The agronomic effectiveness of reactive phosphate rocks 3. A comparison of application strategies for soluble phosphorus and reactive phosphate rock fertilisers

Summary. An investigation into the effectiveness of large (up to 80 kg P/ha), single dressings (capital applications) of different phosphorus (P) fertilisers, compared with smaller annual applications, was undertaken in the National Reactive Phosphate Rock Project. Yield comparisons were made at 23 permanent pasture sites across Australia using triple superphosphate, the highly reactive North Carolina phosphate rock and a partially acidulated form of North Carolina rock. Over 4 years, 19 of 23 sites showed no significant reduction in mean annual pasture yield with a single, large dressing applied in year 1 only, compared with an equivalent amount of total P applied in 4 annual applications. At a site in North Queensland, where the initial soil P level was very low, the large year-1 application of each fertiliser resulted in increased pasture production over that obtained from smaller annual applications because of increased pasture production in the early years. This amounted to an annual increase of between 1500 and 3000 kg dry matter/ha for the capital application strategy. The effectiveness of capital applications depended on soil, pasture and climatic conditions, and on the type of P fertiliser. Soil and environmental factors which appear to be important in determining the effect of capital applications are whether soils allow leaching of P (a function of rainfall and texture), whether they are P-sorbing (a function of clay content and soil mineralogy), the soil P content and how quickly it is being utilised by the pasture. Capital applications can be considered where P is not leached from the soil profile or where P sorption is low, and are most effective where soil P is low and there is a responsive pasture species present. Capital applications of water-soluble P fertiliser should not be considered on free-draining, low P-absorbing soils. Average annual pasture dry matter losses of about 2000 kg/ha occurred with a capital application of triple superphosphate compared with annual applications, at one such site in Tasmania. North Carolina phosphate rock was found to be the most effective P fertiliser for large capital applications, especially on free-draining, low P-absorbing soils.

An economic analysis of the field performance of North Carolina reactive phosphate rock compared with single superphosphate for selected sites from the National Reactive Phosphate Rock Project

Summary. An economic analysis was undertaken using pasture yield data from 8 selected sites from the National Reactive Phosphate Rock Project, that encompassed 7 different performance scenarios for North Carolina phosphate rock. The aims were to determine whether the use of North Carolina phosphate rock in place of single superphosphate might result in a positive financial benefit in the 4th year, and after 4 years of annual applications of fertiliser. The analysis was carried out using annual P applications of North Carolina phosphate rock and single superphosphate, that resulted in pasture yields equivalent to 50, 70 or 90% of the maximum yield response of single superphosphate in the 4th year. Annual pasture dry matter yields, produced by these fertiliser applications, were converted to stocking rates, and dollar incomes were derived by applying appropriate gross margins. The analysis was also undertaken to determine the financial benefit from large, year-1 applications of North Carolina phosphate rock. Single superphosphate was priced at $168/t while North Carolina phosphate rock plus sulfur was priced at $180/t. The economic analysis found that a positive financial benefit with North Carolina phosphate rock occurred for only one scenario where the agronomic performance of North Carolina phosphate rock and single superphosphate were equivalent in the 4th year of annual fertiliser application. For 3 other scenarios where the performance of North Carolina phosphate rock and single superphosphate were also equivalent in the 4th year, the economic performance of North Carolina phosphate rock was poor due to (i) a ‘lag’ effect where pasture yield with North Carolina phosphate rock was generally less than that with single superphosphate in years 1–3, and/or to (ii) a seasonal effect where the autumn–early winter pasture responses with North Carolina phosphate rock were less than those with single superphosphate, necessitating a reduction in annual stocking rates on the North Carolina phosphate rock-fertilised pasture. A key finding was that large, single, year-1 applications of North Carolina phosphate rock generally overcame these seasonal and/or yearly lag effects, and led to positive financial benefits from North Carolina phosphate rock applied in this way. Annual applications of North Carolina phosphate rock were economically viable at sandy, high rainfall sites where water-soluble P from single superphosphate would be readily leached from the root zone, provided that the soil P status was adequate and conditions were conducive to reactive phosphate rock dissolution. However, North Carolina phosphate rock was not an economically viable fertiliser to apply annually at (i) non-leaching sites where the soil P status was low and marked yield penalties occurred in the first few years of North Carolina phosphate rock use, (ii) where the soil had a very high P-sorption capacity, and (iii) where North Carolina phosphate rock dissolution was restricted by low rainfall or high pH. The inclusion of estimated residual P value had only a small impact on the economic outcome for scenarios which had not already performed poorly due to seasonal or lag effects.

