Soil solution concentrations and aluminium species of an eastern wheatbelt acidic soil of WA treated with lime and gypsum.

Key Messages

• Both lime and gypsum applied to acidic soil have led to long term improvement in wheat grain yield. Application of lime increases soil solution pH (pH_Soln) and while the application of gypsum increases soil solution sulphate (S_Soln).
• The increase in pH_Soln converts toxic Al³⁺ to inert gibbsite  while the increase in S_Soln concentration results in the formation of inert aluminium sulphate .

Aims

Soluble aluminium  often abbreviated to Al³⁺ is toxic to plants. The occurrence of this species of Al in the soil layers below 10 cm (subsoil) is a major problem in the south Western Australian agricultural region due to acidic soil, particularly in the subsurface and subsoil (Gazey et al. 2014). As the soil solution pH (pH_Soln) declines below 5.0 Al³⁺ and other species of Al concentrations in the soil solution increase (Rutkowska et al. 2015). That is hydrolysis of Al³⁺ breaks the H-O bond and the release of hydrogen ions results in the formation of; , , . Also Al³⁺ can form complexes with fluorine (F), phosphate (P), sulphate (S) and organic matter (OM) which are not toxic to plants. In general the order of decreasing toxicity of Al species is; , >  >  >  > inert  and  (Rutkowska et al. 2015).

Lime (CaCO₃) and gypsum (CaSO₄) can be used to treat the soil which has toxic levels Al³⁺ increasing wheat grain yield (McLay et al. 1994; Easton 2016). The application of lime has the effect of increasing soil pH as measured using 0.01 M CaCl₂ extracting solution (pH_Ca) resulting in the conversion of Al³⁺ to inert gibbsite  (Rutkowska et al. 2015). While the sulphate component of gypsum is leached into the subsoil resulting in toxic Al species such as Al³⁺ been changed to inert aluminium sulphate () and . There has been a large increase in lime use in the south Western Australia agricultural region over the period 2004-2013 to manage soil acidity (Gazey et al. 2014). However, gypsum is rarely used to treat subsoil Al toxicity in the wheat belt of WA. In contrast, gypsum is commonly used in Brazil for treating subsoil Al toxicity (Tiecher et al. 2018).

This paper aimed to determine whether the concentration of Al in soil solution (Al_Soln) and the species of Al had changed following the application of lime and gypsum to an acidic soil near Bonnie Rock. It was hypothesised that changes in Al concentration and species could explain the increase in wheat grain yield observed when lime and gypsum had been applied (Easton 2016).

Method

An experiment examining the effect of lime and gypsum located near Bonnie Rock (370 km north east of Perth) was conducted over the period 2008-2016 (Easton 2016). The experiment used limesand (effective neutralising value of 80%), dolomite (effective neutralising value 90%) and  an industrial gypsum source (Ca content of 22.4% and an S content of 17.8%). Treatments were applied in 2008 and again in 2013, although with dolomite sourced from Beaufort River (effective neutralising value 60%). This gives eight treatments; control, 2, 4 and 8 t/ha limesand, 2 t/ha gypsum, 4 t/ha lime + 2 t/ha gypsum, 4 t/ha dolomite, 4 t/ha dolomite + 2 t/ha gypsum replicated three times (Easton 2016).

On 21 March 2018 soil samples were collected only from the lime by gypsum factorial combination control, 4 t lime/ha (lime), 2 t gypsum/ha (gypsum) and 4 t lime/ha plus 2 t gypsum/ha (lime plus gypsum) treatments of the experiment. Four soil cores were collected from each plot and divided into increments of 10 cm to a depth of 50 cm. The layer samples of the four soil cores were then bulked together to provide a measurement of the individual treatment plot of each replicate. Soil chemical properties and gravel content were determined using standard soil analytical techniques for the control treatment (Table 1).

Table 1. Summary of measured chemical properties and gravel content of the control treatment measured on 21 March 2018.

^ASBI refers sulphate buffer index which is a modification of the procedure for measuring the S adsorption presented by Anderson and Fillery (2006).

The soil solutions were extracted using Milli Q water at the soil to solution ratios (S/L ratio) of 1/0.4 or 0.4 ml/g for the 0–10 and 10–20 cm soil layers and 1/0.35 or 0.35 ml/g soil for the other three depths. The different ratios were due to the different moisture retention at these depths. These S/L ratios were used to enable the collection of sufficient soil solution to undertake the measurements. An amount of 200 g of soil was added to a 500 mL beaker. The amount of Milli Q water depending upon the moisture retention of soil was added to the soil, covered with glad wrap to reduce the evaporation loss of water and incubated for 16 hours. The soil solution was then extracted by centrifuging the soil for 10 minutes at 10,000 rpm to extract the soil solution. An aliquot sample was passed through a 0.45 µm screen cellulose membrane.

