Comparison of single superphosphate and superphosphate coated with bauxite residue for subterranean clover production on phosphorus-leaching soils.
The coastal sandplains of south-west Western Australia have an average annual rainfall [is greater than] 800 mm where up to 90% falls during the growing season for pasture between April and October. The soils have a coarse, sandy surface of varying depths to impermeable layers of ironstone or clay. The impermeable layers result in periodic inundation for most of the growing season and this results in a combination of subsurface drainage through the topsoil and surface flow when the soil becomes saturated.
It has been necessary to apply dressings of single superphosphate to subterranean clover up to 40 kg phosphorus (P)/ha in the first few years after the clearing of native vegetation (Yeates 1993). Farmers now apply up to 15 kg P/ha to maintain soil P levels. The sandy soils have a low capacity to sorb and retain P and any excess P flows into the waterways. This P is the limiting nutrient for algal growth in the shallow Peel Inlet and Harvey Estuarine System and massive mats of algae have been recurring in the system since 1968 (Cross 1974; Birch 1982; Kinhill Engineers 1988; Hodgkin and Hamilton 1993).
Bauxite residue is a by-product of extraction of alumina from bauxite with caustic soda in the Bayer process. Applying bauxite residue (red mud) has been shown to reduce P leaching from sandy soils in Western Australia (Ward 1983; Summers et al. 1993, 1996a) and increases in plant growth have been attributed to the liming effect and P retention of the bauxite residue (Summers et al. 1996b). However, the expense of haulage limits application of bauxite residue to areas within 30 km of alumina refineries in the Peel-Harvey catchment owing to the large amounts needed to effectively alter the soil characteristics (2-20 t/ha).
To reduce leaching from single superphosphate, Deeley (1989) incorporated bauxite residue into superphosphate in the manufacturing process. This resulted in increased dry matter production of plants relative to single superphosphate when applied to sandy soil in pots under leaching conditions.
To achieve a reduction in leaching it has been proposed that coating superphosphate granules with bauxite residue would achieve similar yield improvements relative to ordinary single superphosphate. This would enable bauxite residue to be used to retain P with less transport cost, because only relatively small amounts of bauxite residue would be needed. Coating the granules as a separate process would enable a range of granulated fertilisers to be modified in this manner for a limited proportion of the market without disrupting the original manufacturing process. The coating process could take place at the point of sale during the blending of fertilisers for customer requirements.
The aim of the experiment was to determine if bauxite-residue-coated superphosphate improved the effectiveness relative to ordinary single superphosphate. It was also aimed to determine the optimum proportion of bauxite residue to coat the superphosphate granule and compare this with bauxite-residue-amended soil, fertilised with single superphosphate. A field experiment was used to relate the effectiveness of the coated superphosphate to ordinary single superphosphate and determine if coating altered the amount of P residues in the soil after annual rainfall.
Materials and methods
Polyvinyl chloride columns (153 mm in diameter and 600 mm deep) were packed with grey Joel sand, Bassendean series (McArthur and Bettenay 1974) from Pinjarra, Western Australia. The soil was coarse, acidic, had a low capacity to retain nutrients (Bettenay et al. 1960), and was classified by the Northcote system as Uc2.33 (US Soil Taxonomic: Aquic Xeropsamment). The soil was put through a 2-mm sieve and about 15 kg was packed to a depth of about 50 mm. The columns were pre-leached with 1500 mm (about 7.5 pore volumes or the equivalent of 1.5 years rainfall) of water and then sown with clover (Trifolium subterraneum cv. Trikkala, 15 seeds thinned to 8 plants). The seed and fertiliser were applied after the pots had been brought to field capacity. The pots were watered with 50 mm of water, 3 times a week for the duration of the 8-week trial. This was a total of 1200 mm total rainfall where 1000 mm is the average rainfall in the coastal south-west of Western Australia and the 200 mm extra served to displace the remaining pore volume. The pots were randomised in replicate blocks alter each watering. A complete, liquid basal fertiliser (excluding P) was applied twice during the experiment, at the beginning and half-way through the trial.
