New Mexico Geological Society Annual Spring Meeting — Abstracts

The creation of accurate small scale soil maps is often very difficult due to time and cost constraints. The traditional methods of soil mapping include using several sets of proxy data to identify boundaries. Aerial photographs are often the most useful due to the amount of detail they often show, from clusters of trees to outcrops of rock. With this data, a soil scientist can infer more detail about soil changes based on the land cover. It is possible for an experienced soil scientist to be able to associate certain landforms with sets of internal soil properties. Using these aids, among others, it is possible to get a preliminary idea of where boundaries will occur. This allows the researcher to have a plan, prior to going in the field, of where he needs to spend the most time. Most time in the field will be spent crossing back and forth across these boundaries to confirm or change them as necessary.

Digital soil mapping is the computer-based operationalization of ideas for predicting soil distribution and has become more feasible with the advance of computational technology. Remote sensing techniques can provide data that is spatially contiguous and spectrally contiguous and it has been used successfully in the past to carry out landuse and landcover surveys. It provides a quicker and more efficient method of mapping soils. Using data from the Landsat 7 processed through the SEBAL model, visual correlations are seen between parameters such as albedo, surface temperature, normalized difference vegetation index, evapotranspiration and soil moisture and soil unit. Visual analysis has been completed at two very different field sites, the Sevilleta National Wildlife Refuge and the Hilton Ranch. In both areas unsupervised classifications have been performed to analyze any correlations with soil unit. More detailed field data has been collected at the Sevilleta, further quantifying the visual correlations.