Statistical and Autoregressive Analysis of Soil Physical Properties of Portsmouth Sandy Loam
1986
Anderson, S. H. | Cassel, D. K.
Reliable estimates of soil physical and hydraulic properties are needed for designing drainage systems for soils with limited natural drainage. The objective of the study was to determine the autocor-relation structure of soil physical properties in a poorly drained soil. Saturated hydraulic conductivity, soil-water characteristic, soil-water diffusivity, texture, and bulk density were measured on undisturbed soil cores sampled from the A, Btg, and Bg soil horizons at 150 points on two transects in a field of Portsmouth sandy loam (Typic Umbraquults). Unsaturated hydraulic conductivity for each of the 450 cores was computed using a quasi-analytical expression utilizing measured soil-water diffusivity data and the derivative of the inverse of the van Genuchten equation for the measured water characteristic. Hydraulic conductivity values at a given soil-water pressure had a high coefficient of variation, ranging from 130 to 3300%, and were lognormally distributed. Evidence of significant short-range autocorrelation, usually ≤2 m, was found for hydraulic conductivity, soil-water characteristic, texture, and bulk density. First-order coefficients for autoregressive equations indicated a decrease in predictability of the observed variation with increasing soil depth. The number of measurements required to estimate the mean of a soil property of the A horizon to within ±10%, ranged from four for sand content to 1985 for saturated hydraulic conductivity. The means for most soil properties in the A horizon, however, can be estimated to within ±10% with 35 observations.
显示更多 [+] 显示较少 [-]