Sulphate leaching under field conditions

Abstract

Sulphate leaching through soil was studied using a number of experimental approaches. The major factors which affect the process of sulphate leaching were examined and selected methods of prediction were tested.

Lysimeters containing undisturbed soil monoliths appeared to be the most reliable method for measuring sulphate leaching losses, provided that preferential flow of water and solute down the edge of the cylinder was prevented. Field experiments provided only indirect measurements of the leaching loss and the level of precision obtained was uncertain. Compared with results from undisturbed lysimeters, leaching studies using columns of repacked soil underestimated initial losses and overestimated final leaching losses.

The amount of sulphate lost by leaching tended to increase with the amount of drainage that occurred. In soil with impeded drainage, it was considered that lateral movement of sulphate may also be important.

Surface-applied sulphate was partitioned into 'readily' and 'sparingly' leachable forms depending on its location in the soil matrix. Readily leachable sulphate was considered to be located in large inter-aggregate pore spaces or in close proximity to macropores and was therefore susceptible to rapid leaching. Sparingly leachable sulphate was considered to be located in micropores within aggregates and was therefore only leached once it had diffused into pores which were contributing significantly to the flow of water. When water was applied immediately after an application of sulphate solution, a high proportion of the sulphate was leached rapidly.

When the water application was delayed for 24 hours, a lower initial leaching loss occurred and this was attributed to the bypass of solute which had diffused into intra-aggregate micropores. The method of water application also affected the rate of leaching with more rapid losses occurring under ponded compared with non-ponded infiltration conditions. More rapid leaching of surface applied sulphate also occurred under continuous compared with intermittent rainfall. These effects were attributed to the degree of soil saturation and hence the proportion and size of macropores involved in water flow and solute transport.

Sulphate leaching was substantially lower in a soil with a moderately high sulphate adsorption capacity compared with a non-adsorbing soil. Adsorption in the subsoil was particularly important and due to a high sesquioxide content of the soil.

Evidence obtained from an experiment using ³⁵S labelled sulphate indicated that sulphate leaching was likely to be influenced by the cycling of sulphur through the soil organic matter. During this particular study, approximately half of the sulphate applied to the soil was incorporated into soil organic sulphur compounds. However, the concurrent release of native soil organic sulphur by mineralisation could not be determined. Thus the magnitude of the effect of sulphur cycling on sulphate leaching could not be assessed accurately.

The accuracy of prediction of three leaching models was tested by comparing calculated leaching results with measured data from the experimental work. The Rose model, a semi-analytical solution of the convective-dispersive equation, gave fairly accurate predictions in poorly structured soil which had a small amount of dispersivity. However, the model was considered to be of limited application on a routine basis because it required a value of dispersivity to be obtained from an initial leaching study. Two empirically based, multi-compartment layer models, the Burns and Addiscott models, were also tested using easily obtainable, local, soil physical data. No leaching experiment was required for calibration.

Although both models gave reasonably accurate prediction under certain conditions, they were not considered to be suitable for immediate application on a routine basis without modification. Some recommendations are made to improve model prediction for sulphate leaching.