National Reactive Phosphate Rock Project —aims, experimental approach and site characteristics

Summary. Field-based cutting trials, which formed part of the National Reactive Phosphate Rock Project, were established across Australia in a range of environments to evaluate the agronomic effectiveness of 5 phosphate rocks, and 1 partially acidulated phosphate rock, relative to either single superphosphate or triple superphosphate. The phosphate rocks differed in reactivity. Sechura (Bayovar) and North Carolina phosphate rocks were highly reactive (>70% solubility in 2% formic acid), whilst Khouribja (Moroccan) and Hamrawein (Egypt) phosphate rock were moderately reactive. Duchess phosphate rock from Queensland was relatively unreactive (<45% solubility in 2% formic acid). Phosphate rock effectiveness was assessed by measuring pasture production over a range of phosphorus levels, and by monitoring bicarbonate-soluble phosphorus extracted from soil samples collected before the start of each growing season. Other treatments included single large applications of triple superphosphate, partially acidulated phosphate rock and North Carolina phosphate rock applied at 2 rates, and the application of monocalcium phosphate and North Carolina phosphate rock sources without sulfur to evaluate the importance of sulfur in the potential use of phosphate rock fertilisers at each site. A broad range of environments were represented over the 30 sites which were based on pastures using annual and/or perennial legumes and perennial grasses. Rainfall across the network of sites ranged from 560 to 4320 mm, soil pH (CaCl2) from 4.0 to 5.1, and Colwell-extractable phosphorus ranged from 3 to 47 µg/g before fertiliser application. Two core experiments were established at each site. The first measured the effects of phosphate rock reactivity on agronomic effectiveness, while the second measured the effects of the degree of water solubility of the phosphorus source on agronomic effectiveness. The National Reactive Phosphate Rock Project trials gave the opportunity to confirm the suitability of accepted procedures to model fertiliser response and to develop new approaches for comparing different fertiliser responses. The Project also provided the framework for subsidiary studies such as the effect of fertiliser source on soil phosphorus extractability, cadmium and fluorine concentrations in herbage, evaluation of soil phosphorus tests, and the influence of particle size on phosphate rock effectiveness. The National Reactive Phosphate Rock Project presents a valuable model for a large, Australia-wide, collaborative team approach to an important agricultural issue. The use of standard and consistent experimental methodologies at every site ensured that maximum benefit was obtained from data generated. The aims, rationale and methods used for the experiments across the network are presented and discussed.

Agronomic effectiveness of a partially acidulated reactive phosphate rock fertiliser

Summary. The agronomic effectiveness of a partially acidulated phosphate rock (PAPR), produced by the 50% acidulation of North Carolina phosphate rock with sulfuric acid, was compared over 4 growing seasons with triple superphosphate (TSP) and the highly reactive North Carolina phosphate rock at 22 permanent pasture sites in the National Reactive Phosphate Rock Project. The performance of PAPR as a phosphorus (P) fertiliser for permanent pasture was determined by calculating the substitution value of TSP for PAPR at 50% of the maximum yield response for TSP from the fitted annual dry matter response curves. PAPR performance varied both between sites, and between years at individual sites. Annual yield responses with PAPR were larger than those with TSP at 1 high rainfall site where water-soluble P from TSP was thought to leach from the root zone. PAPR was superior to TSP at another site and generally similar in effectiveness to TSP at 4 sites with light-textured, low or medium P-sorbing soils with a moderate annual rainfall (500–750 mm). The mean substitution value over the 4 years for these sites was >0.9. PAPR performance at other sites where highly reactive phosphate rocks were effective in the short or medium term was variable: there were equivalent yield responses to TSP in some years but much smaller yield responses in other years. PAPR performed very poorly in a third group of sites where the soil had a high to very high P sorption capacity or where there was a very high demand for fertiliser P due to large legume responses on a P-deficient soil. Although generally inferior to TSP, the PAPR was more effective than North Carolina phosphate rock at the majority of sites during the 4-year study.

