Evaluating soil variability as related to landscape position using different statistical methods

Abstract

The variability of selected upland soils in Brazos County, Texas was studied using a sampling design that consisted of a main grid with 121 points spaced 30 m apart, three subgrids within the main grid, each with 49 points spaced 5 m apart, and a backhoe pit within each subgrid. At each sampling point the surface and solum thicknesses; the depth, designation, color, and texture of each horizon; and the probable soil series were recorded. Samples were taken from the upper 15 cm of the A-horizon, the upper 50 cm of the Bt-horizon, and at the critical depth for determining base saturation for classification. Soil pH, extractable bases, and particle size distribution were determined for each sample and cation exchange capacity was determined for the lower two samples. Data were analyzed using traditional statistical methods and geostatistics. Comparison of the coefficients of variation of each soil property revealed that soil pH was the least variable property while the extractable bases were the most variable; the properties on the summit position were the least variable while those on the footslope position were the most variable; and the properties on the subgrids were less variable than those on main grid. Analysis of variance revealed significant differences for most of the soil properties with respect to both landscape position and sampling depth. Comparisons of the means from the subgrids with those from the main grid using least significant differences revealed that the means for the summit and footslope subgrids were significantly different from those of the main grid. Semivariograms for each soil property revealed that most of them exhibited some spatial dependence with an average range of influence of approximately 65 m. Soil pH had the shortest range of influence with an average of 42 m while the average for surface and solum thickness was 75 m. Most of the soils on the main grid were Paleustalfs which is consistent with the great group classification of the soils mapped there by the Soil Conservation Service. However, only 47% of the soils on the footslope and 60% of the soils on the summit were identified as the series which were mapped on those positions.