Most of the research on forest roads in the southern Appalachians has been conducted at Coweeta Hydrologic Laboratory in North Carolina. This research indicates that of all forest activities, road management poses the greatest risk to water quality, and that roads and skid trails are the primary sources of sediment from forestry-related activities. Decades of research results at Coweeta have been used to develop Best Management Practices (BMPs) for woods road construction in the Appalachians. BMPs deal with chemical pollution and increased water temperature, as well as control of erosion caused by roading, logging, and site preparation. Research at Coweeta has also provided options for treating problem roads.  Problem roads include those not built in the best locations, and those whose design and maintenance do not meet todays standards. Options for treating problem roads include relocation, reconstruction, and closure.

This section focuses primarily on the physical, ecological and socioeconomic effects of roads.  It does not address designing and installing new roads.  Forest road design textbooks are widely available (Walbridge 1997) and have so far not been synthesized as part of this Forest Encyclopedia.

Physical, Ecological and Socioeconomic Effects of Roads

Forest roads have many kinds of direct and indirect effects. The following sections, modified from "Forest Roads: A Synthesis of Scientific Information" by Gucinski and others (2001) review these various effects across the United States.

Rural roads are placed in three classes, which are distinguished by use level and physical condition (Swift and Burns 1999):

• Arterial roads are usually double-lane and paved or have an all-weather gravel surface. This class also includes collector roads, which are single-lane and usually have gravel to provide all-weather access. Arterial and collector roads may be open to traffic year round; they form a connecting transportation network that carries high volumes of traffic.
• Local roads generally are single-lane gravel or dirt roads that usually carry little traffic and may dead-end. They may be open to traffic but impassable in wet weather. Many are rough-surfaced and suitable only for high-clearance or all-wheel-drive vehicles. Local roads are often developed for access to timber sales.
• Orphan roads are abandoned or not maintained by any landowner, government entity, or group of users. They may be passable for off-road vehicles, if at all. They may receive only the minimum maintenance required for passage provided by infrequent users. Because of this lack of attention, sections of orphan roads commonly cause substantial adverse impacts to adjacent water resources.

This section deals exclusively with the economic effects of roads on National Forests.  Future expansion is planned to provide coverage for private and other public lands.

Both benefits and costs are associated with building, maintaining, and continued use of Forest Service roads. Likewise, benefits and costs are associated with removing existing roads. The issues revolve around whether the good things outweigh the bad things and what the extent of roads should be in National Forests.

Some economic activity is supported by building and maintaining roads: economic activity also is supported by decommissioning roads. Analyses for the 1995 RPA program suggest that about 33 jobs nationwide are supported per $1 million expenditure on building and maintaining roads (Alward and others 2000). A reasonable speculation might be that roughly the same rate of employment would be supported by removing existing roads and restoring the land underlying them. Road building and removal represent one-time stimuli to the economy, but maintaining roads is a recurring stimulus.

The major effects of roads on local economies, however, would be expected to result from the economic activity those roads support by providing access to the National Forest and to communities in or near it. On Forest Service roads, that activity includes logging, silvicultural operations, and recreation, among others. Also supported is economic activity that depends on recreation, such as guides, outfitters, and rafting permittees. The roads also provide access for land management and firefighting operations.

Indirect (and approximate) indications of the amounts of economic activity that might be associated with changes in Forest Service roads can be obtained from several sources. Reports indicate that timber harvest from National Forests supports about 16.5 jobs in the local area per million board feet harvested (USDA Forest Service 1996). That estimate is conservative because it is based on summed local-area models. Recreation use of National Forests supports from 1,000 to 2,000 jobs economy wide (nationally) per million trips, depending on the primary activity, based on analyses done for the 1995 RPA program (Alward and others 2000, Archer 1996). Use of public land, in general, follows roads. In Alaska, for example, intensity of use by both hunters and nonconsumptive wildlife users follows road corridors (Miller and McCollum 1997). Further, we hypothesize that more casual users such as scenery gazers, picnickers, car campers, and day hikers that constitute the bulk of National Forest recreationists probably stay closer to the road than do some hunters and backpackers, a minority of National Forest recreationists.

Whenever timber is cut and removed from the forest, roads will be needed; even helicopter logging at some point converts to road use by truck hauling. One issue is the quality of the roads and the length of their lives; that is, whether they are permanent and remain after timber harvesting ceases, or temporary and closed after harvest. Permanent roads are available for other activities, primarily recreation and management activities. Temporary roads are available for timber activity and some incidental activity during harvest, but when the roads are closed, benefits accruing from those roads cease. It is at least conceivable that the cost of maintaining a road over time sometimes outweighs the cost of removing it at the end of one timber harvest cycle and rebuilding it for the next one. Environmental effects (and costs) of multiple entries and decommissioning of temporary roads must be balanced against those of a single permanent road. Permanent roads cost more to build and maintain than temporary ones, with increased potential for degrading the ecosystem, but they can result in more benefits over longer periods than temporary roads because of the access they allow.

Roads affect spatial patterns of forest use. Changes in roads change those patterns. Recreational users are particularly attracted to or driven away from particular areas by the availability and ease of access. With decreased access to the National Forest, some users might drop out and give up outdoor recreation. Others would shift their use to other areas, some on Forest Service land and others off. The result would be reduced economic activity in the locale where forest access was decreased and increased economic activity in areas where displaced users moved. In general, the effects would be reversed if access were increased. Sometimes, however, increased access could lead to decreased use. For example, a new road and associated commercial activity could degrade a landscape for viewers.

Another result of spatial shifts in recreational use could be to concentrate use in areas to which displaced users move. Concentrated use may increase environmental effects as well as decrease the quality of peoples experiences. Crowding imposes costs on existing users by diminishing the benefits they received from their recreational use.

Anything that affects the demand for and benefits received from recreation and other uses of Forest Service land has subsequent economic effects, and it may alter development because land uses drive local economic activity. Forests and local economies will be affected differently, depending on the mix of local activities.

Building or removing Forest Service roads and maintaining existing roads can help mitigate ecosystem degradation associated with roads. Note that the tradeoffs are between the expense of minimizing or eliminating environmental degradation associated with Forest Service roads and the value of access to Forest Service land with associated economic activity. Many roads are or have been funded by the timber program. Benefits accrue from use of those roads for activities other than timber, largely recreation. This contrast presents a classic problem of joint cost allocation, and the accounting problem of attributing cost should not be used as an excuse for looking only at specific programs or components of the Forest Service mission.

