Field study on the local variability of soil water content and solute concentration

• Published in: Journal of hydrology (Amsterdam) 1999-02, Vol.215 (1-4), p.23-37
• Main Authors: Netto, A.M, Pieritz, R.A, Gaudet, J.P
• Format: Article
• Language: English
• ISSN: 0022-1694
• DOI: 10.1016/S0022-1694(98)00259-5

We characterized temporal and spatial variables describing hydraulic and solute transport in a heterogeneous soil. Under field conditions, bulk density, water content and concentration of three applied tracers (conservative Br super(-), nitrate NO super(-) sub(3), and a pesticide BENTAZON) were measured after destructive sampling and compared to continuous measurements. A comparison of destructive sampling and continuous measurements were used to investigate the field. Soil volumetric samples were obtained from trenches on a grid of 1.2 x 1.1 x 0.1 m in a fallow area of approximately 1300 m super(2) to generate two dimensional (2D) interpolated maps. Bulk density, water content, and solute concentration data sets were evaluated by geostatistical analysis in order to choose the best method to build interpolations. Semivariogram analyses for volumetric water content ( theta ), bulk density ( rho sub(d)) and solute concentrations (NO super(-) sub(3) and BENTAZON) showed no autocorrelation possessing only a pure nugget effect. 2D interpolated maps of these soil parameters were constructed using the Radial Basis Function method that allowed calculation of vertical water content and solute concentration profiles and the associated mean ( mu ) and variance ( sigma ) distributions. Vertical neutron and TDR water content profiles were successfully simulated with a discrete geometry approach using the 2D interpolated maps. Destructive sampling for water content and continuous measurements using neutron and TDR probes, revealed consistent results. Comparison of solute concentration from destructive sampling and ceramic solution samplers is less satisfactory at this site owing to both the heterogeneity and the inadequacy of ceramic solution samplers to sample the total soil concentration. Finally, preferential flow pathways can be visualized on the 2D interpolated maps of the concentration of Br super(-), BENTAZON, and NO super(-) sub(3). The observed distribution of tracers provides clear evidence of the influence of both soil hydraulic properties and geochemical heterogeneities which must be incorporated to predict water and solute fluxes.