Investigations of soil extractable aluminium and toxicity in New Zealand soils : A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy at Lincoln University

Abstract

Soil acidity and associated soil aluminium (Al) toxicity severely restrict the establishment, yield and persistence of legumes in New Zealand high and hill country pastures. There is an urgent need to identify soils which are most susceptible to Al toxicity and to determine what factors are involved in regulating soil extractable Al concentrations. This study investigated the relationship between soil chemical, physical and environmental variables and soil extractable Al for New Zealand soils in national, catchment and rhizosphere contexts. The relationship between soil chemical, physical and environment variables and soil extractable AlKCl were investigated using the National Soils Database (NSD). Base saturation (BS), soil pHH2O, cation exchange capacity (CEC), total nitrogen, total carbon and soil order were strongly associated with AlKCl concentrations and relationships differed among the depth zones (0-20 cm, 20-50 cm and 50-120 cm). Soil acidity and high CEC contributed to high AlKCl concentrations, whereas a high BS and total C had a negative effect. Total N decreased with increasing AlKCl in the top 20 cm, likely as a response to Al toxicity induced limitations on biological N fixation by pasture legumes. An AlKCl concentration > 1.0 cmolc/kg, which can be toxic to sensitive plants, occurred across a pHH2O range of 3.8-6.4. Brown Soils and Podzols had the highest mean AlKCl concentrations across all depths and are likely to be more susceptible to Al toxicity. A soil survey in the Ashburton Lakes catchment was conducted to determine which soils are most susceptible to Al toxicity and to identify which factors drive soil pHH2O and extractable AlCaCl2 at a landscape scale. Depth in the soil profile was the strongest explanatory variable for pHH2O and AlCaCl2. Soil pHH2O increased with depth and AlCaCl2 declined. Rainfall and age were significant factors for AlCaCl2, however, there were no systematic patterns of an increase in AlCaCl2 with increasing rainfall and soil age. Differences in pHH2O among landform types were found, in contrast to no difference in AlCaCl2 concentrations. The soil pHH2O ranged from 4.7-6.0 and AlCaCl2 concentrations from 1.2 mg kg-1 to 39.1 mg kg-1, with a mean of 7.8 mg kg-1. Maps of soil pHH2O and AlCaCl2 in the 20 cm depth zone were constructed using the rules established by decision trees. Distinct areas in the landscape were identified which had higher concentrations of AlCaCl2. Higher AlCaCl2 concentrations were found at the wettest sites in the catchment (≥1266 mm), areas that seem to mirror those that were identified as most acidic in the soil pHH2O map. The growth response and nutrient uptake of legumes (Medicago sativa L. and Trifolium ambiguum L.) as bioindicators of Al toxicity were assessed in a range of acidic soils in a glasshouse experiment. Soil extractable AlCaCl2 concentration was strongly associated with lucerne shoot yields. Lucerne shoot yield increased one to six fold with lime application, particularly between the L0P0 and 2 t lime ha-1 treatments. Yield increases were strongly associated with declines in the soil extractable AlCaCl2 concentration to below toxic levels (≤2.5 mg kg-1) and clearly demonstrated the severe plant growth restriction of Al toxicity in these high country soils. On most soils lucerne shoot yields responded more to lime than P applied. In contrast, Caucasian clover (CC) shoot yields were not affected by soil extractable Al concentration, with more consistent yields across the range of pHH2O (5.0-7.5) and the P rates applied than lucerne. This study clearly highlights the potential importance of CC use in the high country, where the growth of more sensitive species such as lucerne is restricted by Al toxicity. The plant effect on Al mobilisation and immobilisation at the root-soil interface and the effects of pHH2O were investigated in a rhizobox experiment with legumes (Lupinus polyphyllus L. and Medicago sativa L.) in an acidic high country soil. The pHH2O was more acidic (0.1- 0.3 pH units lower) and AlCacl2 concentrations were higher (0.5 mg kg-1 and 5.4 mg kg-1) in the rhizosphere of lupin plants compared to the bulk soil. Lucerne plants had a similar soil pHH2O between the bulk and rhizosphere. Hot water extractable organic carbon levels appeared to be consistently higher in the rhizosphere and seemed to increase at the highest lime application. DGT data showed increased mobilisation of Al at the root tip of lupin and depletion along the root axis, indicative of previous removal of Al by the plant from the soil. This is the first study that has shown, in high resolution using DGT and LA-ICP-MS analysis, distinct patterns of soil Al mobilisation induced by the roots of important pasture species. A laboratory investigation was conducted to determine if changing the molarity and extraction time of the standard CaCl2 and KCl soil Al tests altered the Al concentrations extracted. Overall, the Al concentration extracted by the KCl standard test was 16 times higher than the CaCl2 across all soils. The effect of molarity and extraction time on the Al concentrations extracted differed among the five soils tested for the two extraction methods. For the CaCl2 test, the extractable Al concentration increased (P<0.001) with an increase in the molarity (by 8.7-17.7 mg kg-1) of CaCl2 for most soils. Only the Allophanic Soil measured a difference (P<0.001) in extractable AlCaCl2 with an increase in extraction time (decrease in AlCaCl2). The interaction of molarity and extraction time only extracted different (P<0.05) concentrations of Al on two soils for the CaCl2 extraction. For the KCl extraction, on most soils the Al extracted increased (P<0.001) with an increase in molarity to 1 M (by 0.5-1.0 cmolc/kg), with no increase (P>0.05) with a further increase in molarity. For two of the soils, the Al extracted was significantly affected by the extraction time, however, the results were contrasting. On most soils the interaction of molarity and extraction time extracted significantly different concentrations of AlKCl. These findings suggest that the Al concentrations measured by the two extraction methods are affected by specific soil properties in the topsoil related to soil order. This research has identified key variables driving soil extractable Al concentrations in a suite of New Zealand soils, and at different scales. Legume species responded differently to soil extractable AlCaCl2 and influenced pH and the amount of soil Al in the rhizosphere. Soil properties of different soil orders were also found to affect the amount of Al extracted from the soil by the two CaCl2 and KCl extraction methods. The knowledge generated from this thesis has identified specific sets of conditions (environmental, soil chemical and soil order) that have higher concentrations of extractable Al and therefore areas most likely to be susceptible to Al toxicity in New Zealand.