.ds L Burkitt, L Sparrow) pp. 207-210. (Australian Society of Soil Science: Warragul, Vic.)
Bailey G, Hughes B, Tonkin R, Dowie R, Watkins N (2010) Gross soil modification of duplex soils through delving and spading. In 'Soil Solutions for a changing world. Proceedings 19th World Congress of Soil Science', 1-6 August 2010, Brisbane, Qld. (lid. R Gilkes, N. Prakongkep) pp. 26-29. (International Society of Soil Sciences)
Baldock JA. Skjemstad JO (1999) Soil organic carbon/soil organic matter. In 'Soil analysis--an interpretation manual", (lids K Peverill, D Reuter, L Sparrow) pp. 159-170. (CSIRO Publishing: Melbourne, Vic.)
Baldock JA, Skjemstad JO (2000) Role of the soil matrix and minerals in protecting natural organic materials against biological attack. Organic Geochemistry 31, 697-710. doi: 10.1016/S0146-6380(00)00049-8
Betti G, Grant C, Churchman G, Murray R (2015) Increased profile wettability in texture-contrast soils from clay delving: case studies in South Australia. Soil Research 53, 125-136.
Cann MA (2000) Clay spreading on water repellent sands in the south cast of South Australia--promoting sustainable agriculture. Journal of Hydrology 231-232, 333-341. doi: 10.1016/S0022-1694(00)00205-5
Carter DJ, Hetherington RE (1994) Claying of water repellent sands in the Albany district of the south coast of Western Australia. In 'Proceedings 2nd Natural Water Repellency Workshop', 1-5 August 1994, Perth, Western Australia. (Kds DJ Carter. KMWE Howes) pp. 140 144.
Cresswell HP, Hamilton GJ (2002) Bulk density and pore space relations. In 'Soil physical measurement and interpretation for land evaluation'. (Eds N McKenzie, K Coughlan, H Cresswell) pp. 35-58. (CSIRO Publishing: Melbourne. Vic.)
DAFF (2009) 'Soil carbon research program.' (Department of Agriculture, Fisheries and Forestry: Canberra, ACT)
Dalai R, Mayer R (1986) Long term trends in fertility of soils under continuous cultivation and cereal cropping in southern Queensland. II. Total organic carbon and its rate of loss from the soil profile. Australian Journal of Soil Research 24, 281-292. doi: 10.1071/SR986 0281
Davenport D, Hughes B. Davies S, Hall D (201 I) 'Spread, delve, spade, invert: a best practice guide to the addition of clay to sandy soils.' (Grains Research and Development Corporation: Kingston, ACT)
Denef K, Six J, Bossuyt H. Frey SD. Elliott ET, Merckx R, Paustian K (2001a) Influence of dry--wet cycles on the interrelationship between aggregate, particulate organic matter, and microbial community dynamics. Soil Biology & Biochemistry 33. 1599-1611. doi: 10.1016/S00 38-0717(01)00076-1
Denef K, Six J. Paustian K, Merckx R (2001/)) Importance of" maeroaggrcgate dynamics in controlling soil carbon stabilization: short-term effects of physical disturbance induced by dry--wet cycles. Soil Biology & Biochemistry 33, 2145-2153. doi: 10.1016/S0038-0717 (01)00153-5
Desbiolles JM. Fielke JM, Chaplin P (1997) An application of tine configuration to obtain subsoil delving for the management of non-wetting sands. In 'Proceedings Third International Conference on Soil Dynamics (ICSD III)', 3 August 1997, Tiberias, Israel, pp. 201-210. (Faculty of Agricultural Engineering, Technion, Israel Institute of Technology: Haifa)
Ellert BH, Bettany JR (1995) Calculation of organic matter and nutrients stored in soils under contrasting management regimes. Canadian Journal of Soil Science 75, 529-538. doi:10.4141/cjss95-075
Ellert BH, Janzen HH, McConkey BG (2001) Measuring and comparing soil carbon storage. In 'Assessment methods for soil carbon". (Eds R Lal, JL Kimble, JF Follett, BA Stewart) pp. 131-145. (Lewis Publishers: New York)
Fernandez-Ugalde O, Virto I, Barre P, Gartzia-Bengoetxea N. Enrique A. Imaz MJ, Bescansa P (2011) Effect of carbonates on the hierarchical model of aggregation in calcareous semi-arid Mediterranean soils. Geoderma 164, 203-214. doi: 10.1016/j.geoderma.2011.06.008
Hall DJM. McKenzie DC, Macleod DA, Barrett A (1994) Amelioration of a hardsetting alfisol through deep moldboard plowing, gypsum application and double cropping. I. Soil physical and chemical properties. Soil & Tillage Research 28, 253 270. doi: 10.1016/0167-1987(94)90134-1
Hall DJM, Jones HR. Crabtree WL, Daniels TL (2010) Claying and deep ripping can increase crop yields and profits on water repellent sands with marginal fertility in southern Western Australia. Australian Journal of Soil Research 48, 178-187. doi: 10.1071/SR09078
Harper RJ, McKissock I, Gilkes RJ. Carter DJ, Blackwell PS (2000) A multivariate framework for interpreting the effects of soil properties, soil management and landuse on water repellency. Journal of Hydrology 231-232, 371-383. doi: 10.1016/S0022-1694 (00)00209-2
Hatton T, Cork S, Harper P. Joy R. Kanowski P. Mackay R. McKenzie N, Ward T, Wienccke B (2011) Australia state of the environment 2011. An independent report to the Australian Government Minister for Sustainability, Environment, Water, Population and Communities by the State of the Environment 2011 Committee, Canberra, ACT. Isbell RF (2002) 'The Australian soil classification." Revised edn. (CSIRO Publishing: Melbourne, Vic.)
