Spatial information tools for delineating landform elements to support soil/landscape analysis

Soil properties vary laterally with topography, resulting in patterns of soil development. This relationship allows for prediction of soil attributes from landscape position. Slopes are often divided into five segments: summit, shoulder, backslope, footslope and toeslope. These delineations are usually determined using stereo photographs or field observations, based on elevation and profile curvature.; This dissertation investigated new methods to automate this process and extend the utility of the classification for scientists and land managers. Continuous classification (fuzzy logic) methods and unsupervised (ISODATA) classification techniques were used to classify each pixel of a 10-meter resolution digital elevation model (DEM) according to its membership in a landform element class. These classes were determined by the natural clustering of the data in attribute space. Attributes used for the classification, derived from the DEM, were elevation, slope, profile and tangent (related to plan) curvature, compound topographic index (steady-state wetness index), and incident solar radiation. The unsupervised classification assigned pixels to one, and only one, landform class while continuous classification allocated relative class memberships to each pixel. Regression analysis was used to ascertain relationships between depth of silt and the landform characterization. Significant variables proved to be elevation, the wetness index and incident solar radiation.; Work was performed within a geographic information system (GIS) and other commercially available software packages. The GIS facilitated manipulation and visualization of the input data and results, allowing for various combinations of layers of data (e.g., manual delineations, soil map units, residuals of regression analyses) to evaluate the outcomes of the numerical classification methods and statistical analyses. The DEM was used to create a wireframe 2{dollar}{lcub}1over2{rcub}{dollar}-D representation of the study site over which grids of results were draped to provide another dimension to viewing the data.; These numerical classification methods may prove useful to soil scientists for determining sampling designs for field work. An understanding of the landscape is still necessary to properly analyze the results of the classifications, but the techniques do bring a measure of objectiveness to a process that has been largely intuitive and experiential. The ISODATA method is a fairly easy way to quickly delineate landform elements within which to sample. The continuous classification provides more information about the physical nature of sites and may be well suited to statistical analyses.