The jobs and other economic activity supported by building and maintaining roads must be balanced against the cost of building and maintaining those roads, including costs resulting from choosing not to maintain selected roads. The question is: Do the benefits associated with the roads, both direct and indirect from all sources, justify the cost incurred by society, including costs of increased ecosystem degradation from deferred or inadequate maintenance?

The road-related issues associated with energy and mineral resources fall into three overlapping categories: access rights, property rights, and benefits and negative effects. These issues are a consequence of the inherent nature of the resources and their treatment under existing law. The defining characteristic of energy and mineral resources is nonrenewability; energy and mineral resources are finite, so extraction inevitably leads to resource exhaustion. Depleted deposits must be replaced either through domestic exploration and mine or field development or through importation. In many places, National Forest land is underlain by deposits of nonrenewable resources, some of which are privately held. Demand for access to many of these deposits is inevitable.

The extractive industries want, and have certain legal rights to, access to public land to explore for energy and mineral deposits. The access may be on existing forest roads or may require building of new roads. The Forest Service road system facilitates extraction of energy and mineral resources from public lands, which can benefit society. Mineral developments and oil fields in and of themselves can affect the environment negatively, such as by loss of habitat, increased noise, and added particulate emissions in the air and water, but these effects can be attributed only secondarily to roads; that is, without the road, mineral development might not have taken place.

Federal law and Forest Service policy clearly support exploration for and extraction of resources from public land. Leasable resources (that is, metallic minerals found on acquired land and all energy resources) are managed under the Mineral Leasing Act of 1920. Locatable minerals, primarily the metallic ones on public domain land, are managed under the Mining Law of 1872. Saleable minerals (that is, common varieties such as gravel) are managed under the Mineral Materials Act of 1947. These laws predate the National Forest Management Act of 1976 and the Multiple Use Sustained Yield Act of 1960.

Under the Mining Law of 1872, U.S. citizens and firms have the right to explore for and stake claims to selected minerals on all public domain land not specifically withdrawn from mineral entry. Claims are valid in perpetuity or can be converted to private property rights (that is, patented) assuming that appropriate legal requirements are fulfilled. The Forest Service cannot unilaterally deny exploration access to National Forest public domain land, but the agency does have the right to withdraw specific areas from further mineral entry. The agency cannot prevent staking of a claim on this land, and a claim holder is entitled to use the surface for activities attendant to exploring for, developing, and extracting minerals, within the limits set by Federal, State, and local environmental laws. The agency cannot block an otherwise legal patent (that is, deny a claim holder the right to convert the claim to private property). The Congress can, and has, placed a moratorium on new patents, but the moratorium could be lifted in the future. In any event, hundreds of thousands of patented and unpatented claims are already held within the administrative boundaries of the National Forests.

The Forest Service has considerably more control over the location of exploration and development activities for leasable minerals than it has for locatable minerals. For National Forests and Grasslands with completed oil and gas leasing Environmental Impact Statements, petroleum exploration activities are restricted to areas designated as appropriate in those documents. The regions also are taking an active role in directing access for leasable minerals. For example, the Northern Region is attempting to restrict oil and gas exploration to areas relatively near existing roads. This approach is not without potential for controversy, however. Decommissioning of roads could be perceived as a de facto withdrawal of the adjacent lands from exploration. The circuit courts are split on the question of whether failure to offer land for lease is tantamount to withdrawal.

The Forest Service is required by law to provide reasonable access to valid existing mineral rights, regardless of their form, whether unpatented claim, lease, or private property, as a patented claim or subsurface mineral right. An unpatented claim is an implied property right that can be held, sold, or inherited, and access is regulated under the Mining Law of 1872. Patented claims are private property, and access is regulated under the Alaska National Interest Land Conservation Act of 1980 (ANILCA). Coal, oil and gas, and mineral leases also offer a limited form of property right. The rights to individual energy and mineral resources may be held by different legal entities, and the mineral rights may be severed from the surface, which is termed a "split estate." Access to unpatented inholdings, patented claims, leases, and severed mineral rights can be restricted but seldom denied. Access may be by the existing road system or may require new roads. The Forest Service is neither required by law nor expected by industry to build or maintain energy and mineral access roads. Roads built for other reasons (for example, in support of recreation development) might be paid for by the Forest Service but also be used by a mining or energy firm. The firm is always required to maintain the road or to pay for road maintenance called for by its activities; it frequently pays through a reimbursement arrangement with the agency.

The Forest Service can affect the location and design of roads built on National Forest land to support energy and mineral activities. In addition, the agency can sometimes place stipulations on access by limiting road use to certain months, permitting aerial access only, or precluding surface occupancy. Constraints that are unduly expensive to fulfill or so restrictive as to make an otherwise economic mineral deposit uneconomic, however, might well be perceived as denying reasonable access. Temporary roads often are built to facilitate energy and mineral exploration activities. Building plans are subject to review and approval by the agency. If no discovery is made, the exploration firm may be required to obliterate the road. Alternatively, the road may be upgraded to permanent status, depending on the circumstances and legal authority. Public use of the road may sometimes be limited because a road condition acceptable to the mineral industry may be neither acceptable to, nor safe for, the general public. In addition, particularly for exploration, the agency may permit access that does not require roads, including access by helicopter, foot, horseback, and all-terrain vehicles.

The energy and minerals industries use the existing road system in exploration, development, extraction, and reclamation activities. Only a small portion of the entire road system is affected in any given year, but it is reasonable to assume that most roads will be used by the minerals industry over the long term. Designating a subset of the existing road system as having no future benefit to the industry is impractical because geographic targets for exploration and development change in response to technological advances and market fluctuations. Limiting mineral exploration access to areas where minerals have already been or are being extracted could preclude future discoveries. Road closures or decommissionings are controversial. Firms wanting to rebuild obliterated roads could face long delays because of the lengthy approval process now in place for building new roads. Such delays could disrupt multiyear exploration and development plans and financing. The energy and mineral resources produced from National Forests are essential to the manufacturing, farming, building, and power-generating industries. A value of $4.3 billion was estimated in 1995. Forest Service production represents only a small part of the total value of U.S. production, however. For example, the copper produced on National Forests represents only 1 percent of total U.S. copper production. Sometimes, however, production from National Forests is a significant percentage of domestic production; National Forests accounted for 80 percent of domestic lead production in 1995. Significant amounts of coal and molybdenum also are produced on National Forests. These contributions to the domestic economy are made possible by use of the forest road system.

