Three-dimensional GIS cartography applied to the study of the spatial variation of soil horizons in a Swiss floodplain

In this study, we propose to establish a framework for the study of the spa tial variability of the soils found in the floodplain of the Sarine River a nd for the visualisation of soil distribution patterns in two- and three-di mensions (2-D, 3-D). This environment is characterised by a large lateral a nd vertical spatial variability of soils that corresponds to the temporal a nd spatial variations of the fluvial dynamics of the Sarine. The study was carried out using existing Geographical Information System (GIS) functions combined with applications specific to soil cartography. This particular GI S cartography is based on the notion of the soil horizon instead of that of the soil diagnostic profile. A Global Positioning System (GPS) survey was carried out in order to construct a local Digital Elevation Model (DEM) and to ascertain the spatial coordinates for each of the 181 soil obsevation l ocations. All data were stored in a GIS database, and both landform modelin g and soil cartography was undertaken. GIS, ARC/INFO, and Vertical Mapper f or MapInfo were adequate for our linear triangulation interpolation, for co ntour processing and for the creation of cross-sections as well as the corr esponding vertical profiles. These vertical profiles served to illustrate t he superposition of soil horizons along any line across the sampled area A 3-D representation of soil was obtained using the quadratic finite-element method, which is generally employed in geological studies and which we adap ted especially for the representation of soil horizons. 3-D cartography of this type allows the spatial pattern of a given horizon including the varia tion of its thickness, the superimposition of the different soil horizons, the total soil depth, and the number of horizons at any given location - to be followed through space. Our approach, furthermore, facilitates the perc eption of soil horizons and their juxtarelationships as 3-D objects, and pe rmits the visualisation of the relationships that exist between any given h orizon (or sequence of horizons) and the surface topography. In thus enabli ng the realistic representation and easy visualisation of the spatial distr ibution and variability of soils in the landscape, our methodological appro ach provides a powerful instrument for soil scientists, and a useful decisi on-support tool for ecosystem management. (C) 2000 Elsevier Science B.V. Al l rights reserved.