Introduction

Laboratory scale studies have demonstrated that turnover of microbial biomass phosphorus (P) can increase soil P solubility and leaching therefore increasing the potential for dissolved P delivery to surface waters (Seeling and Zasoski 1993; Blackwell et al. 2013). It is much less clear that the microbial biomass contributes to soil P solubility at larger spatial scales because under natural field conditions many additional factors affect P solubility, for example geochemical solubility controls. Riparian vegetated buffer strips present an opportunity to study P solubility in soils of increased organic matter contents but otherwise similar pedological properties compared with adjacent upslope arable field soils. Because microbial concentrations of P have been shown to be strongly correlated with soil organic matter (Joergensen et al. 1995), studying otherwise similar soils but with varying concentrations of organic matter may give insight into the contribution of microbial biomass P to P solubility. The aim of this study was to test the hypothesis that microbial biomass P makes a significant contribution to P solubility in riparian buffer strip soils.

Material and methods

Soil samples were collected from existing buffer strips established on three UK soil associations of differing characteristics within the national Demonstration Test Catchments (Table 1). The buffer strips were established on arable land under either Countryside Stewardship or Environmental Stewardship agri-environment schemes. At four buffer strips on each soil, five soil cores (0–7 cm depth) were collected and bulked from positions within the upslope arable field and 2 and 4 m within the buffer strip from the upslope edge during January 2011.

Table 1 Description of the location, climate, geology and vegetation characteristics of the three study soils (Collins et al. 2012; Soil survey of England and Wales 1983)
Full size table

With the exception of soil total P, sample analyses were carried out in triplicate on field moist soils that were sieved to <2 mm. Soil samples were assayed for basal soil respiration to infer microbial activity and glucose substrate induced soil respiration to approximate microbial biomass size (Campbell et al. 2003). Microbial biomass P was determined to quantify concentrations of P held within the soil microbial biomass (Brookes et al. 1982). Total soil P was measured using an Accuris inductively coupled plasma optical emission spectrometer (ARL/Fisons, Eclubens, Switzerland) after aqua regia acid digestion of air dried soils that were sieved to <2 mm. An agronomic soil test, NaHCO3 extractable inorganic P, originally designed to estimate plant available P but commonly used for determining P leaching risk, was conducted on samples according to the methods of Olsen and Sommers (1982). Phosphorus solubility was determined by extracting 5 g (dry weight equivalent) of soil with 25 ml of deionised water and shaking end-over-end for 1 h before filtration through a 0.45-μm membrane. The concentrations of total P in potassium persulphate-digested filtrates and the concentrations of inorganic P in undigested filtrates were determined by ammonium molybdate colourimetry. Organic P was calculated as the difference between inorganic P and total P concentration.

The variance of transformed data was analysed by linear modelling to determine significant differences between group means and significant relationships between variables (R statistical software version 2.14.1). Sample populations were analysed on the basis of a ‘position’ factor indicating whether samples were from the arable field or positions within the buffer strip and a ‘soil’ factor indicating significance between different soil associations.

Results and discussion

Mean concentrations of determinants within groups and significant differences in means between groups are presented in Table 2. Organic matter and microbial biomass P concentrations were significantly related (R 2 = 0.80, p ≤ 0.001) and means were significantly higher in the 2- and 4-m position groups compared with the field group (Table 2). Mean concentration of water-extractable inorganic P was significantly higher in the 2-m position group compared to the field group and was also increased in the 4-m group (Table 2). In the data as a whole, incorporating variation in soil pedological properties and management caused by the soil factor, water-extractable inorganic P concentration was most strongly related to NaHCO3 extractable inorganic P (R 2 = 0.58, p ≤ 0.001). Within individual position groups, the slopes of this relationship were greater in the two buffer strip position groups compared to the field group (Fig. 1) which confirmed that other factors were contributing to P solubility within the buffer strip soils. Inclusion of microbial biomass P and water-extractable organic P in the statistical model increased R 2 to 0.65. The variation caused by the soil factor was removed by investigating relationships within individual soil groups where water-extractable inorganic P was found to be most strongly related to microbial biomass P (Fig. 1). Therefore, by incorporating soil as a factor in the statistical model for the data as a whole, microbial biomass P was responsible for a significant (p = 0.01) amount of variation in water-extractable inorganic P.