Most of the concerns addressed here apply primarily to the western United States. In much of the East, road networks are well developed and relatively stable because of terrain and vegetation differences. Wildfire interactions are likely to be similar to those described for the West, but the effects are likely to be significantly less. In the Southeast, where use of prescribed fire is widespread, roads are frequently used as firebreaks. Much of this activity is on private land, but a high proportion of the road network is state and county highways rather than Forest Service roads.

The increasing density of road networks in and adjacent to many forest, shrub, and rangeland areas has been an important factor in changing patterns of disturbance by fire on the landscape. Roads provide access that has increased the scale and efficiency of fire suppression, and roads have created linear firebreaks that affect fire spread. These factors can be useful in both fire suppression and prescribed fire operations. In addition, road access has undoubtedly contributed to increased frequency of human-caused ignitions in some areas.

Roads Provide Increased Access for Fire Control and Suppression

That improved road access leads to increased efficiency and effectiveness of fire-suppression activities is a long-held tenet of fire fighting. Much of the effectiveness of past fire-suppression policies probably can be attributed to increased access for ground crews and equipment, particularly under weather and fuel conditions where fire behavior is not severe. Under the severe conditions associated with intense, rapidly spreading fires, the value of forest roads for access or as fuelbreaks is likely to be minimal. Although little has been published in the science literature to quantify these effects, a study in southern California concluded that the road network had been a key factor in determining what suppression strategies were used, both in firefighter access and in use for backfiring and burning-out operations (Salazar and Gonzalez-Caban 1987). Early studies of fuelbreak effectiveness in southern California came to similar conclusions (Green 1977). It is uncertain how much road access increases efficiency, but there is little doubt that there is an increase.

An important issue in the western United States is building new roads to allow harvest and prescribed fire to reduce fuel accumulations in ecosystems where past management (principally fire suppression and harvest) have increased the risk of large, severe wildfires (Lehmkuhl and others 1994). The principal concern here is the tradeoff between reducing the effects of wildfire and increasing the risks of road effects on aquatic habitat. In the Columbia Basin, scientists concluded that "it is not fully known which causes greater risk to aquatic systems, roads to reduce fire risk, or realizing the full potential risk of fire" (Quigley and others 1997). An adaptive management framework has been suggested for addressing this concern (Rieman and Clayton 1997). We currently have few data on how these processes might be affected by road networks, but a study after the 1987 Stanislaus fires in California suggests that cross-slope road networks reduced sediment delivery to debris basins (Chou and others 1994).

Roads Increase Access for Human-Ignited Fires

The benefits that roads provide for fire prevention and fire management carry an associated cost. Roads have increased the frequency of human-caused ignitions, particularly in areas of expanding urban and rural development into wildland interfaces (Hann and others 1997). The high rate of human-caused fires in the Blue Mountains of eastern Oregon is associated with high recreational use in areas with high road densities (Hann and others 1997). In the Southwest, however, numbers of ignitions are less important in fire management than are fuel loadings and climatic conditions are (Swetnam and Baisan 1996). Nevertheless, numbers of ignitions are important determinants of fire risk (Conard and Weise 1998).

Roads Affect Fire Patterns and Fire Regimes

Road networks have changed fuel patterns and fire regimes at the broad scale. If we accept that road networks have been important in effectively suppressing fire and that they alter fire patterns on the landscape, then road systems are, in some sense, linked to changes in fuel patterns and fire regimes. Before fire-suppression activity in the western United States, fuel loads remained relatively low in dry forest types; high fuel loads were restricted to small, isolated patches (Agee 1993). As access increased, areas burned by wildfire declined, at least through the 1960s. As a result of suppression supported in part by access, fuel accumulations increased and areas with moderate to high fuel loadings became larger and more contiguous. This pattern of change has been documented for the entire upper Columbia River Basin. There, scientists assert that fire suppression has generally been more effective in roaded areas, causing those areas in the upper basin to depart further than unloaded areas from unaltered biophysical templates (Hann and others 1997). Roads, along with other human disturbances such as clearcutting, contribute to new disturbance patterns at the landscape scale, both by increasing efficiency of fire fighting and providing barriers to firespread that are different from natural barriers (Swanson and others 1990). Increased emphasis on removing roads in certain environmentally sensitive areas will reduce access for fire suppression and prescribed fires, potentially leading to increased fuel accumulation and increased fire hazard in some areas.

In general, the existence of roads seems to have little effect on forest tree diseases, but there are some examples where building or using roads has caused significant local effects. These problems, where they exist, appear to be specific to the pathogen, host, and site. Nearly always, the negative effects can be ameliorated through simple modifications in how roads are built and used.

Building and maintaining roads may exacerbate root diseases. Wounded trees and conifer stumps created and not removed during road building provide infection courts for annosus root disease, which may then spread through root contacts to kill a patch of trees (Otrosina and Scharpf 1989). Trees can be damaged or stressed by road building through wounding of stems and roots, covering of roots with side castings, or compacting of soil over roots. These trees become susceptible to various tree diseases. Armillaria root disease is an example. In deciduous stands only injured trees are attacked but in coniferous stands an injury can initiate a pocket of disease involving several to many trees (Shaw and Kile 1991). Oak decline is associated with poor sites, older stands, and road building or other disturbance (Wargo and others 1983). Black stain root disease (Leptographium wagneri) attacks conifers stressed by disturbances, especially compaction caused by road building; in pinyon pine (Pinus monophylla), it is associated with roads and campsites (Hansen 1978, Hansen and others 1988, Hessburg and others 1995). Droopy aspen disease is associated with road building and compaction (Jacobi and others 1990, Livingston and others 1979). Sap streak disease in sugar maple is associated with compaction from roads and from direct injury to trees (Houston 1993).

Roads indirectly contribute to disease spread by giving people ways to transport diseased material long distances. New pockets of both oak wilt and beech bark disease (Houston and OBrien 1983) may have resulted from moving firewood from the forest to a homesite (Appel and others 1995, Rexrode and Brown 1983). Pitch canker (Fusarium subglutinans) was recently reported on Monterey pine (Pinus radiata) in California; previously, it had known to cause littleleaf disease on slash pines in the South. A single introduction is thought to have been responsible for the disease occurence in California (Correll and others 1992, Storer and others 1995). Campers who use roads to get to remote sites in Colorado and other states have caused significant mortality by carving on aspen and birch trees.  The wounds provide pathways for various fungi that cause cankers and quickly kill the trees. Many trees are unintentionally damaged, for example, when campers hang a gas lantern on a branch too close to the trunk of a tree, thereby causing heat damage.