Jenkinson DS (1988) Soil organic matter and its dynamics. In 'Russell's soil conditions and plant growth'. (Ed. A Wild) pp. 564-607. (Harlow: London)
Lal R (2007) Carbon management in agricultural soils. Mitigation and Adaptation Strategies for Global Change 12. 303-322. doi:10.1007/ s11027-006-9036-7
Ma'shum M, Oades J, Tate M (1989) The use of dispersible clays to reduce water repellency of sandy soils. Australian Journal of Soil Research 27, 797-806. doi:10.1071/SR9890797
McKenzie N (2008) 'Guidelines for surveying soil and land resources." Vol. 2. Australian Soil and Land Survey Handbook Series. (CSIRO Publishing: Melbourne. Vic.)
McKenzie N, Ryan P, Fogarty P, Wood J (2000) 'Sampling, measurement and analytical protocols for carbon estimation in soil, litter and coarse woody debris.' (Australian Greenhouse Office: Canberra, ACT)
McKenzie N, Henderson B, McDonald W (2002) Monitoring soil change: principles and practices for Australian conditions. Technical Report No. 18/02. CSIRO Land and Water, Canberra, ACT.
McKissock I, Walker EL, Gilkes RJ. Carter DJ (2000) The influence of clay type on reduction of water repellency by applied clays: a review of some West Australian work. Journal of Hydrology 231-232, 323-332. doi:10.1016/S0022-1694(00)00204-3
Rebbeck M, Lynch C, Hayman PT, Sadras VO (2007) Delving of sandy surfaced soils reduces frost damage in wheat crops. Australian Journal of Agricultural Research 58, 105-112. doi:10.1071/AR06097
Saidy AR, Smernik RJ, Baldoek JA, Kaiser K, Sanderman J, Macdonald LM (2012) Effects of clay mineralogy and hydrous iron oxides on labile organic carbon stabilisation. Geoderma 173-174, 104-110. (In English] doi:10.1016/j.geoderma.2011.12.030
Sanderman J. Farquharson R, Baldoek J (2010) Soil carbon sequestration potential: a review for Australian agriculture. A report prepared for Department of Climate Change and Energy Efficiency, Australian Government. CSIRO Sustainable Agriculture Flagship, Canberra, ACT.
Sanderman J, Baldoek J, Hawke B, Macdonald L, Massis-Puccini A, Szarvas S (2011) National Soil Carbon Research Programme: field and laboratory methodologies. Technical Report. CSIRO Sustainable Agriculture Flagship, Canberra, ACT.
Schipper LA, Sparling GP (2000) Performance of soil condition indicators across taxonomic groups and land uses. Soil Science Society of America Journal 64, 300 311. doi:10.2136/sssaj2000.641300x
Shukla MK, Lai R, Ebinger M (2004) Principal component analysis for predicting corn biomass and grain yields. Soil Science 169, 215-224. doi: 10.1097/01.ss.0000122521.03492.eb
Singh SS, Tiwari SC. Dhar MS (2001) Evaluation of soil degradation using pliysicochemical, biochemical and biological parameters in humid tropics of Arunchal Pradesh, India. Annals of Forestry 9, 287-292.