Effect of fertiliser type on cadmium and fluorine concentrations in clover herbage

Summary. This study investigated whether changing phosphatic fertiliser type affects the accumulation of cadmium (Cd) and fluorine (F) in pasture herbage. North Carolina phosphate rock, and partially acidulated fertilisers derived from this rock, generally have higher Cd and F concentrations compared with single superphosphate currently manufactured in Australia. Clover herbage from sites of the National Reactive Phosphate Rock trial was collected and analysed for concentrations of Cd (11 sites) and F (4 sites). A comparison was made between pastures fertilised with 4 rates of single superphosphate, North Carolina phosphate rock, and partially acidulated phosphate rock having Cd concentrations of 283, 481 and 420 mg/kg P respectively, and F concentrations of 170, 271 and 274 g/kg P respectively. One site used Hamrawein (Egypt) phosphate rock (HRP) having a Cd and F concentration of 78 mg Cd/kg P and 256 g F/kg P respectively. To help identify differences in herbage Cd concentrations between sites, unfertilised soils from each site were analysed for total and extractable Cd contents. At one site Cd concentrations in bulk herbage (clover, grasses and weeds) were related to infestation of the pasture by capeweed (Arctotheca calendula L. Levyns). There were no significant differences between F in herbage from plots fertilised with either single superphosphate, partially acidulated phosphate rock or North Carolina phosphate rock, or between sites. Concentrations of F in herbage were low, generally less than 10 mg/kg. However, there were large differences in Cd concentrations in herbage between sites, while differences between fertiliser treatments were small in comparison. The site differences were only weakly related to total or extractable (0.01 mol CaCl2/L) Cd concentrations in soil. Significant differences in Cd concentrations in clover due to fertiliser type were found at 5 sites. North Carolina phosphate rock treatments had significantly higher Cd concentrations in clover compared with single superphosphate at 2 sites. Partially acidulated phosphate rock treatments had significantly higher Cd concentrations in clover compared with single superphosphate at 4 sites. At the site where Hamrawein was tested, this treatment had significantly lower Cd concentrations in clover compared with both single superphosphate and North Carolina phosphate rock treatments.

The agronomic effectiveness of reactive phosphate rocks 5. The effect of particle size of a moderately reactive phosphate rock

Summary. The effect of particle size on the agronomic effectiveness of a moderately reactive phosphate rock (from the Hamrawein deposit in Egypt) was investigated at 4 of the field sites in the National Reactive Phosphate Rock Project. The aim of these experiments was to determine whether the agronomic effectiveness of this fertiliser was increased by removing particles greater than 0.5 mm in diameter (these coarser particles constituted 28% of the fertiliser). The more reactive North Carolina phosphate rock was used as a reference fertiliser. Soils from 3 of the field sites were also used for glasshouse experiments. The forms of Hamrawein phosphate rock used in these experiments were the less than 0.50 mm particle size fraction (as was used in the field experiments), the 0.075–0.15 mm fraction, and the less than 0.50 mm particle size fraction finely ground. Subterranean clover was used as the test plant and North Carolina phosphate rock and monocalcium phosphate were used as reference fertilisers. A series of laboratory experiments were also undertaken to investigate the effect of particle size. The agronomic effectiveness of Hamrawein phosphate rock was not affected by the different particle size treatments used in the field and glasshouse experiments. The Hamrawein phosphate rock was about equal in effectiveness to North Carolina phosphate rock at the high rainfall leaching site that had acidic sandy soil, and was less effective than North Carolina phosphate rock at the other 2 field sites. The Hamrawein phosphate rock was similar in effectiveness to North Carolina phosphate rock and monocalcium phosphate in the glasshouse experiments with the 2 low P-sorbing soils and was much less effective than North Carolina phosphate rock and monocalcium phosphate in the high P-sorbing soil. Laboratory studies found that there were minimal differences in the amount and rate of dissolution of different particle size fractions of the Hamrawein phosphate rock in moist incubated soil. Free carbonate, most likely in the form of dolomite, was found in all particle size fractions of the Hamrawein phosphate rock and was particularly concentrated in the particle size fractions less than 0.15 mm. This fine dolomite fraction may have suppressed the extent that Hamrawein phosphate rock dissolves in soil, and may have negated any potential beneficial effect that reducing particle size by sieving or by grinding may have had on the agronomic effectiveness of Hamrawein phosphate rock in the field and glasshouse experiments.