Road building also may set the stage for an insect attack that further stresses the trees, creating conditions for a disease outbreak (Boyce 1961).

One benefit of roads is to provide access for silvicultural activities that limit the damage from tree diseases (Bull and others 1997). Road building can be planned to help reduce the spread of some forest tree diseases: mistletoe is spread by the forcible ejection of the mistletoe seeds. In young plantations or pole-sized stands, roads can subdivide an area to prevent mistletoe seeds from reaching a healthy stand (Hawksworth and Wiens 1996). In Texas, roads could be planned to separate a portion of a stand with oak wilt from healthy trees. The act of building the road (if extensive enough) severs root connections and prevents tree-to-tree movement of the pathogen (Appel and others 1995, Rexrode and Brown 1983). In other areas, new or established roads may have the unintended effect of breaking the continuity of host roots and thus halting the spread of laminated root rot (Phellinus weirii) and other root diseases (Hadfield 1986, Thies and Sturrock 1995).

Among the benefits that roads provide is access for research, timber and non-timber forest inventories, and monitoring. Although the economic scale of these tasks is low compared to some other activities, the effects of access on inventory and monitoring are not trivial, and the need for these activities is obvious.

In the region, contracting cost for inventories run about $600 per plot when roads allow access to within 0.25 mile of the plot. In the same region, cost rises to $1,300 per plot in roadless areas open only to foot access. In the Pacific Northwest, survey costs were $1,460 per wilderness plot and $1,174 per nonwilderness plot. In Alaska, roadless areas are vast, but helicopter access is permitted. The average cost per plot for roadless areas in interior Alaska has averaged $4,000 per plot for 170 plots.

Roads fragment habitats by changing landscape structure and by directly and indirectly affecting species. Habitat effects of roads on the landscape include:

• Dissecting vegetation patches.
• Increasing the edge-affected area and decreasing interior area.
• Increasing the uniformity of patch characteristics, such as shape and size (Reed and others 1996).

Whenever forest roads are built, changes in habitat and modified animal behavior will lead to changes in wildlife populations (Lyon 1983). Large mammals such as elk (Cervus canadensis), bighorn sheep (Ovis canadensis), grizzlies (Ursus arctos horribilis), caribou (Rangifer tarandus), and wolves (Canis lupus) avoid roads. Avoidance distances of 300 to 600 feet are common for these species (Lyon 1985). Road usage by people and their vehicles has a significant role in determining road avoidance by animals. In a telemetry study of movement by black bears (Ursus americanus), they almost never crossed interstate highways, and they crossed roads with little traffic more frequently than those with high traffic volumes (Brody and Pelton 1989). Bobcats (Lynx rufus) crossed paved roads in Wisconsin forests less than expected, possibly to minimize interactions with vehicles and people (Lovallo and Anderson 1996).

A few studies have related genetic changes in populations simply to the presence of roads (Forman and others 1997), but the distribution of roads in the environment also must be considered. Road density is a useful index of the effect of roads on wildlife populations (Forman and others 1997). Wolves in Wisconsin are limited to places with pack-area mean road densities of 0.7 mile/square mile or less (Mladenoff and others 1995). Some studies have shown that a few large areas of low road density, even in a landscape of high average road density, make suitable habitat for large vertebrates (Rudis 1995).

See also:  Effects of Roads on Habitat

According to the 1995 draft RPA program, about 46.2 million acres of National Forest land are considered suitable for livestock grazing. Producing livestock can be an important part of local economies, and livestock grazing is deeply rooted in the culture of the American West. Grazing was first authorized on National Forest land by the Organic Administration Act of 1897 and confirmed by many later appropriations acts (USDA Forest Service 1989). The Public Rangelands Improvement Act of 1978 reinforced a national policy that public rangelands were to be "managed...so that they become as productive as feasible for all rangeland values."

Essentially no scientific information exists that analyzes the ecological, administrative, or economic effects of roads on administering the Forest Service range-management program. Preliminary unpublished analyses from the interior Columbia River Basin ecosystem management project addressed the road issue from the perspective of ecological responses to the presence or absence of roads. The analyses found correlations between changes in vegetation composition, riparian functioning, and fire regimes and the presence of forest roads. They could not conclude any cause-and-effect relations from these correlations, however. The program also found higher road densities to be associated with diminished ecological integrity, including those based on range criteria.

To assess the importance of National Forest roads for administering the grazing program, as well as their economic value to permittees, an ad hoc interdisciplinary team was formed to provide a nominal assessment. However, the results of the interdisciplinary-team assessment are heavily weighted towards the Rocky Mountain Region (Colorado, Kansas, Nebraska, South Dakota, and eastern Wyoming) and thus may not represent a national perspective. The findings below reflect the input of the team:

• Roads on national forest lands have both positive and negative effects on rangelands and the administration of the grazing program. Roads have mostly replaced driveways as a means for transporting sheep and cattle to and from mountain allotments. As a result, these driveways have dramatically improved in rangeland health. Until the 1970s, livestock driveways were considered "sacrifice areas" in the range-management discipline (Stoddart and Smith 1955). Thus, national forest roads can promote ecosystem management objectives along alternative transportation corridors, which they replace.
• Roads in national forests are essential for administering the grazing program, allowing timely access to allotments. Compliance enforcement was mentioned in particular as an activity greatly benefiting from forest roads. The principal reasons cited were that agency downsizing has resulted in high workloads for remaining range conservationists, which does not allow them sufficient time to carry out their duties; guard stations have been closed; Forest Service personnel no longer have the option of spending nights in the field in some places; and many allotment plans incorporate Forest Service roads into their approved grazing system or as drive-ways to and from the allotment; for example, in the Black Hills, all driveways are along roads.
• Roads can reduce operating costs by providing motorized access to allotments. The team estimated that, if all National Forest roads were closed, permittee costs would increase by three to five times. These costs would accrue from increased riding time, cost of horses and riders, and added equipment costs (such as horse trailers). The grazing program derives benefit from only part of the road system, however, and if arterial and collector roads remained open, the expected cost increases would be less, ranging from none to a twofold increase.
• Roads can heighten conflicts among users of national forests, such as cattlemen and recreationists, although some evidence shows that concerns about road conditions actually can cause some forest visitors to slightly, but measurably, shift their focus of attention from grazing encounters to roads (Mitchell and others 1996).

