Hydrologic Effects of Roads

Roads have three primary effects on water:

• They intercept rainfall directly on the road surface and road cutbanks and intercept subsurface water moving down the hillslope.
• They concentrate flow, either on the surface or in an adjacent ditch or channel.
• They divert or reroute water from flow paths that it would take were the road not present.

Most hydrologic and geomorphic consequences of roads result from one or more of these processes. By intercepting surface and subsurface flow, for example, and concentrating it through diversion to ditches, gullies, and channels, road systems effectively increase the density of streams in the landscape. This increase changes the amount of time required for water to enter a stream channel, altering the timing of peak flows and changing the shape of the hydrograph (King and Tennyson 1984, Wemple and others 1996). Similarly, concentration and diversion of flow into headwater areas can cause incision of previously unchanneled portions of the landscape and initiate slides in colluvial hollows (Montgomery 1994). Diversion of streamflow at road-stream crossings is a key factor contributing to road failure and erosional consequences during large floods (Furniss and others 1998, Weaver and others 1995).

Hydrologically, different parts of the road system behave differently. All roads do not perform the same during a given storm, and the same road segment may behave differently during a given storm of different magnitudes. Recent, detailed examination of hydrographs at stream crossings with culverts shows that during the same storm, some road segments contribute substantially more flow to channels than others, primarily owing to differences in the amount of subsurface water intercepted at the cut bank (Bowling and Lettenmeier 1997, Wemple and others 1996). As storms become larger or soil becomes wetter, more of the road system contributes water directly to streams. Slope position has a profound effect on the magnitude of hydrologic change caused by roads. Discharge from hill slopes, height of cut bank, density of stream crossings, soil properties, and response to storms all differ with slope position.

Research on the Hydrologic Effects of Roads

Although hydrologic effects of roads have been studied for more than 50 years, systematic studies with long-term measurement of the full range of potential interactions between water and roads are few. Most studies have emphasized geotechnical issues, including road design, culvert size and placement, and erosion control from road surfaces (see Reid and others 1997, for bibliography; Swift 1988). Of those studies that have attempted to look at the hydrologic behavior of roads, most have been part of small (typically 0.3 to 2 square miles) watershed experiments, where roads were a component of the experimental treatment, which often included other silvicultural practices. Key studies of this type include those by Swank and others (1982, 1988) in the southern Appalachians, Helvey and Kochenderfer (1988) in the central Appalachians; and Hornbeck (1973) and Hornbeck and others (1997) in the northern Appalachians. Very few studies have focused on the hydrologic behavior of roads alone. Most studies have compared streamflow hydrographs before and after road building, with little ability to identify key processes. Exceptions include the work of Megahan (1972), Keppeler and others (1994), and Wemple (1994) on subsurface flow interception and the work of Luce and Cundy (1994) and Ziegler and Giambelluca (1997) on road-surface runoff.

Even fewer published studies have explicitly considered how road networks affect the routing of water through a basin. We therefore have little basis to evaluate the hydrologic functioning of the road system at the scale of an entire watershed or landscape. Few published studies to date have identified how roads in different landscape positions might influence the movement of water through a basin. Montgomery (1994) looked at the effect of ridgetop roads on channel initiation, and Wemple (1994) documented the magnitude of drainage network enlargement caused by roads in different slope positions.

