What actually is RPR?
The expression ‘RPR’, or reactive phosphate rock describes a phosphate rock that – with certain provisos like soil pH of 6 or less, and an annual minimum of 800mm rainfall and/or irrigation – will adequately supply the maintenance P requirements of (in New Zealand’s case) a vigorous pasture. Given the knowledge that exists in 2010, to state that ‘not all RPRs are the same’, as some self-appointed experts do, could be interpreted by as a deliberate attempt to confuse farmers.
Various laboratory tests have been developed to predict whether a particular phosphate rock is soluble or ‘reactive’ enough to maintain vigorous pasture growth. The intention of these tests is to help avoid avoid the lengthy time, trouble and very high costs of doing scientific, replicated field trials. Solubility in citric acid is one. If the solubility of the P in a phosphate rock in the citric acid laboratory test is at least 30% of the total, such a product is labeled an ‘RPR’, and it is highly likely that it will maintain a vigorous pasture, at least from the third year onwards. In the first 2 years, the use of a blend of soluble P and RPR is usually recommended.
RPRs were all formed on the sea floor originally, by the gradual absorption of phosphate present in sea water into dead sea organisms, which are largely calcium carbonate. The level of carbonate remaining in the crystal lattice largely determines the solubility of the product. The deposits that are economic to mine are those that are now above sea level, due to earthquakes or sea-level changes.
Some countries in which RPR exists contain phosphate rock deposits varying in reactivity from medium to very high. Egypt is one example. Others contain only one, such as North Carolina, USA. In other cases, the same deposit is spread over what are now two or more countries. The Tunisian and Algerian deposits are essentially one and the same, but with the cadmium (Cd) content rising from very low in Algeria(10 ppm) to very high (80ppm) at the eastern extremes in Tunisia.
Many deposits consist of more than one layer, often with clay, limestone or dolomite in between, that were laid down at different geological periods. The Sechura deposit contains one layer that is far less ‘reactive’ than the others, and there is also a wide variation in Cd levels.
North Carolina is no longer sold as an RPR for direct application, partly because its cadmium level of 48ppm is considered too high by most fertiliser retailers, and partly because its owners regard using to produce phosphoric acid (from which most of the Cd can be stripped) to be more profitable. The phosphoric acid is used to make DAP, MAP and TSP.
Risks in relying solely on laboratory tests to assess ‘reactivity’
A problem with the citric acid test is that you can generate an artificially high test if you grind some semi-reactive phosphate rocks very finely…
A problem with the citric acid test is that you can generate a ‘satisfactory’ test if you grind some semi-reactive phosphate rocks (like Moroccan BG4) very finely, or test just the fine fraction, or blend them with a highly citric-soluble RPR such as Sechura . Sechura has, for chemical reasons (the presence of hydroxide in the crystal lattice), an atypically high solubility in citric and the similar formic acid tests. While it is unquestionably an effective RPR, it generally performs no better in field trials than the other RPRs such as Tunisian, Algerian and North Carolina. One of New Zealand’s largest fertiliser companies was recently accused of doing this.
On the other hand, naturally-occuring free lime or dolomite mixed with a true RPR can artificially reduce the solubility in the test a bit, even though it doesn’t adversely affect its field performance. Some true RPRs from Egypt can show this effect, as the deposits were laid down with dolomite.
So, neither citric acid or formic acid solubility tests are perfect!
The unbeatable test; the a-axis measurement
By far the single best test to define whether an RPR is actually an RPR is the x-ray diffraction measurement of the a-axis of the crystal lattice. This can be done quite easily with modern x-ray diffraction equipment , but the equipment is very expensive. They have one at the renown International Fertiliser Development Centre in Alabama, USA, visited many times by the author. This instrument demonstrated that the Algerian, Tunisian and North Carolina RPRs all have identical a-axis dimensions.
The IFDC have published comparisons of plant responses of different RPRs, showing that, provided the a-axis is the same, they perform the same. End of argument.
The IFDC have published comparisons of plant responses of different RPRs, showing that, provided the a-axis is the same, they perform the same. End of argument. The one and only exception to this is the Sechura RPR from Peru. As mentioned before, Sechura RPR is a bit different chemically, because it has hydroxide replacing some carbonate in the crystal lattice. This changes its solubility in citric acid, and its a-axis, but it still performs the same in the field as other RPRs. The smaller the a-axis, the more resistant a phosphate rock is in the soil, to the point where it will not be capable of releasing sufficient P annually to maintain vigorous growth, and is therefore not worthy of the ‘RPR’ title. The IFDC have stated all this, and they know what they are talking about.
It is concerning that a phosphate rock such as Moroccan BG4 can continue for years to be advertised as an RPR, in the complete absence of scientifically robust field trial data, or the independent measurement of the a-axis dimension.
