A study of the leaching of non-reactive solutes and nitrate under laboratory and field conditions : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Soil Science at Massey University
Leaching of solutes such as nitrate from soil to surface water and groundwater is of environmental and economic concern. Leaching experiments were conducted both in the laboratory using large intact soil cores (230 mm diameter; 250 mm depth) and in the field using a mole-pipe drained Tokomaru silt loam soil under pasture. In the laboratory experiments 'tracers' (tritium, bromide or chloride) were applied as pulse or step-change inputs to the soil surface during steady flow. A transfer function model, based on a probability density function (pdf), which characterised solute travel between inlet and outlet surfaces in terms of cumulative drainage, was used to predict solute movement. Using tracer model parameters, leaching of indigenous chloride was reasonably predicted, but the leaching of indigenous nitrate could not be modelled satisfactorily. This was apparently due to the dynamic nature and spatial variability of the biological transformations to which nitrate is subject in soil. In the field experiment solid sodium bromide and urea were applied in autumn 1990 to adjacent drained paddocks, each 0.125 ha in area. Soil, suction-cup and drainage samples were collected regularly during the drainage seasons of 1990 and 1991. The average amounts of drainage collected were 250 mm in 1990 and 320 mm in 1991, but the average amounts of nitrate leached were 47 and 20 kg N/ha, respectively. The results indicate the importance of source-strength for nitrate in N leaching loss. The nitrate-N concentration was around 35 g N m-3 in the early drainage, well above the WHO limit of 10 g N m-3, but dropped to around 2 g N m-3 later in the drainage season. About 8 % of the applied N, but 52 % of the applied bromide, was leached during the 1990 drainage season. This shows the important effect that biological reactions such as immobilization can have in reducing nitrate leaching. Comparisons were made between solute concentrations of suction-cup solution, soil extracted solution, and the drainage. For non-reactive solutes such as bromide (an applied solute) and chloride (an indigenous solute) die suction cup data provided better estimates of the solute concentration in the drainage than did the soil solution data. For nitrate, neither of these two measurements could estimate accurately solute concentrations in the drainage. The solute leaching data obtained in the field were modelled using transfer functions. The bromide and chloride data were used to calculate the pdf of solute travel times. For chloride, an exponential pdf fitted the data slightly better than a lognormal pdf, despite it having only one rather than two fitted parameters. The chloride pdf appeared to be similar for both 1990 and 1991. For bromide, the inferred pdf conformed to a log-normal distribution and was quite different from the pdf derived from the chloride data. It seems that assuming a pulse (Dirac delta) flux input for a surface-applied solid fertilizer is not valid, and that this is the reason for the discrepancy between the pdfs obtained using the bromide and chloride data. When the pdf derived from the chloride data was used to model nitrate leaching, the result was generally disappointing.