Phosphorus was applied as single superphosphate coated with a range of thicknesses of bauxite residue. The proportion of bauxite residue to superphosphate was 0, 5, 10, 15, 20, 25, 30, 35, and 40% by weight. Each of these fertilisers was applied at rates equivalent to 20 and 40 kg P/ha. The coating of the superphosphate was achieved during mixing with dry bauxite residue in a cement mixer and applying a fine spray of water to promote adhesion. Less than 10% water by weight was added, after which the pellets were air-dried. The bauxite-residue-coated superphosphate was split to near the final weight with a riffle splitter to produce a representative particle-size distribution. Two extra treatments of 20 t/ha of bauxite residue (a rate that has been used in previous soil amendment studies; Summers et al. 1996b) plus 20 kg P/ha as ordinary single superphosphate and 20 t/ha of bauxite residue plus 40 kg P/ha as ordinary single superphosphate were added. The treatments included: 9 ratios of bauxite residue to superphosphate plus one treatment of 20 t/ha bauxite residue as a soil amendment by 2 rates of superphosphate by 3 replicates = (9+1)x2x3: 60 pots.
The clover tops were harvested at the end of the trial, dried, weighed, and analysed for P content after digestion in sulfuric acid and hydrogen peroxide (Yuen and Pollard 1954). The P concentration in the digest was measured by the molybdovanadium-phosphate method (AOAC 1975).
The volume of leachate from the columns was measured before each irrigation and the leachates from 3 replicates were mixed together (bulked) for P, Mn, Zn, Cu, Fe, S, Mg, Na, Ca, and K concentration determination using inductively coupled plasma (ICP) mass spectrometry on unfiltered samples. The pH of the soil in the bauxite residue treatments was measured in a 1:5 ratio of soil to water and a 1:5 soil to 0.01 M Ca[Cl.sub.2] solution. The pH was measured in 20-mm increments down the column to 100 mm depth, and at intervals of 100-150, 150-200, 200-300, 300-400, and 400-500 mm.
The field experiment was conducted in Denmark, Western Australia (Table 1), on humic sandy podzol soils which are classified as Uc2.21 (Northcote 1979), a haplohumod (Soil Survey Staff 1975), or a humic aquic podosol (Isbell 1996). The site is 365 km south-west of Perth, and receives an annual average rainfall of 1000 mm including 83% or 830 mm during the growing season (April-October).
Table 1. Soil properties of the field site at Denmark, Western Australia
<2 mm fraction of soil pH (1:5 soil: 0.01 M Ca[Cl.sub.2]) 4.3 Bicarbonate-extractable P (Colwell 1965) ([micro]g/g soil) 2 % Clay (<2 [micro]m) 2 % Silt (2-20 [micro]m) 1 % Sand (20-2000 [micro]m) 97
The design incorporated 3 replicates of the treatments with 9 replicates of the nil treatment which received no addition of phosphorus fertiliser. The plots were 4 m by 4 m. The 2 fertiliser treatments were ordinary single superphosphate and single superphosphate coated with 25% bauxite residue by weight. The P was applied at 0, 5, 10, 20, 40, and 80 kg P/ha. The trial was over 3 years where a third of the plots were treated in the first year, a third in the second, and a third in the last year. The number of plots can be calculated by: (3 replicates)x(5 rates of P)x(3 years)x(2 P treatments)+(9 nils) = 99 plots. The site was sown to the clover Trifolium subterraneum var. Trikkala.