Table 2 Means and standard errors of determinants measured within soil and position factor groups with overall factor significance
Full size table
Fig. 1
figure 1

Relationships between water-extractable P fractions and NaHCO3 extractable inorganic P within individual position groups and relationships between water-extractable P fractions and microbial biomass P within individual soil groups. Shapes indicate that groups are from within the soil factor where squares, circles and triangles denote samples from the Ardington, Burlingham and Clifton soil association groups, respectively. Shading within shapes indicates that groups are from within the position factor where transparency, hatching and opacity denote samples from field, 2- and 4-m position groups, respectively

Full size image

The significant relationships between NaHCO3 extractable inorganic P and water-extractable inorganic P concurs with the findings of previous studies on the relationship between such agronomic soil P tests and P concentrations in more soluble fractions (Pote et al. 1996) and suggests saturation and subsequent desorption of P. However, the combination of NaHCO3 extractable inorganic P, microbial biomass P and water-extractable organic P explained a greater amount of variation in inorganic P solubility. As well as desorption, inorganic P released from the microbial biomass and mineralisation of soluble organic P both also contribute to the soluble inorganic P pool. The significant relationships between water-extractable inorganic P and microbial biomass P within the soil groups shows how, when soil pedological properties and management are held relatively constant, variations in microbial biomass P concentrations can be directly responsible for significant variations in soil P solubility. Both of these findings suggest that the soluble inorganic P pool is partially independent of soil P determined by agronomic soil P tests which may not be sensitive to small but environmentally significant changes in P solubility. Mobilisation of P from the microbial biomass could therefore be responsible for previously reported variations in P solubility and leaching from soils with similar agronomic soil P concentrations but different concentrations of organic matter (Stutter et al. 2009; McDowell and Sharpley 2001). Elucidating the exact mobilisation mechanisms by which microbial biomass P contributes to P solubility will require targeted approaches and the novel experimental design will guide these future studies. Given the stable temperatures and soil moisture conditions during the period of sampling, the increased solubility is most likely due to microbial turnover of P during basal mineralisation at stable respiration rates. Under stable soil conditions, the soluble organic and inorganic P pools would be constantly maintained by microbial turnover as a consequence of microbial death and P mobilisation coupled with simultaneous multiplication and P immobilisation (Oberson and Joner 2005). Subsequent biological or biochemical mineralisation of soluble organic P would also contribute to the soluble inorganic P pool. Phosphorus turnover would also be enhanced during unstable soil conditions such as soil drying or freezing where large quantities of P could be mobilised in riparian buffer strip soils, due to microbial cell lysis and subsequent release of P (Blackwell et al. 2010).

Microbial biomass P contributed to variation in P solubility within the data as a whole and within data for the individual soil associations tested. Phosphorus solubility is therefore partially independent of agronomic soil P concentrations and depends on a range of processes which suggests that agronomic soil P testing alone will not accurately predict dissolved P leaching risk. Combining these soil tests with simple analyses for example, organic matter, clay mineral contents and water-extractable P, would greatly aid the prediction of P leaching risk at appropriate catchment management scales. While the variation in organic matter provided by the experimental system served well to study the microbial driver of P solubility, this variation also has implications for P delivery to surface waters. Riparian buffer strip and other riparian agricultural soils showing increased organic matter and microbial turnover of P may bring a dissolved P leaching risk at a critical landscape location due to increased soil P solubility. In order to reduce this risk, management of P mobilisation may be required and in the case of riparian vegetated buffer strips, occasional vegetation removal and/or tillage could help to slow organic matter build up. A better understanding of these processes and their contribution to P solubility and delivery at larger spatial scales will facilitate the development of these management strategies.