Soil solution was analysed for alkalinity or the amount of acidity required to lower soil solution pH_Soln to 4.3 (mg/L), Ca_Soln (mg/L), Fe_Soln (mg/L), K_Soln (mg/L), Mg (mg/L), Na_Soln (mg/L), P_Soln (mg/L), S (mg/L), Si_Soln (mg/L) were measured using an ICP–AES. Redox potential or Eh_Soln (mV), pH_Soln and F_Soln (mg/L) were measured using an ion specific electrode. Nitrogen concentration as NH₃ (mg/L), NO₂ (mg/L) and NO₃ (mg/L) was measured calorimetrically. Dissolved organic carbon (DOC_Soln) (mg/L) was determined using a modified Walkley and Black (1934) method for soil solution using a sample size of 5 ml. The measured nutrient concentrations were converted to concentration in the soil solution with a water content of 14% for the 0–10 cm soil layer and 10% for the soil layers below 10 cm. These values were then entered into the chemical equilibrium model PHREEQC (Parkhurst and Appelo 1999) to estimate the composition of chemical species. Individual replicate samples were used to extract soil solution but only treatment average concentrations have been used in PHREEQC.

Analysis of variance using Genstat® was conducted on the soil solution chemical properties of the soil profile to a sampling depth of 50 cm. The analysis of variance used a treatment structure of treatment and soil sampling depth and a block structure of replicate to determine the significance of treatment, soil depth, and treatment x depth interaction. In the analysis of variance soil depth was used as a covariate. Differences between treatments, soil sampling depth and treatment x sampling depth interaction were determined using l.s.d. values at P = 0.05.

Results

Note that all of the results reported below relate to soil samples collected in 2018, that is ten and five years respectively after the original and follow-up treatments were applied.

Soil solution concentrations

The pH_Soln for the lime treatment was higher than the control treatment in the 0–10 cm soil layer (6.9 versus 4.6) and in the 10–20 cm soil layer (5.4 versus 4.3) (Figure 1a). In contrast, gypsum application did not affect pH_Soln. A comparison between soil and solution pH measurements illustrate, soil pH_Ca was strongly related to the pH_Soln (pH_soln= 1.45 x pH_Ca – 1.91, r² = 0.91) with the difference between pH_Ca and pH_Soln increasing with increasing pH_Ca.

The Al_Soln concentration for the lime treatment was lower than the control treatment in the 0–10 cm soil layer (0.1 mg Al/L versus 1.0 mg Al/L) and in the 10–20 cm soil layer (0.5 mg Al/L versus 1.7 mg Al/L) (Figure 1b). In contrast, Al_Soln concentration for the gypsum treatment was 2.3 mg Al/L in the 0–10 cm soil layer, 3.2 mg Al/L in the 10–20 cm soil layer, 2.0 mg Al/L in the 20–30 cm soil layer and 1.1 mg Al/L in the 30–40 cm soil layer. While Al_Soln concentration for the lime plus gypsum treatment was 0.0 mg Al/L in the 0–10 cm soil layer, 0.4 mg Al/L in the 10–20 cm soil layer, 0.6 mg Al/L in the 20–30 cm soil layer and 0.5 mg Al/L in the 30–40 cm soil layer.

The S_Soln concentration for the gypsum treatment was higher than the control treatment in the 0–10 cm soil layer (1426 versus 182 mg S/L), in the 10–20 cm soil layer (285 versus 80 mg S/L), in the 20–30 cm soil layer (214 versus 62 mg S/L) and in the 30–40 cm soil layer (200 versus 53 mg S/L) (Figure 1c). In contrast, S_Soln for the lime plus gypsum treatment was 639 mg S/L in the 0–10 cm soil layer and between 317-585 mg S/L in the 10–40 cm soil layer.

Figure 1. Impact of lime and gypsum treatments on soil solution (a) pH_Soln, (b) Al_Soln mg Al/L (c) S_Soln mg S/L and (d) Ca_Soln (mg Ca/L) measured on 21 March 2018.

