| Popis: |
Interpolation inaccuracies can occur when spatially mapping bulk soil conductivity (ECa). These inaccuracies are compounded when surveying large acreages using mobile equipment. As a mobile survey is typically more time extensive and covers more acreage than a pedestrian survey, allowances for temporal effects and instrument orientation are necessary. To evaluate these inaccuracies, data were gathered to determine the influences of short-term conductivity shifts occurring over a data collection period, as well as from travel route patterns that determine instrument orientation. Using root-mean-squared error (RMSE) to quantify the transformation accuracy of interpolated ECa maps, a data collection method and an appropriate geostatistical model were determined for the loessial soils of West Tennessee. Analysis showed that parallel travel paths, rather than crossing travel paths, helped to insure improved map quality; and interpolation inconsistencies were reduced when measurements were obtained in a manner by which all adjacent measurements were temporally contiguous as possible. For our specific application, ordinary kriging (OK) gave acceptable map transformation predictions, as did the inverse distance weighted (IDW) and radial basis function (RBF) interpolation models. |
