SPATIAL DEPENDENCE OF SOIL-WATER POTENTIALS ASSOCIATED WITH SEPTIC SYSTEMS

Abstract

Soil properties are commonly continuous to some extent and mayi exhibit spatial dependence. Structural analysis was performed on soil-water pressure potentials associated with two designs of septic systems: a single-line-source filter field fed by gravity, and a triple-line-source filter field fed by pressure. Semivariograms were calculated and found to be anisotropic and similar for both septic systems. The nugget was zero for all days tested. The semivariance in the horizontal direction was smallest, assumed to be isotropic, and considered to represent the true spatial dependence. The true spatial dependence was described by the linear model with slopes ranging from 0.02-2.74 m2 of potential per meter of distance for the single-line system and from 0.006-1.45 m2 of potential per meter of distance for the triple-line system. The semivariance in the vertical direction was greatest due to drift present in these data. Sensitivity analysis of the effects of drift removal was performed with universal kriging. Quadratic drift was intuitively relevant; however, three expressions for the drift were analyzed, i.e., quadratic drift, linear drift, and no drift. Soil-water pressure potentials were jackknifed and compared to the measured values by calculating residuals. Quadratic drift had the lowest mean residual and was within 1% of the mean of the measured values. However, the deviations from the linear and no-drift cases were also small and within 7% of the measured mean. Overlaying contour lines of the kriged soil-water pressure potentials revealed that practical differences in the three drift expressions were negligible. When a visible difference existed, it was less than the standard deviations of the estimations and therefore not significant. It was concluded that in kriging soil-water potentials for these septic systems drift can be ignored.