PS 50-154
Using high-resolution aerial photogrammetry to 3-dimensionally model soil erosion in rangeland ecosystems

Wednesday, August 7, 2013
Exhibit Hall B, Minneapolis Convention Center
Jeffrey Gillan, USDA ARS Jornada Experimental Range, Las Cruces, NM
Jason W. Karl, Jornada Experimental Range, USDA Agricultural Research Service, Las Cruces
Nichole N. Barger, Ecology and Evolutionary Biology, University of Colorado Boulder, Boulder, CO
Michael C. Duniway, Southwest Biological Science Center, U.S. Geological Survey, Moab, UT
Ahmed Elaksher, Engineering Technology and Surveying Engineering, New Mexico State University, Las Cruces, NM

Soil erosion accelerated by land manipulation and overuse is a leading cause of degradation of rangelands that we value and depend on for livestock production, watershed services, biodiversity, and recreation. Multi-scale monitoring programs that follow erosion trends are essential if we are to diagnose and take corrective action. Remote sensing methods for detecting and quantitatively measuring soil erosion (scouring, deposition, redistribution) are currently being sought for monitoring programs, especially at broad-scales where efficiencies are available. To test the feasibility of this approach, this study compared image-based estimates of soil erosion with field-based estimates. In a 31 ha Piñon-Juniper woodland in southeastern Utah, a suite of fuel reduction vegetation treatments were carried out with the goal of comparing erosion rates between the treatment methods. Wind and water erosion were measured with wind aspirated dust samplers (BSNEs) and 1 m silt fences. Using aerial photogrammetric methods we created high-resolution 3D digital terrain models of the study area before and after the vegetation treatments. By differencing the elevations of these two models, we were able to map and volumetrically measure soil movement in each of the treatment plots.


Results from the field data show that plots treated with prescribed fire had a 11-32 fold increase in windborne sediment transport as compared to masticated and untreated control plots in the first two growing season after treatment. Water erosion, as measured by sediment in the silt fences was 20-40 times higher in plots with prescribed fire compared with masticated and untreated control plots. Results from the image analysis echo the findings of the field data and show that burned plots experienced significantly more soil movement one year after treatment compared with masticated and untreated plots. In addition, the remote sensing approach provided a comprehensive survey of the soil movement while the field methods represented a sample. The locations of soil scours and deposits could be easily identified in the imagery which promoted a better understanding, through visualization, of the erosive processes specific to this site. These methods could also be used to identify and estimate development rates for rills, gullies, pedestals and terracettes which are indicators of rangeland degradation related to soil and site stability. Overall, photogrammetric based modeling of rangeland erosion is a promising technique for monitoring large landscapes.