Six J, Bossuyt H, Degryze S, Denef K. (2004) A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil & Tillage Research 79, 7-31. doi: 10.1016/j.still.2004.03. 008
Skjemstad JO, Janik LJ, Head MJ, McClure SG (1993) High energy ultraviolet photo-oxidation: a novel technique for studying physically protected organic matter in clay- and silt-sized aggregates. Journal of Soil Science 44, 485 499. doi: 10.1111/j.1365-2389.1993.tb00471.x
Tisdall .IM, Oades JM (1982) Organic matter and water-stable aggregates in soils. Journal of Soil Science 33, 141 163. doi:10.1111/j.1365-2389. 1982.tb01755.x
Ward P, Oades J (1993) Effect of clay mineralogy and exchangeable cations on water repellency in clay-amended sandy soils. Australian Journal of Soil Research 31, 351-364. doi: 10.1071/SR9930351
Wilson BR, Barnes P. Kocn TB, Ghosh S, King D (2010) Measurement and estimation of land-use effects on soil carbon and related properties for soil monitoring: a study on a basalt landscape of northern New South Wales, Australia. Australian Journal of Soil Research 48, 421-433. doi: 10.1071/SR09146
Appendix 1. Results of two-way analysis of variance (treatment x depth) of SOC content at two sites Factor South East Eyre Peninsula Probability (P-value) Treatment <0.001 <0.001 Depth <0.001 <0.001 Treatment x depth <0.001 <0.001 ED
Amanda Schapel (A,B,C), David Davenport (B), and Petra Marschner (A)
(A) School of Agriculture, Food and Wine, The University of Adelaide, SA 5005, Australia.
(B) Primary Industries and Regions SA, Rural Solutions SA, SA 5000, Australia.
(C) Corresponding author. Email: email@example.com
Caption: Fig. 1. Example of determination of sample number required for reliable estimation of organic carbon content for the South East site clay spread treatment by plotting variance against increasing sample number (1, 1 + 2, 1 + 2 + 3, until all 20 samples collected in the grid are accounted for). Based on 20 randomised runs.
Caption: Fig. 2. Organic carbon concentrations for clay modification treatments in 0-10, 10-30 and 30- 50cm depths at (a) South East and (b) Eyre Peninsula sites. Capped lines show standard error and means followed by the same letter are not significantly different (P> 0.05) based on the treatment x depth interaction.
Caption: Fig. 3. Relationship between organic carbon concentration and clay concentration for clay-modified sites (Eyre Peninsula and South East) in different depths increments: (a) 0-50cm, (b) 0-10cm, (c) 10-30cm and (d) 30-50cm (B2 clay).
Table 1. Characteristics and management history for each site in the South East (SE) and Eyre Peninsula (EP), South Australia C, Cereal; P. pulse legume; O, oilseed; u.a., not applicable Clay No. of years Depth to Crop rotations modification since clay undisturbed (past 3 years) treatment modification clay ([B.sub.2]) horizon (cm) SB EP SE EP SE EP Control n.a. n.a. 45-60 45-60 C-C-O C-P-C Clay spread 9 14 30-50 60-75 C-P-C C-P-C Delved 9 3 50-70 40-50 C-P-C C-P-C Spaded 4 3 45-65 25-40 C-C-O C-P-C Clay No. of years Estimate of modification since annual organic treatment pasture material incorporated with spading (t [ha.sup.-1) SE EP SE EP Control 6 10 Clay spread 8 4 Delved 8 4 Spaded 6 10 (A) 6 (B) 10 (A) (A) Lucerne hay. (B) Barley stubble. Table 2. Example of sample numbers required from the control and delved treatments for reliable estimation of organic carbon (OC) concentration with 5 20% precision values in a 25-m grid at a = 10% (90% confidence), using Scheffe's formula Sampling from the South East of South Australia. Bold values indicate the optimum sample number following the rules of precision based on OC concentration Soil depth Mean OC Precision (d) (cm) (%) 5% 10% 15% 20% Control 0-10 0.82 37 10 5# 3 10-20 0.29 168 42 19 11# 20-30 0.12 98 25 11 7# 30 50 0.26 230 58 26 15# Delved 0-10 0.94 19 5# 3 2 10-20 0.69 41 11 5# 3 20-30 0.40 146 37 17 10# 30-50 0.25 49 13 6 4# Note: Bold values indicate the optimum sample number following the rules of precision based on OC concentration indicated with #. Table 3. Sample numbers required for reliable estimation of organic carbon concentration and bulk density in a 25-m grid for clay-modified soils at the South East (SE) and Eyre Peninsula (EP) sites Organic C Bulk density SE EP SE EP Control 15 16 4 1 