The yield of the pasture was measured using a rising plate meter (Earle and McGowan 1979) in spring each year. The meter readings were calibrated with dry matter pasture yield cut from within quadrats. The pasture was cut to ground level and dried at 70 [degrees] C in a forced-draught oven before weighing. Yields were measured on 11 October 1996, 11 November 1996, 14 October 1997, 5 August 1998, 16 September 1998, and 28 October 1998. After each measurement, the pasture was cut with a mower set at 30 mm above the ground to simulate grazing. The cut herbage was removed from the plots. Before mowing on 11 October 1996 and 11 November 1996, samples of pasture tops were collected from random locations within each subplot. The samples from each subplot were bulked, dried, and used to determine the concentration of P in the plant tissue. Samples of the ground, dried plants were digested in sulfuric acid and hydrogen peroxide (Yuen and Pollard 1954). The P concentration in the digest was measured by the molybdovanadium-phosphate method (AOAC 1975).
Soil samples were collected in February 1997 using a metal tube 20 mm in diameter and 100 mm deep. Ten cores were collected from random locations within each subplot and then these cores from each subplot were combined. The soil was analysed for bicarbonate-extractable P using an automated version of the Colwell procedure (Colwell 1965). In addition, the amount of P extracted by sulfuric acid (hereafter called `total P') was also extracted and determined colorimetrically (Murphy and Riley 1962). The soil pH was determined in 1:5 soil :0.01 M Ca[Cl.sub.2].
The P concentrations in the single superphosphate and the bauxite-residue-coated superphosphate are listed in Table 2. The single superphosphate made in Western Australia included granules ranging from 0.5 to 5.0 mm diameter. The coating used in the field trial was 25% by weight and produced by the same method described in the glasshouse experiment above.
Table 2. Phosphorus concentrations (%w/w) in the fertilisers used Values in parentheses are percentages of total P
Single superphosphate Uncoated single coated with superphosphate bauxite residue(C) Total P 9.1 7.14 Water-soluble P(A) 7.3 (80) 2.76 (39) Citrate-soluble P(B) 1.5 (16) 4.13 (58)
(A) Measured by standard AOAC (1975) procedures.
(B) In neutral ammonium citrate, as measured by standard AOAC (1975) procedures.
(C) 25% w/w.
Basal fertilisers (Table 3) were applied in early April and mid-August to ensure that P was the only element limiting yield. Basal and P fertilisers were applied to the surface, the common practice for fertilising pastures in Western Australia.
Table 3. Basal fertilisers applied (kg fertiliser/ha) (kg/ha of element in parentheses)
Element April Compound (%) 1996 Ca[SO.sub.4] 18 S 100 (18 S) 23 Ca (23 Ca) KCl 50 K 100 (25 K) [Na.sub.2][BO.sub.4] 11 B 3 (0.33 B) Mg[SO.sub.4] 10 Mg 50 (5 Mg) Mn[SO.sub.4] 32 Mn 30 (9.6 Mn) Cu[SO.sub.4] 25 Cu 10 (2.5 Cu) ZnO 80 Zn 2 (1.6 Zn) Mo[O.sub.3] 67 Mo 0.2 (0.134 Mn) Element April Mid-August Compound (%) 1997 and 1998 each year Ca[SO.sub.4] 18 S 100 (18 S) 100 (18 S) 23 Ca (23 Ca) (23 Ca) KCl 50 K 100 (25 K) 100 (25 K) [Na.sub.2][BO.sub.4] 11 B Mg[SO.sub.4] 10 Mg Mn[SO.sub.4] 32 Mn Cu[SO.sub.4] 25 Cu ZnO 80 Zn Mo[O.sub.3] 67 Mo
Analysis of data
The data from the glasshouse experiment were tested using an analysis of variance (ANOVA) to detect significant pH differences between treatments using the Fisher's Protected Least Significant Difference (l.s.d.), shown in Fig. 4 as an error bar indicating the 95% confidence level. The data from the leaching, P uptake, and yield were analysed using a linear regression analysis.