The effect of phosphate buffering capacity and other soil properties on North Carolina phosphate rock dissolution, availability of dissolved phosphorus and relative agronomic effectiveness

Summary. The dissolution of North Carolina phosphate rock (NCPR) in soil was investigated in a laboratory study using surface soils sampled from 28 permanent pasture sites. The relationships between phosphorus (P) dissolved, P availability and various soil properties were investigated using simple and multiple linear regression and the findings related to the relative effectiveness of NCPR for pasture production at the sites. The extent of dissolution of NCPR was positively correlated to P buffering capacity (r2 = 0.42). Phosphorus buffering capacity and titratable acidity together accounted for 72% of the variance in dissolution. Bicarbonate-extractable P (‘available’ P) generally increased as dissolution increased. However, the increase in available P was consistently lower for soils with higher P buffering capacity. The proportion of dissolved P that was available also decreased with increasing P buffering capacity (r2 = 0.63). Consequently, the increase in available P was highest for soils with very low to low P buffering capacity. This suggests that the effectiveness of NCPR as a fertiliser may be more closely related to the availability of dissolved P, than to the amount of NCPR dissolved in a soil. Consistent with this laboratory finding, the agronomic effectiveness of NCPR relative to superphosphate measured in the field tended to decrease with increasing P buffering capacity. The agronomic effectiveness of NCPR was comparable with superphosphate only at certain sites, and with some noted exceptions, most of these had surface soils with very low to low P buffering capacity. The high relative effectiveness of NCPR at these sites was mostly attributed to the loss of superphosphate by leaching. Since NCPR dissolves much more slowly than superphosphate, only a small amount of the P applied as NCPR would be lost during leaching events. Slow dissolution of the remaining NCPR probably supplied a small amount of dissolved P over an extended period of time, and due to the low P buffering capacity, much of this was available to plants.

Uptake of phosphorus from different sources by Lotus pedunculatus and three genotypes of Trifolium repens .2. Forms of phosphate utilised and acidification of the rhizosphere

Improvements in the ability of legumes to utilise applied phosphate (P) could lower fertiliser requirements. Earlier studies have found differences between legume species and genotypes of Trifolium repens in their ability to take up P from P sources that vary in solubility. Acidification of the rhizosphere may account for different P utilisation, especially from reactive phosphate rocks. To investigate differences in the ability of legumes to acidify their rhizosphere and extract various forms of P, Lotus pedunculatus and 3 genotypes of white clover that varied in their ability to respond to added P were grown in root study containers with either unfertilised soil, or soil fertilised with either monocalcium phosphate (MCP) or North Carolina phosphate rock (NCPR). After 6 weeks of plant growth, the containers were dismantled and 10 thin soil slices were taken 0-8 mm from the rhizoplane. These slices were fractionated for various P forms and pH was measured. All 4 legumes, especially lotus, acidified their rhizosphere. This resulted in negligible depletion of the small, H2SO4-P-i fraction (Pi, inorganic phosphorus) of unfertilised and MCP-fertilised soil, but caused direct dissolution of NCPR. The predominant forms of P depleted in the MCP-fertilised soil were the resin-Pi and NaOH-Pi forms, whereas in the NCPR treatment the H2SO4-Pi fraction was depleted. No significant depletion of any P fraction was observed in the unfertilised soils as plants were very small and P change in the soil slices was below that which could be measured using the experimental technique. Suggestions are made on how the experimental technique could be improved.

An isotope injection technique to measure agronomic effectiveness of different phosphate sources in undisturbed soils

This paper describes the use of an isotope injection technique for assessing the agronomic effectiveness of phosphate sources in undisturbed soils in field and glasshouse experiments. A special injection apparatus consisting of 20 syringe needles linked to a common reservoir was used to label soil cores 150 mm in diameter and 150 mm in height, without significantly disturbing the soil structure and established perennial ryegrass (Lolium perenne) root patterns. Variations in the amount of phosphorus (P) taken up by plants from single superphosphate (applied at 30, 60, and 100 kg P/ha), from 30% acidulated North Carolina phosphate rock (60 g P/ha), and from 'as received' North Carolina phosphate rock (60 kg P/ha) were indicated by the specific activity of the plant P. Percentages of plant P derived from the fertilisers were calculated on the basis of the specific activity of the fertilised treatments and the controls. The technique provides a simple method of comparing the agronomic effectiveness of phosphate sources in undisturbed soils and, unlike total dry matter yield and P uptake, is relatively unaffected by environmental factors.