EM 1110-2-2907
1 October 2003
These calculations provided the test required to determine the effectiveness of the smooth-
ing and resampling techniques. The visual analysis of the image overlaid onto the 10-m
resolution DEM revealed little noise. The USPED model is described as being "very sensi-
tive to artifacts in a DEM as it is a function of second order derivatives (curvatures) of the
elevation surface." With the reduced noise in the data, the USPED model is predicted to ac-
curately assess soil erosion and deposition.
(2) Sediment flow rates were calculated for a subset area from within a 36-km2 area of
Fort McCoy, Wisconsin. The rate was determined with the use of the SIMWE, which solves
for the continuity of mass equation. The results indicated high sediment flow rates in valley
centers and varying flow rates in adjacent areas. The SIMWE model compared well with the
USPED model results
(3) The USPED and SIMWE models were also compared in an analysis of soil trans-
port in the Fort McCoy, Wisconsin, area. Topographic potential for erosion and deposition
were estimated with the USPED model using a 30-m and a 10-m DEM. The 10-m data were
again derived from the 30-m data by a smoothing and resampling technique.
(4) The GIS map is based on the 10-m data denoting areas of high potential for soil
erosion, typically shown to be hilly areas adjacent to streams. This landscape model showed
areas of temporary deposition, where soil and sediment resided before entering the main
stream. The map created with the 30-m data inadequately predicted the areas of soil loss; it
was suggested this was the result of concentrated flow in valleys. Furthermore, artificial
waves of erosion and deposition were shown in flat areas. This was due to the vertical reso-
lution of up to 1 m in the 30-m pixel size DEM. The 10-m data maintains a lower 0.1-m
vertical resolution.
(5) When the 10-m resolution DEM was used with the USPED model, intense erosion
was predicted in the hilly regions adjacent the main streams and tributaries. Deposition con-
tinued to be evident in the concave areas. Distinct from the map derived with 30-m DEM,
the 10-m resolution DEM GIS map indicated high erosion in areas with concentrated flow
that could reach the main streams. The artificial pattern of erosion/deposition along nearly
flat contours was not depicted in the 10-m GIS data.
e. Conclusions. The CASC2D, USPED, and SIMWE soil erosion models significantly
advanced the simulation of runoff, erosion, and sediment transport and deposition. With the
application of factors relating to three dimensions, these models better predict the spatial
distribution and motion of soil and sediments in a watershed. The 10-m resolution was
shown to be most advantageous in revealing the detail required to model soil erosion and
deposition. The 10 m resolution was easily developed from 30-m pixel sized data with the
use of software resampling tools followed by a smoothing algorithm. In summary, this work
potentially improves land management and should reduce land maintenance and restoration
costs.
Point of Contact: Steven Warren; swarren@cemml.colostate.edu
6-24
Previous Page