[Figure 4 ILLUSTRATION OMITTED]
The relationship between yield and the level of P applied was described by the Mitscherlich equation:
y = a - b exp(-cx)
where y is the yield of dried pasture (kg/ha), x is the level of P applied (kg P/ha), the coefficient a provides an estimate of maximum yield (kg/ha), the coefficient b quantifies the yield response (maximum yield minus the mean yield for the nil fertiliser application, as estimated by the fitted equation) (kg/ha), and coefficient c describes the curvature or shape of the response curve such that as c increases, the response curve moves to the left and less P is required to produce the same yield. Estimates of a, b, and c were obtained by fitting the Mitscherlich equation to the data by generalised nonlinear least squares.
The effectiveness of the P treatment was calculated relative to the effectiveness of freshly applied single superphosphate. This was done using the initial slope (bc) of the fitted Mitscherlich equation. It is not valid to use the curvature coefficient c, because the bauxite residue coating and the ordinary single superphosphate produced different maximum yield plateaux (Barrow and Campbell 1972). Relative effectiveness (RE) values were calculated by dividing the bc value for each P fertiliser added, either that year or in previous years, by the bc value for freshly applied single superphosphate. From this, the RE of freshly applied single superphosphate is always 1.00. The RE was compared within the years using a paired t-test.
Results and discussion
The heavier the bauxite residue coating, the greater the amount of P retained in the soil. The majority of the water-soluble P leached rapidly (Fig. 1) before all the clover had germinated or grown past the cotyledon stage. The superphosphate coated with 20-35% by weight of bauxite residue leached equivalent amounts of P to the 20 t/ha of bauxite residue treatment. Both of these treatments leached only half of the P that leached from the untreated column with ordinary single superphosphate (Fig. 2). From the initial peak of leaching in Fig. 1 it seems that there is a 70% retention of P based on the apparent area under the curves. However, when the areas of the tail of the curves that cannot be clearly seen are taken into account, the sustained leaching from the bauxite-residue-coated superphosphate brings this down to 50%. These results are in agreement with those obtained by Summers et al. (1996a), where columns treated with 20 t/ha of bauxite residue leached half of the P leached from untreated soil. Deeley (1989) mixed bauxite residue and superphosphate as a slurry in a proportion of 33%, and noted that the solubility of P in water was reduced to about 40% of the solubility of ordinary superphosphate.
[Figures 1-2 ILLUSTRATION OMITTED]
Increasing the coating of bauxite residue around the granule increased the P uptake and gave a small but not significant increase in dry matter yield, for the 40 kg P treatments (P [is less than] 0.0001 and P = 0.0927, respectively). As the coating of the bauxite residue increased, the 20 kg P treatments showed a significant (P = 0.0065) increase in P uptake and a slight but not significant (P = 0.387) increase in dry matter yield (Fig. 3). The pH of the top 100 mm of soil was unchanged by the bauxite residue coating of superphosphate. The pH [is less than] 4.0 was probably limiting the growth or nodulation of the clover and may have reduced the potential yield. Treatments of the soil with 20 t/ha of bauxite residue as a soil amendment increased the pH of the top 100 mm (from 3.9 to 6.2 in 0.01 M Ca[Cl.sub.2]). When compared with the untreated soil columns, in increments of 20-50 mm, the bauxite residue increased the pH of the topsoil to a depth of about 60 mm (Fig. 4). Similar changes in topsoil pH have also been found with the use of bauxite residue in field experiments (Summers et al. 1996b).
[Figure 3 ILLUSTRATION OMITTED]
Of the other elements that were leached, only potassium showed any variation. The small amounts of bauxite residue applied in the coatings ([is less than] 160 kg bauxite residue/ha) produced no change in the leaching of potassium. However, the 20 t bauxite residue/ha leached 65% of the potassium of the other treatments (128 mg and 193 mg, respectively). This may have been due to the ability of the zeolite in bauxite residue noted by Wong (1990) to specifically exchange cations of sodium and potassium. The potassium displaces sodium from the binding sites of the zeolite, resulting in greater retention of potassium ions. This has implications for the use of bauxite residue on poor sandy soils as a tool to reduce the leaching of potassium. It may also contribute to the increase in the dry matter production from bauxite residue treated sites. The combination of potash with bauxite residue or another pure zeolite may improve the effectiveness of the fertilisers under leaching conditions.