Based on studies of small watersheds, the effect of roads on peak flows is detectable but relatively modest for most storms; insufficient and contradictory data do not permit evaluation of how roads perform hydrologically during the largest floods. Roads do not appear to affect annual water yields, and no studies have evaluated their effects on low flows. In some basins, roads produced no detectable change in flow timing or magnitude (Rothacher 1965, Wright and others 1990, Ziemer 1981), but in other basins, average time to storm peak advanced and average peak magnitude increased after road building for at least some storm sizes (Harr and others 1975, Jones and Grant 1996, Thomas and Megahan 1998). Helvey and Kochenderfer (1988) concluded that typical logging operations in the central Appalachians do not increase flows sufficiently to require larger culverts to accommodate them. Forest harvesting without roads in the southern Appalachians increased stormflow volumes by 11 percent and peak flow rates by 7 percent (Hewlett and Helvey 1970, Swank and others 1988). Harvesting an adjacent watershed with 4 percent of the area in roads increased stormflows by 17 percent and peak flows by 33 percent. Four years later, peak flows dropped to a 10-percent increase after 40 percent of the road system was closed and returned to forest (Douglass and Swank 1975, Douglass and Swank 1976). Collectively, these studies suggest that the effect of roads on basin streamflow is generally smaller than the effect of forest cutting, primarily because the area occupied by roads is much less than that occupied by harvest operations. Generally, hydrologic recovery after road building takes much longer than after forest harvest because roads modify physical hydrologic pathways, but harvesting principally affects evapotranspiration processes. The hydrologic effect of roads depends on several factors: (1) including the location of roads on hillslopes, (2) the characteristics of the soil profile, (3) subsurface water flow and ground-water interception, (4) the design of drainage structures (ditches, culverts) that affect the routing of flow through the watershed, and (5) proportion of the watershed occupied by roads.

Causes of Road-Related Problems During Floods

Most road problems during floods result from improper or inadequate engineering and design, particularly at road-stream crossings but also where roads cross headwater swales or other areas of convergent groundwater. Road redesign that anticipates and accommodates movement of water, sediment, and debris during infrequent, but major storms should substantially reduce road failures and minimize erosional consequences when failures occur.

Recent studies after large floods in the Pacific Northwest highlight the importance of plugged culverts and ditches in contributing to road-related failures (Donald and others 1996, Furniss and others 1997). A typical failure resulted from culverts sized only to accommodate the flow of water, but not the additional wood and sediment typically transported during major floods. The culverts became obstructed and diverted water onto the road surface, and onto unchanneled hillslopes or into neighboring drainages that were unable to adjust to the increase in peak flow from the contributing basin. Cascading failures were common, where diversion or concentration of flow led to a series of other events, ultimately resulting in loss of the road or initiation of landslides and debris flows.

Analysis of the probability of large floods and how they relate to the design life of roads indicates that most road crossings are likely to have one or more large floods during their lifetimes. Consequently, designing roads with large storms in mind is prudent and well within the reach of current engineering practices (Douglass 1977, Furniss and others 1991, Furniss and others 1997, Helvey and Kochenderfer 1988).

Decreasing the Negative Hydrologic Effects of Roads

Although the ability to measure or predict the hydrologic consequence of building or modifying a specific road network might be limited, general principles and models can be provided that may decrease the negative hydrologic effects of roads. These principles will be useful during upgrading or decommissioning of roads to meet various objectives.  A partial list of principles includes:

• Locate roads to minimize effects; carefully examine the geology of all proposed road locations.
• Design roads to minimize interception, concentration, and diversion potential; include measures to reintroduce intercepted water back into slow (subsurface) pathways by using outsloping and drainage structures rather than attempting to concentrate and move water directly to channels.
• Evaluate and eliminate diversion potential at stream crossings.
• Design road-stream crossings to pass all likely watershed products, including woody debris, sediment, and fish in addition to water.
• Consider landscape location, hillslope sensitivity, and orientation of roads when designing, redesigning, or removing roads.
• Design with failure in mind. Anticipate and explicitly acknowledge the risk from existing roads and from building any new roads, including the probability of road failure and the damage to local and downstream resources. Decisions about the acceptable probabilities and consequences of failures should be based on explicit risk assessments. The many tradeoffs among road building techniques to meet various objectives must be acknowledged. For example, full bench road construction may result in lower risk of fill slope failure, but it also may increase the potential for groundwater interception. Outsloping of the roadbed may reduce runoff concentration on the road surface but it also may increase driving hazard during icy or slippery conditions.

See also: 

• Drainages and the Hydrologic Cycle
• Designing Roads to Control Water