Forests and Water: A State of the Art Review for Colorado

Home Browse Search

The hydrologic significance of roads is due to their potential to alter both the processes by which runoff is generated and the rate at which water is delivered to stream channels. The change in runoff processes is due to the fact that forested areas typically have high infiltration rates and most or all of the precipitation and snowmelt infiltrates into the soil; overland flow is rarely observed (Hewlett, 1982; Troendle, 1987b). In contrast, the infiltration rates for unpaved roads are usually no more than 1 mm or 0.04 inches per hour (e.g., Luce and Cundy, 1994). The low infiltration rate means that most of the rainfall and snowmelt on road surfaces will rapidly run off as infiltration- excess overland flow. If the roads are insloped, the road surface runoff is directed into an inside ditch, and these ditches commonly drain directly into swales or stream channels. This means that roads not only generate more runoff as a result of their low infiltration rates, but also provide a pathway to rapidly deliver this runoff to the stream network (Montgomery, 1994; Wemple et al., 1996). Roads that are cut into the hillslope also can affect the amount and timing of runoff by intercepting the water that normally flows downslope through the soil mantle (e.g., Megahan, 1972). This intercepted subsurface stormflow is often collected by inside ditches and deliv- ered to the stream network in the same way as the runoff from the road surface. This interception of subsurface flow is important because it transforms slower- moving subsurface flow to faster - moving surface runoff, and rapidly delivers this intercepted subsurface stormflow to the stream network (LaMarche and Lettenmaier, 2001). The additional runoff generated from roads can initiate channels where they normally would not be present (e.g., Montgomery, 1994), and the road ditches also can be di- rectly linked to the new and pre - existing channels. The resulting increase in the drainage density (defined as the total length of channels per unit area) increases the rate at which water is delivered to the stream network and this can substantially increase the size of peak flows (La Marche and Lettenmaier, 2001). While these changes in runoff generation and routing have been documented at the site scale and several studies have concluded that roads can increase the size of peak flows at the water- shed scale (e.g., Jones and Grant, 1996; Jones, 2000; La Marche and Lettenmaier, 2001), the effect of roads on the size of peak flows is still controversial and largely undocumented at the watershed scale (e.g., Thomas and Megahan, 1998). Most of the research on road runoff has been conducted in rain - dominated areas, but roads should have similar effects in snowmelt- dominated areas. However, there are several reasons why the hydrologic effects of roads in snowmelt - dominated areas may be less than in rain - dominated areas. First, snowmelt rates are much less than peak rainfall rates, so proportionally more of the snowmelt on unpaved road surfaces will infiltrate. Sec- ond, the lag between peak snowmelt and peak runoff for small to moderate -sized basins is typically around 5 -12 hours, so the more rapid delivery of water from roads to the stream network will not necessarily coincide with peak snowmelt runoff from the rest of the basin. Data from paired -basin experiments in snowmelt -domi- nated areas have failed to demonstrate a change in the size of peak flows due to roads. In both the Fool Creek and Coon Creek experiments the roads were built a cou- ple of years before the timber was harvested, and in each case there was no detectable change in runoff. Since the road network occupied less than 2% of the total water- shed in each case, it should not be surprising that road construction caused no detectable change in runoff in the short monitoring period prior to forest harvest ( Troendle et al., 2001). In addition to roads, surface runoff can be generated from roofs, hiking or off -road vehicle trails, and com- pacted areas such as skid trails and landings. Runoff from these areas is more likely to be routed onto forested slopes rather than into ditches, in which case it has a greater likelihood of infiltrating into the soil. To the ex- tent that this runoff can infiltrate into the soil and is not routed to the stream, one generally would expect these other sources of overland flow to have less of an effect on watershed -scale runoff rates than forest roads. In summary, unpaved roads, compacted areas, and im- pervious surfaces can generate overland flow. Road cuts will intercept subsurface stormflow, and the road drain- age network may rapidly route this runoff into the stream network. The impact of these changes on runoff at the watershed scale depends on the proportion of the water- shed that is affected (Harr, 1986) and the drainage design of the compacted areas. Insloped roads with ditches are usually of greatest concern because they typically col- lect the runoff and deliver it directly to the stream net- work. The hydrologic changes due to roads have been documented at the road segment and hillslope scale, but have not been confirmed at the catchment scale. Roads and skid trails have an even greater effect on sediment production than runoff, and this issue is discussed in Chapter 3. 16