The Ca_Soln concentration for the lime plus gypsum treatment was higher than the control treatment in the 0–10 cm soil layer (481 versus 197 mg Ca/L), in the 10–20 cm soil layer (250 versus 71 mg Ca/L), in the 20–30 cm soil layer (192 versus 50 mg Ca/L) and in the 30–40 cm soil layer (175 versus 47 mg Ca/L) (Figure 1d). Ca_Soln for the gypsum treatment was the highest, with 700 mg Ca/L in the 0–10 cm layer, 115 mg Ca/L in the soil layer 10–20 cm layer and 75 mg Ca/L in the soil layer 20–30 cm layer. Application of lime resulted in a Ca_Soln concentration of Ca_Soln 298 mg Ca/L in the 0–10 cm layer and 139 mg Ca/L in the 10–20 cm layer

Modelled Al species within the soil solution.

The main modelled Al_Soln species present for the control treatment were Al³⁺,  and  which was predicted to account for 77–96% of Al_Soln (Figure 2). Al³⁺ was predicted to account for 38% of Al_Soln in the 0–10 cm soil layer increasing to 63% in the 40–50 cm soil layer. The  was predicted to account for 20–23% of Al_Soln in the 0–20 cm soil layer and 4–11% in the 20–50 cm soil layer. Al_Soln as  and  declined from 32% in the 0-10 cm soil layer to 15% in the 40-50 cm soil layer.

Gypsum application had a low percentage of Al_Soln as Al³⁺, 18% in the 0–10 cm and to 37% in the 40–50 cm. In contrast, the percentage of Al_Soln as  and  was high (53-67%) in the soil profile. While the percentage of Al_Soln as  was stable in the 0–40 cm soil layer (6–22%) with only 6% present in the 40–50 cm soil layer.

Lime application had low percentage of Al_Soln as Al³⁺ (0-3%) and a high the percentage of Al_Soln as  in the 0-20 cm soil layer (87%) and in the 20-50 cm soil layer (40-60%). The next most common Al species was  accounting for 7% of Al_Soln in the 0–20 cm soil layers increasing to 18–26% in the 20–50 cm soil layers. There were low amounts of  and  accounting for 0-4% of Al_Soln and  accounted for 6–15% of Al_Soln in the 0–50 cm soil layers.

Lime plus gypsum application also had a low the percentage of Al_Soln as Al³⁺ to 0-1% in the soil profile and high levels of  in the 0-20 soil layer (91%). There was a high percentage of Al_Soln as  (42-60%) and  (16-22%) in the 10-50 cm soil layers. There were also low amounts of  and  accounting for only 4-10% of the Al_Soln.

Figure 2. Impact of lime and gypsum treatments on Al species present in the soil solution ten years after initial application. Al species are presented in order of decreasing toxicity with Al³⁺ (■), >  (■), >  (■), >   (■),  (■),   (■),   (■) and     (■) in the soil layers to a depth of 50 cm calculated using PHREEQC. The red shaded species indicate the forms toxic to plant while the green shaded species are not toxic.

Conclusion

Lime and gypsum application changed the Al_Soln concentration (Figure 1b) and Al species distribution in the soil solution (Figure 2). The lime application increased pH_Soln and Ca_Soln concentration and decreased Al_Soln (Figure 1) resulting in the conversion of Al³⁺ to the less toxic species ,  and  (Figure 2). The effect of lime application on pH_Soln and Al_Soln was measured to a depth of 20 cm (Figure 1) while the effect on Al species distribution occurred to a depth of 50 cm (Figure 2). Indicating assessing the Al species distribution is a more sensitive measure of the impact of lime than either Al_Soln or pH_Soln. In contrast, gypsum did not affect pH_Soln but the increased S_Soln and Al_Soln concentration (Figure 1). The increase in S_Soln occurred to a depth of 50 cm. An increase S_Soln results in toxic Al³⁺ being predicted to be converted to inert  (Figure 2, Sumner 1990). However, because gypsum application has increased Al_Soln concentration (Figure 1b), a lime application was also required to remove Al³⁺ from the soil profile (Figure 2). These changes in Al_Soln species distribution likely contributed to the increase in wheat grain yield over the control of 4-32% when gypsum was applied, of 3-33% when lime was applied and of 17-48% when both lime plus gypsum was applied (Easton 2016).