Roads fragment habitat and create habitat edges, favoring species that use edges. Edge-dwelling species generally are not threatened, however, because the human-dominated environment has provided ample habitat for them. Road building introduces new edge habitat in the forest. The continuity of the road system also creates a corridor by which edge-dwelling species of birds and animals can penetrate the previously closed environment of continuous forest cover. Species diversity can increase, and increased habitat for edge-dwelling species can be created.

Roads and their adjacent environment qualify as a distinct habitat and have various species, population, and landscape-scale effects (Baker and Knight 2000, Dawson 1991, van der Zande and others 1980). Some research has attempted to describe habitat modifications caused by roads, but most of this work is species and site specific (Lyon 1983). Surveys of songbirds in two National Forests of northern Minnesota found 24 species of birds more abundant along roads than away from them (Hanowski and Niemi 1995). Close to half these species were associated with edges, including birds like crows (Corvus brachyrhynchos) and blue jays (Cyanocitta cristata) that use roads as corridors to find food. Turkey hens (Megapodiidae) in North Carolina nested near closed and gated logging roads and used them extensively in all stages of brood development (Davis 1992). One study showed that habitat in the road-side right-of-way supports a greater diversity of small mammals than do adjacent habitats (Adams and Geis 1983), but this finding may not apply to forest roads with only narrow cuts and fills on either side.

The similarity between forest roads and transmission-line rights-of-way may be important in assessing the contribution of roads to habitat. Studies have shown that wide transmission-line corridors support grassland bird communities of species not found in the forest, and narrow corridors produce the least change from forest bird communities (Anderson and others 1977). The same study notes that increasing edge diversity of birds, for instance, may negatively affect abundance of interior species (see Roads and Biological Invasions).

See also:  Roads and Fragmentation

In addition to satisfying the American penchant for sightseeing by car and other forms of recreation requiring auto travel, roads and their features themselves sometimes have heritage value because of historic significance or architectural features. Roads also may affect areas considered sacred by American Indians or other religious groups. These issues can affect the legal and political framework for Forest Service road policy and management because important historical, social, and cultural values are often part of developing, maintaining, or decommissioning roads. Forest planning for transportation and for individual roads should incorporate information on heritage and cultural values for both roaded and unroaded areas.

Cultural Value of Roads

Roads and associated features are part of the history of the nation. Some features are significant for their association with exploration and settlement, others for accomplishments in engineering, and still others for reasons of local history and culture. Roads and other transportation features figured prominently in the early nonindigenous settlement and development of the nation. Roads that were or are significant in this way include early Spanish roads, such as El Camino Real (the Royal Highway) in California and New Mexico; those that follow the routes of American Indian trails (Davis 1961); military roads such as Cooks trail, which crosses the forests of northern Arizona (Scott 1974); and some early routes established for commerce, such as the Santa Fe Trail, which crosses the Cibola National Forest. Given their historical role, such roads (many still in use) often are eligible for the National Register of Historic Places. Of equal importance, historic roads often have special meaning to people who live near them or have used them. Route 66, for example, which crosses the Kaibab National Forest, is considered historically valuable for its role in establishing regular, all-season east-west automobile transportation to California (Cleeland 1988, Cleeland 1993).

Features forming part of or associated with a road may be historically or culturally valuable for their own merits (Fraser 1987). Bridges and other features built by the Civilian Conservation Corps often are fine examples of engineering and considered eligible for the National Register of Historic Places (Throop 1979). Many such bridges are on Forest Service roads. Roads also may have heritage value as part of a cultural landscape, such as the landscapes associated with homesteading, ranching, or logging. Even roadside advertising can have local cultural significance, such as the hand-painted message along an abandoned highway in the Cibola National Forest that claims "Curandera cures all." The National Park Service and the U.S. Committee of the International Council on Monuments and Sites recognized the heritage value of transportation corridors in a conference held in 1993 (USDI 1993).

Effects of Roads on Cultural Sites

Building, maintaining, and decommissioning roads can affect historical and cultural values. Roads often directly affect historical and archaeological sites. Building, maintaining, or decommissioning roads can damage or destroy archaeological sites (Spoerl 1988) with earthmoving equipment used on buried and surface remains, such as structures and other cultural materials. Roads also affect sites indirectly by increasing erosion or by making sites accessible to vandals. Less tangibly, but no less important, roads often affect areas that American Indians consider sacred, may limit their ability to conduct ceremonies that require privacy, and may even diminish the sacred qualities of such places.

See also:  The Cultural Landscape

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 

The design and construction of, and soil loss from, forest roads have been researched by Coweeta Hydrologic Laboratory since the 1930s.

Roadbank Stabilization Tests: 1934-1958

Early roadbank stabilization tests at Coweeta were aimed at using natural materials and labor-intensive methods to stabilize eroding slopes. These tests involved trials using mulch, poles, planted grass, or natural vegetation to slow erosion on exposed areas. Results showed that shrubs and trees provided relatively permanent stabilization (Hursh 1945, Swift 1988).

Exploitative Logging Demonstrations: 1941-1956

In 1940, Coweeta demonstrated the effects of exploitive logging and poor land management practices. Most of the roads constructed for this demonstration were adjacent to streambeds, skid trails and spur roads were often steep, logs were skidded downslope, and little effort was made to divert storm waters off the roads or to vegetate disturbed soil. These practices severely degraded water quality. About 408 m³ of soil were lost from each kilometer of road length (860 cubic yards/mile) (Lieberman and Hoover 1948). Sediment concentrations peaked at 5700 ppm during a storm in 1947. Observers concluded that watershed damage had little to do with the poor silviculture of exploitative logging, but was principally due to road design and methods used to remove logs from the woods (Swift 1988).

Integrated Forest and Watershed Management Demonstration: 1954-1955

In the mid-50s, Coweeta began to demonstrate road-building techniques that could be used to protect water quality. In this demonstration, skid trails were not permitted, logs were winched to roads, and downhill skidding was discouraged. Contour roads crossed streams at right angles, streams were carried through corrugated metal pipe, and open-topped culverts or narrow water bars were used for surface drainage. Roads were relatively narrow (10 feet/3 m wide) and slightly outsloped without an inside ditch.  Roads were seeded with grass after logging was completed. This road system met water-quality goals, but high maintenance costs and high initial costs discouraged acceptance by managers and loggers (Swift 1988).