The ‘National Series’ of RPR Trials
During his time (1981-84) as Technical Advisor to the Directorate of MAF’s Agricultural Research Division (which subsequently became the AgResearch Crown Research Institute), and subsequently as Chief Scientist for Soil Fertility at the Ruakura Research Center in Hamilton (1984-87), the author had the responsibility for designing and overseeing the New Zealand-wide series of field trials comparing RPR with superphosphate, otherwise known as the ‘National Series’ of RPR trials.
This series of trials was conducted on 19 sites for 6 years, from 1981-86. A few of the sites continued for a further 2 years, purely looking at residual effects in the absence of fertiliser applications.
Before leaving the MAF in late 1987,the author presented a summary of the trial results to a farmer conference at Ruakura, including a simple predictive model for predicting how RPR would compare to superphosphate in various situations.
Subsequently, the late Dr Allan Sinclair, who had been closely involved with the trials throughout, wrote up the results for publication in scientific journals, with the assistance of several of the other scientists involved.
So, what exactly did this series of trials establish? Here are the key points, all of which assume that the RPR has had blended with it sufficient fine elemental sulphur to meet maintenance S requirements-
Provided that (a) the initial Olsen P in the soil was not well below the recommended range for a high-producing pasture on a given soil, and (b) that the soil pH was not much higher than 6.0, and (c) that the average annual rainfall (or rainfall plus irrigation) was at least 800mm, then RPR would usually equal the performance, or in the worst case come to equal the performance over a 3 year period, of superphosphate.
With the same provisos as above, the average pasture production with RPR was 3% less (range 0-5%) in year 1, 1% less (range 0-3%) in year 2, and no different in year 3 onwards. This ‘lag-effect’ is caused by the time required for annual applications of RPR to build up a sufficiently large P reserve in the soil so that each year, sufficient RPR was being dissolved into plant-available form by soil acidity to provide sufficient plant-available P to maintain the growth of a vigorous ryegrass-clover pasture.
The explanation is very simple. Essentially, any one application of RPR took 2-5 years depending on soil type, soil pH and rainfall (average 3years) for most of it to be dissolved into plant-available form. So, after 3 years of application, you reached an equilibrium situation where a year’s maintenance requirement became available in each year. It’s not rocket science.
Where the initial soil Olsen P was high enough on a given soil to mean that there would have been no response to superphosphate for a year or more if withheld, there was no difference in pasture production between RPR and superphosphate right from the start. This is simply because the reserve of P in the soil avoided any drop in production from the ‘lag-effect’ from RPR.
Where the soil Olsen P was in the low or medium (‘maintenance’) range at the start, any lag-effect in pasture production from RPR could be totally avoided by either of two practices-
(a) Using a 70-30 blend (by P) of RPR and a low gypsum-content form of water-soluble P (this excludes superphosphate) for the frst 2 years, before switching to straight RPR. These blends of P could be either a physical blend of RPR and TSP, DAP or MAP, or what is called a partially-phosphoric acidulated reactive phosphate rock, or PAPR. Blends of RPR and superphosphate do not work as well initially. This is believed to be due to the calcium ‘common-ion’ effect, whereby the Ca dissolving from the soluble gypsum of superphosphate (55% by weight), inhibits the dissolution of the Ca from the RPR, and therefore the release of the P into plant-available form.
(b) Where cash-flow permits, an application of RPR equivalent to 3 years maintenance, in one initial application at the start, and then reverting to maintenance application in year 4 onwards.
All RPRs performed the same. Although only Sechura RPR was included at multiple rates to produce a full ‘response curve’, North Carolina RPR and Chatham Rise phosphorite were included at 0.75x maintenance rates in virtually all the trials. These invariably performed no better, and no worse, than Sechura RPR. Other work has shown than if the RPR is not sufficiently fine, or has too much free lime or dolomite occurring with it, initial production can be lower.
Other trials have shown that if a less reactive phosphate rock is used, most accurately determined by measuring its crystal a-axis dimension, it is likely that it will never be able to maintain the same level of production as a true RPR or superphosphate, because it simply cannot release enough P each year, unless the soil is extremely acid, and soil aluminium toxicity is not present.
There is anecdotal evidence over the long period of time (25 years plus) that some farmers have been using RPR that eventually, maintenance P requirements start to drop significantly (by up to 10 kgP/ha/yr) with RPR. This is likely to be partly because of the scientifically proven reduction of P run-off into waterways when RPR is used, and partly because of reduced fixation onto soil clay particles.
A review of all New Zealand RPR trials has been published in a series of scientific articles and conference papers by Dr Quin, named RPR Revisited (1)-(4), 2013/14.