Clay Spread 11 10 5 1 Spaded 11 9 5 1 Delved (stratified) 10 10 2 2 Delved: delve line 11 6 2 2 Delved: off-delve 20 12 2 3 Table 4. Example of sample numbers required for reliable estimation of organic carbon concentration at different depths when using Scheffe's test, the variance method and coefficient of variation, from the control and clay spread treatments, South East of South Australia Scheffe's formula is used at [alpha] =10%, following proposed rules for precision; variance (0.001) and coefficient of variation (5%) arc plotted against sample number Method Soil depth Highest 0-10 10-20 20-30 30-50 value cm cm cm cm Control Scheffe's [alpha] = 10% 5 1 1 7 15 15 Variance 9 11 3 9 11 Coefficient of variation 8 17 13 17 17 Clay spread Scheffe's [alpha] =10% 5 6 11 5 11 Variance 9 13 14 7 13 Coefficient of variation 6 13 17 12 17 Table 5. Mean values, significance (Tukey's lest) and coefficient of variation (CV) for organic carbon concentration and bulk density, and clay content for control, clay spread, delved and spaded sites from the South East of South Australia Within columns, means followed by the same letter are not significantly different (P>0.05) Soil dep Organic C (cm) concentration (n = 20) Mean Tukey's CV (%) significance (%) Unmodified control 0-10 0.82 fg 22 10-20 0.29 bc 50 20-30 0.12 a 38 30-50 0.26 ab 58 Clay spread 0-10 0.99 h 20 10-20 0.69 ef 52 20-30 0.43 cd 78 30-50 0.32 ab 45 Delved 0-10 0.94 gh 15 10-20 0.69 ef 23 20-30 0.4 bcd 46 30-50 0.25 ab 26 Spaded 0-10 0.84 fgh 15 10-20 0.54 de 38 20-30 0.37 bc 43 30-50 0.35 bc 27 Soil dep Bulk density Clay (cm) (n = 3) (%) Mean Tukey's CV (g [cm.sup.3) significance (%) Unmodified control 0-10 1.53 abc 11 3 10-20 1.44 ab 12 2 20-30 1.60 abc 5 1 30-50 1.71 abc 6 19 Clay spread 0-10 1.42 abc 7 14 10-20 1.71 ab 7 12 20-30 1.59 abc 13 20 30-50 1.63 abc 1 42 Delved 0-10 1.70 c 10 10 10-20 1.69 abc 19 7 20-30 1.66 abc 11 8 30-50 1.75 bc 17 29 Spaded 0-10 1.36 a 5 9 10-20 1.46 ab 15 11 20-30 1.61 abc 5 18 30-50 1.56 abc 4 31 Table 6. Mean values, significance (Tukey's test) and coefficient of variation (CV) for organic carbon concentration and bulk density, and clay content, for control, clay spread, delved and spaded sites from the Evre Peninsula, South Australia Within columns, means followed by the same letter are not significantly different (P>0.05) Soil depth Organic C concentration (n-20) (cm) Mean Tukey's CV (%) (%) significance Unmodified control 0-10 0.38 ef 39 10-30 0.18 abc 48 30-50 0.3 cde 29 Clay spread 0-10 0.53 gh 14 10-30 0.15 ab 47 30-50 0.06 a 48 Delved 0-10 0.61 hi 20 10-30 0.24 bed 48 30-50 0.33 de 20 Spaded 0-10cm 0.84 j 18 10-30cm 0.61 i 33 30-50cm 0.46 fg 29 Soil depth Bulk density Clay (cm) (%) Mean Tukey's CV (g [cm.sup.3]) significance (%) Unmodified control 0-10 1.56 ab 6 1 10-30 1.70 bcd 1 1 30-50 1.72 bcd 5 23 Clay spread 0-10 1.51 a 1 9 10-30 1.76 de 2 2 30-50 1.73 bcde 2 1 Delved 0-10 1.68 abed 2 5 10-30 1.70 bed 7 3 30-50 1.92 e 5 38 Spaded 0-10cm 1.57 abc 2 11 10-30cm 1.75 cde 2 18 30-50cm 1.68 abed 1 38 Table 7. Annual change in organic carbon (OC) stock since clay modification for samples collected in the South East and Eyre Peninsula of South Australia Increase in stock is the difference between the modified treatment and the unmodified control, n.a. Not applicable OC stock Increase No. of (0-30cm) in OC years ([[LAMBDA]]) stock since since modification modification (t[ha.sup.-1]) South East Unmodified control 19.8 n.a. n.a. Clay spread 33.5 13.7 9 Delved 34.0 14.3 9 Spaded 27.3 7.5 4 Eyre Peninsula Unmodified control 12.0 n.a. n.a. Clay spread 13.3 1.3 14 Delved 18.3 6.3 3 Spaded 34.5 22.5 3 OC stock change (t[ha.sup.-1]) [year.sup.-1]) South East Unmodified control n.a. Clay spread 1.5 Delved 1.6 Spaded 1.9 Eyre Peninsula Unmodified control n.a. Clay spread 0.1 Delved 2.1 Spaded 7.5 ([LAMBDA]) In soil mass of 5000 Mg [ha.sup.-1].
|Printer friendly Cite/link Email Feedback|
|Author:||Schapel, Amanda; Davenport, David; Marschner, Petra|
|Date:||Mar 1, 2017|
|Previous Article:||Effects of tillage on the soil water retention curve during a fallow period of a semiarid dryland.|
|Next Article:||Effects of land use and topography on spatial variety of soil organic carbon density in a hilly, subtropical catchment of China.|