Management Tests: 1956-1960

The cost and practicality of the Coweeta road design were tested on National Forest ranger districts. In one district, loggers found the road costs acceptable because savings in equipment maintenance and higher work efficiency compensated for the higher initial construction investment (Swift 1988).

Multi-Resource Management Demonstrations: 1962-1964

In the 60s, Coweeta began studies to demonstrate the concept of multiple-use management in which all resources in the watershed were to be made available. Long-term access to the entire basin was needed. Goals were: (1) to improve earlier designs so that maintenance requirements such as frequent cleaning of narrow-based water bars could be reduced, and (2) to demonstrate that timber access roads are permanent investments and not temporary expedients. A solution was the broadbased dip, a design feature that has become a part of nearly every set of forest road guidelines in the eastern United States. The broad-based dip is a gentle roll in the centerline profile of a level or climbing road. Where the roadbed can be drained by outsloping and broadbased dips, the problems associated with inside ditches can be avoided. Also, dips permit use of vertical banks, which are less expensive because less right-of-way clearing is required, less soil is moved, and smaller fills are created. At Coweeta, vertical cuts up to 6 1/2 feet (2 m) high have stabilized naturally on moist, fertile sites (Swift 1988).

Transportation Planning

One outgrowth of the multiresource management demonstration at Coweeta was a realization that long-range planning of a forest transportation system should include intermittent-use or local roads along with fully engineered forest development roads. With the recognition that even the lowest class of road could be a permanent capital investment came the understanding that planning was necessary to assure that each mile of road was constructed at the best possible location (Swift 1988).

Best Management Practices

The knowledge of road design accumulated by years of studies at Coweeta contributed strongly to the development of Best Management Practices (BMPs). Although BMPs deal with chemical pollution and increased water temperature, the greatest effort is directed at erosion control caused by roading, logging, and site preparation. Almost without exception, BMP guidelines for forest access roads include design features based on Coweeta experience (Swift 1988).

A comprehensive understanding of the economic effects of roads in National Forests must include both effects that can be measured in dollars (market effects) and those with no direct dollar values (nonmarket effects). The influence and importance of market values to land management decisions is obvious, and measuring and comparing effects of management decisions that affect market values are relatively simple. For example, the cost of building and maintaining a road into a forest can be readily compared to the income generated from harvesting the timber accessed by that road. Also important, but far more difficult to measure and compare, are the things people care about for which no market exists, such as access for hunting, bird watching, and wilderness experience.

Economists generally classify nonmarket values as either active or passive. The term "active-use value" applies to goods and services used in some activity like recreational fishing, skiing, or camping. The term "passive-use value" includes two categories (Peterson and Sorg 1987, Randall 1992): (1) things people appreciate without actually using them or even intending to use them (like a distant wilderness or an endangered plant or animal) are called "existence values"; and (2) things people want to remain available for others (such as their descendants) to use and appreciate are called "bequest values." Natural resource economists have invested much effort over the last several decades to develop and test methods for estimating nonmarket values. The methods can produce useful information, but the analyses are costly and their validity has not yet been demonstrated sufficiently to satisfy many economists (Arrow and others 1993, Cambridge Economics 1992, Mitchell and Carson 1989, Portney 1994).

People assign passive-use value to natural resources, especially roadless areas and natural areas with unique characteristics. And the passive-use value often exceeds the active-use value of road access (Bengston and Fan 1997, Brown 1993, Driver and others 1987, Driver and others 1996, Payne and others 1992, Walsh and others 1984, Walsh and others 1990). Building roads in roadless areas may reduce passive-use value significantly; decommissioning roads may increase such value.

Decisionmakers need to consider all these tradeoffs. Individuals and affected groups often disagree aggressively about the passive-use value of specific roaded and roadless areas and the effects of building or decommissioning roads (Bengston and Fan 1997). Thus, questions of equity must be considered: Whose desires should the Forest Service fulfill when stakeholders values conflict? What criteria should be used to decide among them? What approaches can be taken to resolve the conflict?

Forest Service officers responsible for road policy and management need to know various things: (1) the forest landscape conditions to which people assign passive-use or other nonmarket values, (2) how such values differ among individuals and groups of people, (3) the strength or significance of the value assigned, (4) how changes in the landscape affect the nonmarket values, and (5) how such values trade off with other forest-related values assigned by affected people. Under regulations of the Comprehensive Environmental Response, Compensation and Liability Act of 1980 (CERCLA), as amended, 42 U.S.C. 9651 (c), a United States Court of Appeals for the District of Columbia ruled in 1989 that passive-use values "...reflect utility derived by humans from a resource and thus, prima facie, ought to be included in a damage assessment." Thus, if Forest Service roads significantly alter passive-use value, whether positively or negatively, such value needs to be considered in road policy and management decisions. Failure to include these nonmarket values in an economic evaluation, when such values are judged to be important, could lead to inefficient and unfair allocation of resources.

The effects of roads on passive-use values differ by location and circumstance. Differences in the quality and uniqueness of landscapes modify the effects on passive-use value of building or decommissioning roads. The relation between supply and demand also affects the extent and strength of a passive-use value. For example, if many substitutes for a given roadless landscape exist, building a road in that landscape may have little or no effect on its passive-use value, just as the hunters killing of a single elk does not reduce the passive-use value of elk if the species is still abundant. Likewise, if an abundance of roads are provided to resources that people want for active use, decommissioning or closing one road will have little effect. People with strong attachments to a special place, use, or road may suffer loss, however, unless they can find and adapt to a substitute.

Validly and reliably measuring changes in passive-use and other nonmarket values are costly and can sometimes exceed the cost of being wrong. Managers of National Forest roads, however, must understand such values and the circumstances under which they are significant decision factors, to assure that the values can be included where appropriate. A survey-based method called contingent valuation can be helpful. Contingent valuation generally uses surveys or interviews to determine how much people say they would be willing to pay for some nonmarket good. Peoples willingness to pay for nonmarket values can provide a useful indication of their relative magnitude. Applying the method to passive-use value of public goods is, however, may produce flawed results, criticism, and controversy. Studies must be designed and applied carefully, and the results must be interpreted cautiously.

