| This project tests the hypothesis that soil mapping in patterned landscapes is confounded. Typically, development of a regional soil map requires the sampling of soils, a description of their geostatistical properties (i.e., how do they vary in space) and then the construction of an interpolated map that can be used to visualize the patterns over large areas. The presence of local scale patterning at or below the minimum map resolution confounds the construction of maps made in the typical way because much of the variance in, for example, total soil P, was shown to be due to very fine scale geographic patterns that are entirely missed by a coarse regional sampling domain. We also observed that the strong correspondence between soil elevation and total P was most pronounced in areas where the ridge-slough landscape was intact, and where the landscape has been degraded by large scale hydrologic modification, this association is markedly weakened. This suggests that P may play an important role in the self-organized pattern development. |
