QUANTIFYING SOIL VARIABILITY IN GIS APPLICATIONS .1. ESTIMATES OF POSITION

Geographical information systems (GIS) commonly make use of soil infor mation which consists of soil maps, descriptive text, and tables of at tribute values. However for optimal use, published soil maps first nee d to be digitized and registered and subsequently the variability asso ciated with tabular soil attribute information needs to be quantified. Validity of individual field observations depends on combined accurac y of an attribute estimate, sampling position location, and interpolat ion or extrapolation error among measured values. In this paper (part I), we evaluate the accuracy of sampling position location by comparin g Global Positioning System (GPS) field measured values with visual es timates obtained from aerial photographs. Results indicate that such v isual estimates of position, registered using standard procedures, can be in considerable error. We concluded, that best estimates of positi on on a GIS map are obtained when coordinates of position are computed from kriged estimates, or a projection type registration procedure.