Other methods, such as value juries (Brown and others 1995), focus groups, public hearings, and other forms of public participation also can provide useful information. Quantitative measures should be taken only when the scale of the problem justifies sufficient investment for scientifically rigorous results. If fully and correctly disclosed, the cost of opportunities foregone by preserving a roadless landscape can serve as the price to be paid for the values served by preservation. Preserving a roadless area may sometimes cause an opportunity cost in the form of alternative uses foregone, such as timber harvest, developed recreation, or fire suppression. If the opportunity cost has been fully disclosed to the decisionmaker, a decision to preserve a roadless landscape is a policy acknowledgment that the value created exceeds the opportunity cost. In a decision about whether to designate an area as roadless, opportunity cost can sometimes serve as the price to be paid for whatever values, including intangibles, are served by the designation. Stakeholders and decisionmakers can then decide by judgment, negotiation, or analysis whether the gain is worth the price (Bell 1996, Fight and others 1978, Fight and others 1979, Randall and others 1979).

See also:  Economics of Roads

Consumer forces, changing social climate, and expanding global markets are contributing to the increasing development of nontimber forest products such as medicinals, botanicals, decoratives, and natural foods. These products are becoming viable economic options for sustaining rural communities. Ginseng (Panax quinquefolius), goldenseal (Hydrastis canadensis), coneflower (Echinacea angustifolia), and St. Johns wort (Hypericum perforatum)-- all plants found on National Forests-- are major contributors to a multibillion-dollar herbal and botanical industry. How roads will affect the survival and sustainability of nontimber forest products and how access to nontimber forest products will be influenced are important issues.

Plants are harvested from the wild by local residents or contract crews brought in from elsewhere. Particularly for the local harvesters, who operate under permit systems, access by road to the resource is a critical cost factor. Market value of nontimber forest products is related to cost; increasingly difficult access as plants become scarce may be factored into market value.

Maps indicating roads that offer access to nontimber forest products often act as a means of pinpointing the desirable harvesting areas. For example, in the special forest products inventory (Karen Theiss and Associates 1996) created for Trinity County, California, roads were used extensively to describe how to find areas where wildcrafters could harvest a particular species.

Habitats and plant community structure of some commercially harvested species are linked to roads. Roads create openings important to maintaining diverse species in abundance. From an assessment of 45 commercial species in Oregon, 30 percent can be found in openings and along roadsides. It also is well known that certain species require undisturbed mature forest and would not benefit from the gaps and disturbance caused by roads. Because of the specific habitat requirements of, for example, wild ginger, pitcher plants, and shade-loving mosses, roads would not directly benefit these plants. Some of these species are listed as sensitive, and ready access threatens their survival.

Illegal collection of nontimber forest products is considered a problem in many areas. Roads play a role in illegal taking, as well as in monitoring harvest activities.

A special forest products inventory created for Trinity County, California, suggests that harvesters stay away from roadsides because some Bureau of Land Management and Forest Service districts routinely spray herbicides and pesticides.

Generalizing the need for roads or road decommissions for nontimber forest products is impossible. Some populations of harvestable species benefit from the disturbance caused by building and maintaining roads, and other populations are harmed. Although enforcement of illegal harvest might be hampered, so would legal harvest. But market adjustments for reduced harvest (product scarcity) are unpredictable, and whether any increased value would be transferred to the harvester is not known.

Management of most nontimber forest product species would benefit from information and models that predict regional and general effects from building or closing roads. Information is needed on the economic effects of roads on various components of the industry from harvesters overhead to product price. These questions must be answered to determine how building or decommissioning roads affects the sustainability of individual commercial species and hence the sustainability of the economies reliant on them.

Forest roads create corridors by which predators, especially people, can enter the forest environment and affect wildlife populations.

There is strong evidence that human predation, either legally in game management programs or illegally, is greatly facilitated by roads and can significantly affect populations of animals. High road densities are associated with a variety of negative human effects on several wildlife species (Brocke and others 1988). People directly affect snakes by collecting, harassing, and killing them (Wisdom and others 2000). Increases in illegal hunting pressure, facilitated by roads, also negatively affect populations. Moose, wolves, caribou, pronghorn antelop, mountain goat, and bighorn sheep are particularly vulnerable to this kind of predation (Lyon and Marzluff 1985, Wisdom and others 2000).

Roads may increase nest depredation of song-birds by predators attracted to edges. Evidence for edge effects, however, is highly variable (Paton 1994). Although evidence has been found for local edge effects in cowbird parasitism and nest depredation, their effects on bird populations are not documented. Geographic location and large-scale patterns in the amount of forest and nonforest habitats may be more important in determining the reproductive success of forest songbirds (Donovan and others 1997, Robinson and others 1995).

Forest carnivores apparently travel on roads in winter when snow is deep, and thus the road system alters and enhances their ability to move (Paquet and Callaghan 1996). Wolves and grizzly bears are two key species that have chronic, negative interactions with people, and roads are a key facilitator. Repeated, negative interactions of these two species with people increase mortality of both species and often cause high-quality habitats near roads to be population sinks (Wisdom and others 2000).

See also: Fragmentation, Roadkill, Recreation.

Little science-based information is available on the relations between roads and private inholdings. Therefore, the following propositions are offered as hypotheses based on judgment, rather than scientific findings.

• The Forest Service is required by law to permit access to private inholdings.
• The Forest Service can require private inholding owners or lessees to comply with official regulations and standards that apply to building roads on or through National Forest land. The regulations and standards are documented in writing as official policy, but they are subject to interpretation and application in specific cases by agency line officers.
• The Chief of the Forest Service may consult appropriate National Forest officers about the sources of scientific documentation used in practice and in official regulations, standards, and procedures applicable to roads on or through National Forest lands that provide access to private property.
• In general, the scientific documentation of ecological and human effects of roads on or through National Forest land provided elsewhere in this synthesis applies to roads that provide access to private inholdings.
• No scientific basis exists for stating propositions about whether the Forest Service subsidizes access to private inholdings or the effect, if any, of Forest Service roads on the various values of private inholdings.
• The Chief needs inventory information about the type, number, acreage, location, use, value, and so on of private inholdings on National Forest land and the extent to which private inholders use National Forest roads for access. At present, no systematic inventory procedure or documentation can provide comprehensive and valid information of that type.

These propositions do not necessarily apply to inholdings dedicated to mineral and energy exploration or extraction, which are covered in Energy and Mineral Resources.

