Incremental P release in alkaline Vertosols

Chris Guppy, Karl Andersson and Matt Tighe (UNE)

GRDC code

GRS10029

Introduction

Soil testing for available P status commonly involves a Colwell test on the 0-10 cm layer. Recent GRDC funded research has highlighted the role of poorly soluble P forms in alkaline soils (the two pools at the bottom of Figure 1) in maintaining higher than expected available P values considering the amount of P removed by crops over time. It is increasingly common to measure these poorly soluble phosphates using the BSES procedure, a dilute acid extraction that dissolves calcium phosphates. These forms of P can be abundant in the alkaline Vertosols in the northern grains region (NGR), however the plant availability of the P forms dissolved in the BSES test may vary depending on soil pH buffering capacity, P mineral solubility and strength of other processes that remove P from soil solution (sorption/precipitation/immobilisation). This paper summarises some recent work that attempts to quantify the differences in availability of sparingly soluble P sources as affected by soil acidification.

Figure 1. Dominant P pools and processes in Vertosol soils.

Sparingly soluble P reserves

The poorly soluble P sources are widespread in the NGR.  This was shown in partnership with David Lawrence’s group in Qld, where we analysed P forms in more than 680 sites from southern and central Queensland that were collected for paddock comparisons of soil organic matter and health. Soil P status was determined to help plan fertility strategies and pasture development.

Two thirds of the soils collected would likely respond to starter P applications based on a surface Colwell P value of less than 25 mg/kg. However, in more than half of the soils collected, Colwell P values were greater than 16 mg/kg and this coincided with these soils containing the majority of the sparingly soluble P (Table 1). As available, Colwell reserves decrease, poorly soluble P sources are increasingly used and depleted if they were originally present. In the very P deficient sites, reserves of P are insignificant.  Hence, understanding how the reserves present in more than half of soils release P is important.

Table 1. Sparingly soluble P reserves of a range of Queensland soils sorted by available surface P

P released in alkaline soils during acidification

Calcium and magnesium phosphates in soils occur as a variety of native minerals or fertiliser reaction products and these all dissolve to differing extents depending on the soil pH. These minerals dissolve so slowly that they can’t sustain yield within a crop cycle, but may be able to contribute to P nutrition within a year and partially replenish Colwell P. It is also possible that they dissolve in the root zone of plants that acidify the rhizosphere.

We selected six soils (Table 2) from the NGR to examine the solubility of P forms in soils that are broadly similar (high P Vertosols) but vary in some properties (Colwell P and pH buffering capacity). pH buffering capacity reflects the ability of a soil to resist change in pH; this is related to very high CEC or the presence of free lime in the profile. Two of the highly buffered soils had free lime. 

Table 2. Selected soil properties of the Vertosols.

We incrementally acidified the soils and recovered the P with anion exchange membranes (AEM). These membranes recover P from the solution, mimicking plant uptake and minimising competitive reactions. As the soil was acidified, the pattern of P release was high to begin, as the Colwell P was extracted. In soils that don’t strongly resist change in pH, there were increases in P release around pH 7.0, 6.0 and 5.5 with large amounts of P released at higher (as well as lower) pH values.  In the strongly buffered soils, considerable P release was not observed until acidification was high and pH had fallen to 6.0.  The pH values at which P dissolved loosely correspond with the solubilities of known P minerals.  However, what is more important is that it suggests that plant roots do not have to acidify the soil as much to mobilise P from sparingly soluble sources where the pH buffering capacity is not high. The cumulative P recovered using AEM were close to the BSES concentrations for all soils other than Soil 2 and Soil 6. Soil 6, the Fernlee soil from Central Qld, contained massive reserves of sparingly soluble P and it was unlikely that this would have been depleted over the duration of the experiment (Table 2).  

Figure 2. Incremental P recovered by AEM v pH for 6 vertosols (labelled 1 – 6) with high reserves and varying Colwell values as affected by pH buffering capacity.

The reasonably complete release of P upon acidification is not unexpected.  Recently published work by Dr Tim McLaren suggested that where the BSES Ca/P ratio is less than 74, the P sources they represent can replenish and supply labile pools of P in soil (Figure 3).  This corresponded to BSES P values in soil that are greater than 60 mg/kg.  Hence, soils with <60 mg/kg BSES P may not be able to replenish available P pools easily, and this corresponds with two thirds of those soils investigated in the Queensland study. All 6 soils in our study had lower BSES Ca/P ratios than 74 and more than 60 mg/kg of BSES P.

Figure 3.  A regression plot for the ratio of BSES-Ca to BSES-P versus the ratio of resin to Colwell P (Published McLaren et al. SSSAJ )

The availability of these P forms to plants will be affected by the amount of acidification required to counteract the soil pH buffering capacity in the rhizosphere to dissolve Ca-P minerals. Note in Figure 4 that whilst P was released with minimal acid input in the less buffered soils, the amount of acid needed to release the P in the highly buffered soils is much higher. In practice this means that the ability of these sources to replenish Colwell pools is limited

Figure 4. Cumulative P recovered by AEM versus acid added (uM/g).

Conclusion

The release of sparingly soluble P sources does vary considerably, and is affected by pH buffering capacity.  Soils with free lime may release far more slowly than less buffered soils.  If the sample of Queensland soils presented is indicative of wider NGR soils, it may be that 2/3 of soils have on average depleted these reserves to the point where they no longer replenish available P supplies in a timely manner. Targeted BSES investigations may identify soils that have a large supply of this slowly available P inherent in the profile.

Contact details

Chris Guppy
University of New England
Ph: 02 6773 3567
Fx: 02 6773 3238
Email: cguppy@une.edu.au

Reviewed by

Tim McLaren
The University of Adelaide
Ph: 08 8313 0394
Email: tim.mclaren@adelaide.edu.au