Distributed Hydrology-Soil-Vegetation Model (DHSVM)
Forest Management Application
Forest roads have received attention recently for their potential effects on watersheds and wildlife habitat. The Clinton Administration ordered a moratorium in February 1999 on road construction in some national forests. Runoff production on road surfaces and the interception of subsurface flow by road ditches and cut-banks may alter flow production and routing within a watershed. Greater amounts of water may enter some channels at a faster rate, altering channel morphology and fish habitat.
A road reach begins to intercept subsurface flow when grid cell water tables rise above the elevation of the associated road drainage ditches (Figure 3). The road drainage network also receives runoff directly from the surface of in-sloped roads. Impacts of the road network on channel flows are modeled by GISWA explicitly; road location and geometry, along with soil moisture determine the volume of water intercepted, while road drainage characteristics and culvert placement determine the path and travel time to the channel system.
Figure 3. Subsurface flow interception by a logging road in a model grid cell. Click for a larger version.
GISWA was applied to the experimental Carnation Creek watershed to evaluate the influence of logging road design on channel flow. Interception of subsurface flow by the road network alters significantly the distribution of soil moisture and runoff generation in many areas of the basin. The road network increases the contributing area of some channel segments and decreases the area draining to others. This effect may change with time as watertables rise and more subsurface flow is intercepted. These effects are illustrated in Figure 4 where discharge in segment 121 is increased greatly by a road network when cut depths are equal to the full soil depth. The same road network with cut depths equal to one-half the soil depth does not begin to impact channel flow until later in the storm when water tables rise above these shallower road cuts. By the time significant flow interception occurs, the rainfall has reached its peak and decreases rapidly, causing a slight increase in discharge relative to the non-road case.
Figure 4. Basin-averaged precipitation for December 15-17, 1972 (upper) and simulated discharge with and without roads for channel segments 121 (middle) and 130 (lower). Simulated discharge is presented without roads, with road cut depths one-half the soil thickness, and roadcut depths equal to the total soil thickness. Click for a larger version.
Channel segment 130 provides an example where the road network diverts water away from the channel, reducing streamflow. In the first two hours of the storm discharge is the same in all three scenarios because the area contributing directly to the channel has not extended upslope to the road system. Later in the storm the contributing area in the non-road case continues to expand beyond the road locations and discharge continues to increase. With a full cut depth the channel contributing area can not extend above the road locations, resulting in lower discharges than under natural conditions. The one-half cut depth results in flows close to, but lower than the non-road case.
Recent applications of the model have been used to evaluate the impacts of forest harvesting and road construction on watersheds in western Washington (Storck et al. 1995; Bowling and Lettenmaier, 1997; Lamarche and Lettenmaier, 1998), eastern Washington (Wetherbee and Lettenmaier, 1997), and western British Columbia (Wigmosta and Perkins, 1997).