The responses of pasture herbage yield to P applied were defined by the Mitscherlich equation and the values of the relative effectiveness (RE) from the Mitscherlich equation have been grouped by the time since fertiliser application and presented graphically in Fig. 5. The RE of the bauxite-residue-coated superphosphate was consistently greater than the ordinary single superphosphate and when compared within each time of assessment; this difference was significant (P = 0.0323, paired t-test). The differences in RE varied between assessments during each year and between years. The RE of the bauxite-residue-coated superphosphate increased within each year relative to ordinary single superphosphate as each year progressed. For example, the freshly applied bauxite-residue-coated superphosphate was 6% less effective than ordinary single superphosphate when measured in August 1998, 14% more effective in September 1998, and 42% more effective in October 1998.
[Figure 5 ILLUSTRATION OMITTED]
Variation in RE within and between years has also been observed in a comparison of ordinary single superphosphate and partially acidulated rock phosphate by Bolland et al. (1995). The trend for the RE to increase with successive harvests in the same year was also found by Bolland et al. (1995). It is likely that the water-soluble P in single superphosphate rapidly dissolved and was immediately available for plant uptake in both this experiment and the experiment by Bolland et al. (1995). The glasshouse experiment showed that this soluble component of single superphosphate rapidly leaches into the soil beyond the root-zone. The single superphosphate becomes increasingly less effective with increasing time since application, even in the year of application. In contrast to the single superphosphate examined here and the sources examined by Bolland et al. (1995), the reduction in water solubility created by the bauxite residue coating releases P throughout both the growing season and subsequent seasons.
It is concluded that the bauxite residue coating maintains the supply of P near plant roots in a manner that is more effective than single superphosphate in both the year it is applied and subsequent years. However, there is still further potential for increases in effectiveness as shown by the leaching of P from the bauxite residue in the glasshouse trial, albeit greatly reduced in comparison with the single superphosphate.
The sandy soils on the coastal plains of west and south Western Australia have residues from applied superphosphate in P that is sorbed, present in old superphosphate granules, and P present in organic matter in the soil. The reduced water solubility induced by coating with bauxite residue is likely to contribute to all of these components and greatly reduce the leaching of P from both the freshly applied fertiliser and its residues.
The results show that relative to single superphosphate, a bauxite residue coating of superphosphate reduces P leaching and increases production on sandy soils susceptible to P leaching. The increased effectiveness continued for at least 2 years after application.
The bauxite residue may also be of value in coating other fertilisers. The value of bauxite residue to manipulate potassium leaching is noteworthy and deserves further investigation because potash is an expensive component of pasture production on high rainfall, sandy soils and also leaches rapidly. If potassium is limiting plant growth then the optimum use of P is restricted and may result in growers erroneously applying excessive superphosphate. This may lead to further leaching of P.
The field experiment was conducted at Denmark Agricultural High School where staff provided encouragement and helped in many ways. The Chemistry Centre of Western Australia measured plant and soil samples. Funds were provided by the Western Australian Government. Dr M. D. A. Bolland provided helpful comments.
AOAC (1975) `Official methods of analysis.' 12th edn. (Association of Official Agricultural Chemists: Washington DC)
Barrow NJ, Campbell NA (1972) Methods of measuring residual value of fertilizers. Australian Journal of Experimental Agriculture and Animal Husbandry 12, 502-510.
Bettenay E, McArthur WM, Hingston FJ (1960) The soil associations of part of the Swan Coastal Plain, Western Australia. CSIRO Australian Division of Soil, Land Use Series No. 35.
Birch PB (1982) Phosphorus export from coastal plain drainage into the Peel-Harvey estuarine system of Western Australia. Australian Journal of Marine and Freshwater Research 33, 23-32.
