A comparison of the fate of elemental sulphur and sulphate sulphur based fertilizers in pasture soils : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Soil Science at Massey University

Abstract

Nitrogen fixation by legumes has a particular requirement for adequate soil sulphur status. Sulphur (S) is a mobile nutrient and is easily leached from aquic soil environments, therefore regular topdressing with S fertilizer is required to maintain legume vigor and pasture production in most New Zealand pasture soils. Escalating fertilizer costs have focused attention on the efficiency of use of S fertilizers, particularly superphosphate (SSP) and alternative elemental S (S0) based fertilizers less liable to leaching loss in this aquic environment. Field and glasshouse trials, using the resident clover/ryegrass sward on undisturbed soil cores (150 mm diameter, 100 mm depth) , were undertaken to determine the comparative short-term fate of SSP and different particle sizes of S0. Methods for manufacturing radioactively labelled (35s) fertilizers were developed. In addition, the effect of sheep dung on the short-term immobilization of soil and fertilizer S was also investigated. A simple computer simulation model explaining the observed transformation of soil sulphur and 35s labelled fertilizer was developed. Initially, the effect of sheep dung on the short-term immobilization of soil and fertilizer S was investigated. Very small amounts (about 2-5%) of plant (clover/ryegrass pasture) S and P, within 1 00 mm of the area surrounding the dung pellet, were derived from the dung. Under the experimental conditions that prevailed, dung S behaved as a slow release S form causing neither greater immobilization of soil or fertilizer S nor mineralization of soil organic S. It was concluded that the impact of dung return on short-term (< one year) S fertilizer fate need not be considered. An initial field trial comparing the fate of microfine S0 (< 0.010 mm) relative to sulphate-based SSP was undertaken on Tokomaru silt loam, a New Zealand yellow-grey earth (Fragiaqualf). The microfine S0 oxidized within 30 days of application but initially (up to 60 days) was slightly less effective than SSP in terms of plant uptake. Over longer periods of time (150 days) their performances were comparable. Final cumulative plant uptake at 150 days accounted for 13.6% of microfine S0 and 16.3% of the SSP-sulphate. The major transformation of 35s from microfine S0 and 35s belied gypsum In SSP to soil organic 35s forms occurred in the first 30 days after application. The organic 35s activity formed from 35s0 was twice that formed from sulphate-based fertilizer and was mainly carbon- bonded 35s in the top 33mm of the pasture soil profile. The amount of organic 35s remaining as carbon-bonded 35s decreased with soil depth and the reverse occurred for the estersulphate 35s. By 1 50 days, greater activity from the microfine 35s0 remained in the soil organics fraction than from the sulphate-35s fertilizer, indicating that more soil organicS reserves may be formed through the use of fine S0 fertilizer than from the sulphate-based fertilizer. This also indicated the advantage of using S0 in minimizing the S leaching losses in this aquic environment. An inverse dilution technique using an isotope injector developed at Massey University to uniformly label undisturbed soil cores with carrier-free 35so4= solution was used to measure the impact of S0 and sulphate-based fertilizers on the fate of soil S. Results were consistent with the labelled fertilizer technique and both techniques indicated rapid incorporation of 35s into soil organic S and that the carbon-bonded S formed was likely to be a subsequent source of mineralized S available to plants. Soil samples from the preliminary field study were used to evaluate soil preparation and extraction techniques. Soil sampling and preparation techniques were evaluated on the basis that an extract sampling the plant available S pool in soil should have the same 35s specific activity as plant growing on that soil. The average 35s specific activity in a calcium dihydrogen phosphate (CaP-S) (0.04 M) extract from a freeze-dried sample of the top 60 mm of a pasture soil was most closely related to the 35s specific activity of plants growing on that soil. CaP-S extracts from field-moist soil or 0.01 M CaCI2 extracts from field-moist or freeze-dried soils had higher specific activities than plants. lt was concluded that plants were able to extract soil S from soils which was not exchangeable with added 35so4= fertilizers during either the field experiment or extraction with 0.01 M CaCI2. The second series of field and glasshouse trials were conducted to investigate the fate of 35s labelled SSP, gypsum and S0 of varying particle sizes (<0 . 1 50 mm, 0.1 50-0.250 m m and 0.250-0.500 mm, in granulated and non-granulated forms) in two pasture soils contrasting in mineralogy and fertility status. Under glasshouse conditions, 50 mm of simulated rainfall was applied to each of the undisturbed soil cores during the first 56 days after fertilizer application. For the remainder of the period, cores were watered from below using a saucer. Field cores remained subject to the local climate. Both the rate of oxidation in soil and the efficiency of plant use of S0 was improved by decreasing its particle size. Relative to soluble so4=-s applied as gypsum or SSP, the plant utilization of oxidized SJ was similar. Granulation of finer S0 with or without finely ground phosphate rock had little effect on the long-term ( 180 days) oxidation rate or the efficiency with which, after oxidation, finely ground S0 was taken up plants. Apart from S0 of large particle size (>0 . 1 50 m m) which had not oxidized, the major fate of fertilizer 35s, either under glasshouse or field conditions, was again in soil organic matter mostly formed in the top 33 mm of the soil. Applications of gypsum and SSP caused 35s to move to the 33-1 00 mm soil depths but there was no additional influence of P on the depth to which so4= was leached. A preliminary computer simulation model describing the fate of 35so4 =-s fertilizer was developed. The model provided a very accurate method of predicting plant uptake of S from both SSP fertilized and u nfertilized soil cores. The model also indicated that, at any particular soil depth, on average, actual rates of mineralizatio n a nd i m mobilization may exceed root uptake of S by 1.5 to 2 fold (mg S turned over per unit of S taken up by plants). The accuracy of the estimated turnover rate could not be validated because the model gave relatively inaccurate predictions of the measured movement and transformations of 35s tracer added to the soil as SSP. There was, however, relative similarity between the observed and predicted proportional distribution of 35s between soil and plant S forms. Such a distribution supported the concept of using root activity as a modifier of mineralization and immobilization rates in order to describe the extent of these processes at different soil depths. The study has emphasized the greater importance of the surface few millimeters of pasture soil in S transformations, important in the fate of fertilizer and pasture plant nutrition. There appears to be scope in manipulating S0 particle size to improve the efficiency of the S fertilizer used.