The ability of S to reduce Al toxicity in soils depends on the S_Soln concentration achieved by the gypsum application (Cameron et al. 1986). The S adsorption properties of the soil control the S_Soln concentration with soils having higher S adsorption capacity increases the amount of gypsum requirement to treat subsoil Al³⁺ (Sumner 1990). The soil profile in this experiment had a relatively low capacity to adsorb S as indicated by the low SBI values of the soil profile (Table 1). The low S sorption ability means the soil had a relatively high S_Soln concentration (Figure 1c) for the measured content of S_Soil. This interaction between S_Soil, S_Soln and percentage of S_Soil as is illustrated by the changes in these soil properties with increasing soil depth for the control treatment. As soil depth increases, both S_Soln concentration and the percentage of Al_Soln as was was predicted to decrease while S_Soil was observed to increase. Hence, the sulphate adsorption capacity of the soil increased with soil depth resulting in a decrease in the percentage of Al_Soln as . The S content of the soil (S_Soil) as measured by the KCl-40 extraction technique of the 0-30 cm soil layer (Table 1) was much greater than the critical value of 2.8 mg S/kg required for crop nutrition (Anderson et al. 2013). Therefore, crop responses to gypsum application were highly unlikely to be associated with improved S nutrition.

Gypsum application increased the concentration of Ca_Soln (Figure 1d) which would result in the displacement of Al³⁺ from the exchange sites (Mora et al. 1999) and an increase in the concentration Al_Soln (Figure 1b,). Nevertheless, for the gypsum treatment was predicted to have lower Al3+ than the control treatment due to the formation of (Figure 2).

Gypsum with S content of 17.8% applied at 2t/ha was observed to be effective in treating subsoil Al toxicity on a soil profile with relatively low sulphate adsorption capacity as indicated by the SBI measurement. The low SBI values of the soil indicate gypsum-S would have been leached rapidly into the soil profile to produce a wheat grain yield response in the first year of application (Easton 2016). But leaching of S_Soln over time was observed to reduce the effectiveness of the applied gypsum. Further research should be conducted to examine the effectiveness of gypsum in treating subsoil Al toxicity over time and on soils with higher SBI.

Further information Geoff Anderson 08 96902104 or geoff.anderson@dpird.wa.gov.au

Acknowledgements

GRDC project number DAW00242

Reviewed by Dr James Fisher

References

Anderson G, Peverill K, Brennan R (2013) Soil sulfur – crop response calibration relationships and criteria for field crops grown in Australia. Crop and Pasture Sciences 64, 523–530

Anderson GC, Fillery IRP (2006) Sulfate sorption by agricultural soils in Western Australia.  World Soil Science Congress, Philadelphia, USA, 9th – 15th July 2006.

Cameron RC, Ritchie GSP, Robson AD (1986) Relative toxicities of inorganic aluminium complexes to barley. Soil Science Society America Journal 50, 1231-1236.

Easton J (2016) Soil Acidity – Can Gypsum Increase Wheat Yields on Aluminium Toxic Soils? In 2016 GRDC Research Updates 29-30 February 2016 .

Gazey C, Davies S, Master R (2014) Soil acidity – A guide for WA farmers and consultants, Second edition. Department of Agriculture and Food, Western Australia, Bulletin 4784. ISSN 1833-7236

McLay CDA, Ritchie GSP, Porter WM, Cruse A (1994) Amelioration of subsurface acidity in sandy soils in low rainfall regions. I. Responses of wheat and lupins to surface-applied gypsum and lime. Australian Journal of Soil Research 32, 835–846

Mora ML, Schnettler B, Demanet R (1999) Effect of liming and gypsum on soil chemistry, yield, and mineral composition of ryegrass grown in an acidic Andisol. Communications in Soil Science and Plant Analysis 30, 9-10.

Parkhurst DL, Appelo CAJ (1999) User’s guide to PHREEQC (Version 2) A computer program for speciation, batch–reaction, one dimensional transport, and inverse geochemical calculations: U.S. Geological Survey Water–Resources Investigations Report 99–4259, 0-310.

Rutkowska B, Szulc W, Hoch M, Spychaj–Fabisiak E (2015) Forms of Al in soil and soil solution in a long–term fertilizer application experiment. Soil Use and Management 31, 114–120.

Sumner ME (1990) Gypsum as an ameliorant for the subsoil acidity syndrome. Publication No 01–024–090. Florida Institute of Phosphate Research, Bartow.

Tiecher T, Pias OHC, Bayer C, Martins AP, Denardin LGO, Anghinoni I (2018) Crop response to gypsum application to subtropical soils under no-till in Brazil: a systematic review. Rev Bras Cienc Solo. 42:e0170025.