Bolland MDA, Clarke MF, Yeates JS (1995) Effectiveness of rock phosphate, coastal superphosphate and single superphosphate for pasture on deep sandy soils. Fertilizer Research 41, 129-143.
Colwell JD (1965) An automatic procedure for the determination of phosphorus in sodium hydrogen carbonate extract of soil. Chemistry and Industry, 893-895.
Cross WJ (1974) A look at pollution problems of Peel Inlet. Australian Marine Science Association Bulletin 48, 7-9.
Deeley DM (1989) Phosphorus fertilizer dissolution and leaching losses from acid grey sands. MSc Thesis, University of Western Australia.
Earle DF, McGowan AA (1979) Evaluation and calibration of an automated rising plate meter for estimating dry matter yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry 19, 337-343.
Hodgkin EP, Hamilton BH (1993) Fertilizers and eutrophication in southwestern Australia: setting the scene. Fertilizer Research 36, 95-103.
Isbell RF (1996) `The Australian soil classification.' Australian Soil and Land Survey Handbooks. (CSIRO: Australia)
Kinhill Engineers (1988) Peel Inlet and Harvey Estuary management strategy environmental review and management programme Stage 2. Department of Marine and Harbours, Department of Agriculture. Report and recommendations of the Environmental Protection Authority. Part I and II. EPA Bulletin 363, November 1988.
McArthur WM, Bettenay E (1974) The development and distribution of soils of the Swan Coastal Plain, Western Australia. CSIRO Australia Publication No. 16.
Murphy J, Riley JP (1962) A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta 27, 31-36.
Northcote KH (1979) `A factual key for the recognition of Australian soils.' 4th edn. (Rellim Tech. Publishers: Glenside, S. Aust.)
Soil Survey Staff (1975) Soil taxonomy: a basic system of soil classification for making and interpreting soil surveys. United States Department of Agriculture Handbook No. 436. (Govt. Printing Office: Washington DC)
Summers RN, Guise NR, Smirk DD (1993) Bauxite residue (red mud) increases phosphorus retention in sandy soil catchments in Western Australia. Fertilizer Research 34, 85-94.
Summers RN, Guise NR, Smirk DD, Summers KJ (1996b) Bauxite residue (red mud) improves pasture growth on sandy soils. Australian Journal of Soil Research 34, 569-581.
Summers RN, Smirk DD, Karafilis D (1996a) Phosphorus retention and leachates from sandy soil amended with bauxite residue. Australian Journal of Soil Research 34, 555-567.
Ward SC (1983) Growth and fertiliser requirements of annual legumes on a sandy soil amended with fine residue from bauxite refining. Reclamation and Revegetation Research 2, 177-190.
Wong J (1990) Sodium release characteristics and revegetation of fine bauxite refining residue (red mud). PhD Thesis, Murdoch University, Western Australia.
Yeates JS (1993) Soils and fertilizer use in south-western Australia. Fertilizer Research 36, 123-125.
Yuen SH, Pollard AG (1954) Determination of nitrogen in agricultural materials by the Nessler reagent. II. Microdeterminations in plant tissue and soil extracts. Journal of Science Food and Agriculture 5, 364-369.
Manuscript received 25 June 1999, accepted 24 December 1999
Robert Summers, Martin Clarke, Tim Pope, and Tim O'Dea Agriculture Western Australia, PO Box 376, Pinjarra, WA 6208, Australia.
Robert Summers, Corresponding author; email: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Summers, Robert; Clarke, Martin; Pope, Tim; O'Dea, Tim|
|Publication:||Australian Journal of Soil Research|
|Date:||May 1, 2000|
|Previous Article:||A modified chromium-reducible sulfur method for reduced inorganic sulfur: optimum reaction time for acid sulfate soil.|
|Next Article:||Accessibility of subsoil potassium to wheat grown on duplex soils in the south-west of Western Australia.|