Comparison of topographic and physiographic properties measured on the ground with those derived from digital elevation models

摘要

With the widespread availability of digital elevation models (DEM) and regional surveys of soils, topographic and physiographic features of landscapes are now more easily characterized. Within southwestern Oregon 391 field plots were registered within a geographic information system (GIS) to digitized topographic and soils coverages and properties extracted from the digital coverages compared with those estimated in the field. The initial comparison showed major differences in estimates of aspect, slope, and maximum available soil water content (#theta#), although the location of plots showed general agreement with elevations recorded on the maps. To extrapolate climatic data and interpret hydrologic responses accurately, an automated search procedure was developed whereby the initial location of each plot was, if necessary, shifted within specified bounds to give closer agreement with field estimates of aspect, slope, and #theta#. Specifically, the search routine sequentially identifies the nearest 100 m-resolution cell within a search radius of 3 or 5 cells in which differences are within +-22.5 deg of aspect, +-20% of slope, and in closest agreement with field estimates of 0. The search procedure resulted in improved agreement with field estimates: r~2s = 0.82 for aspect, 0.56 for slope, 0.54 for #theta#. To obtain these improvements required that the initial plot locations be shifted, on the average, 289m within the 3-pixel search radius, and 435 m within the 5 pixel radius. With the terrain analysis procedures developed in this paper, it is possible to overcome many problems associated with registering the precise location of field plots upon digitized topographic and soil maps. The procedure is particularly appropriate in situations where the environmental regimes associated with a specified field location are to be extrapolated across landscapes. The approach also permits a wealth of historical survey plot data to be incorporated into a GIS format and to be spatially extended.