| This study was conducted to refine the genetic or landform approach to terrain analysis for soils investigations in the Las Cruces area of New Mexico, by characterizing in a more detailed and explicit manner the image interpretation elements for landform analysis (i.e., boundary, topography, drainage, erosion, tone, and vegetation/land use). Landform units were visually interpreted from 1:40,000 scale panchromatic and 1:112,000 scale color airphotos with stereoscopes, and from Landsat multispectral scanner (MSS) images through color aditive viewing. The delineated units were compared with each other and with published soil survey mapping units, using a geographic information system (Map Analysis Package). Based on the comparison results and field data, the generally accepted landform units in arid regions (i.e., alluvial fan, filled valley, playa, and sand dune) were subdivided into 23 refined units. Each unit had generally consistent grain size distributions, reflecting its ultimate source material. Delineations from both scales of airphotos generally agreed well with each other, and both generally agreed with delineations from soil survey maps. These results supported a greater use of remote sensing for soil surveys of arid regions, and of small scale aerial (and possibly space) photography for terrain analysis. Although less useful for delineating landform units, Landsat MSS images provided fairly accurate information for some soil mapping units. Topography was the dominant interpretation element for characterizing the landform units' image patterns, for delineating the units' boundaries, and for determining the degree of agreement among the different data types. This topographic dominance resulted in an overall redundancy of the interpretation elements, suggesting that some elements could probably be combined or eliminated without much loss of information. It follows that stereoscopic coverage is crucial (whereas, although color information is desirable, it is not crucial) and that the incorporation of digital terrain data should improve the results from Landsat analysis. Several potential remote sensing indicators of caliche and of relative age of landform units were identified. The results of this study were generally applicable to two other study areas, near Yuma, Arizona and Nipton, California; and should be applicable to other arid regions of the world, where adjacent to bedrock mountains. |