Almost all types of public recreation on National Forests depend in one way or another on roads for access. Whether, when, and where various recreational uses occur depend on access to, and the extent and location of, the road system. Altering this system is likely to have widespread and differing effects on different types of uses. The relations between roads and recreation on National Forests are highly complex and include many direct, indirect, and secondary links that are not well understood. Indirect evidence and related research provide the following insights and hypotheses:

• Roads provide corridors of access to a variety of National Forest sites, settings, and viewing opportunities for widely diverse users. Almost all recreation use in National Forests depends to some degree on road access. Sightseeing, driving outdoors for pleasure, and developed camping are examples of activities that directly use roads as a part of the recreation experience. Backpacking, white-water boating, and birdwatching are examples of activities that usually occur away from roads, but roads provide access to jumping off point for those users. Altering road systems can disrupt long-established access and use patterns and, at least in the short run, result in not meeting visitors expectations. Less road mileage or maintenance can lead to uneven shifts in recreation opportunities.
• Roads provide staging access to remote areas and wilderness, but the presence of roads can also reduce opportunities for solitude and perceptions of wildness.
• As demands for forest recreation opportunities grow locally, regionally, and nationally, even a stable amount and condition of forest roads likely will result in increased congestion, lowered satisfaction, and user conflicts. Participation in outdoor recreation has grown strongly across a wide spectrum of activities and segments of the American public (Cordell and Bergstrom 1991). Projections show this growth is likely to continue well into the future for all nature-based activities except hunting (Bowker and others 1999). At the same time, access to private land is continuing to decrease and be limited to lessees and friends of the owners (Cordell and others 1999). Public lands are likely to be the destinations of choice for increasing numbers of people looking for high-quality experiences in natural settings. Several National Parks already have limited motorized access to bus tours or other public transportation as one way to address increased congestion from private cars. Continued growth in demand without increases in road systems or limits on use of private cars likely will lead to lowered satisfaction and more conflicts at the more popular National Forests (Tarrant and others 1999). Changes in satisfaction likely will differ significantly by setting (Tarrant and others 1999). Direct recreation access, the character of and access to scenic views, and provision of increasingly sophisticated visitor services (including rescue and medical services) will depend on the character of the road system in place.

For the most part existing databases and literature have only indirectly addressed hypotheses that deal specifically with the relations between roads and recreation (for example, Knight and Gutzwiller 1995). Substantial research is needed to better understand direct and indirect relations between road-system characteristics, recreational use, and ecosystem conditions, including issues such as the introduction of exotics, soil erosion, habitat fragmentation, forest-product harvesting, wildlife disturbance, riparian vegetation, and fire.

An estimated 1 million vertebrates a day are killed on roads in the United States (Lalo 1987). However, these documented roadkill rates are significant in reducing populations of only a few rare species in North America, and these kills generally are on high-speed highways (Forman and others 1997).

Studies show that the number of collisions between animals and vehicles is directly related to the position of the nearest resting and feeding sites (Carbaugh and others 1975). Because most forest roads are not designed for high-speed travel, relatively few large mammals are killed by vehicles on forest roads (Lyon and Marzluff 1985). Exceptions are forest carnivores, which are especially vulnerable to road mortality because they have large home ranges that often include road crossings (Baker and Knight 2000).

Forest roads pose a greater hazard to small, slowly moving, migratory animals, such as amphibians, which are highly vulnerable as they cross even narrow forest roads (Langton 1989). Nearly all species of reptiles use roads for cooling and heating, so many of them are killed by vehicles. Highways and other roads that support moderate- to high-speed traffic are population sinks for many species of reptiles, and result in reduced and increasingly isolated populations (Wisdom and others 2000). Predators and scavengers are killed while they feed on road-killed wildlife, as are other species attracted to roads because of salts or vegetation, or because roads facilitate winter travel (Baker and Knight 2000).

Despite the large body of data documenting annual roadkill, little research has focused on how to mitigate the effects on wildlife populations.

See also: Habitat Fragmentation and Predation

Forest roads can significantly reduce site productivity by removing and displacing topsoil, altering soil properties, changing microclimate, and accelerating erosion. The direct effects of taking land out of production by removing trees and displacing soil has been estimated to range from 1 to 30 percent of the landscape area in managed forests (Megahan 1988). Studies in eastern U.S. forests have consistently found that 4 to 5 percent of the total forested area is taken out of forest production by building roads during logging operations. Although more than 50 percent of this area may be reforested within 8 years, growth rates and productivity are reduced. Total road length required to support logging operations depends on the harvesting and silvicultural system and topographic configuration, but the area disturbed may be surprisingly consistent (Douglass and Swift 1977, Robinson and Fisher 1982, Swank and others 1982, Swift 1988).

Causes of Reduced Rree Growth Near Roads

Trees can grow on any portion of a closed road, but they can grow only on cut and fill slopes on open roads. Sites are harshest and soils are poor or nonexistent on road cuts and the cut portion of roadbeds. Measurable declines in tree growth are common where soil is excavated to build the road crown. Causes for these declines may be a combination of the following factors:

• Changes in soil physical properties. Road building changes soil physical properties including depth, density, infiltration capacity, water holding capacity, gas exchange rate, nutrient concentrations, and microclimate. Fertile topsoils, often containing most of the organic matter and plant nutrient capital of a site, frequently are buried under road fills or sidecast and may be rendered inaccessible to plant roots.
• Nutrient loss. Loss of nutrient capital is inevitable with soil disturbance from road building (Swanson and others 1989), but isolating this effect from other site changes has proved difficult. An indirect indication of nutrient loss is the marked growth response of plants on road fills after fertilizer is applied. Fertilizer applied to a granitic road fill in Idaho increased growth of vegetation by 32 to 116 percent (Megahan 1974a), but such increases are not documented after fertilizer is applied on undisturbed soils.
• Increased erosion. Both surface and mass-erosion rates increase after road building, and roads often accelerate erosion on the slope below. Downslope damage generally is associated with mass erosion when a landslide originates from a road and causes scour on lower slopes or gullies related to concentrated road drainage (Megahan 1988). This problem is widespread on steep slopes of the Pacific States and in the northern Rocky Mountains (Burroughs 1985, Swanson and others 1981), but Megahan (1988) estimates that productivity is reduced on only about 0.3 percent of forested land at a broad scale. The persistence of these effects may range from decades (Ice 1985) to more than 85 years (Smith and others 1986).
• Soil compaction. Roadbeds are highly compacted compared to natural soils, but soil compaction is not a productivity issue so long as roads are open and the running surface is bare. Road decommissioning must take compaction into account in restoring productivity, and various "ripping" treatments are routinely applied to decompact road surfaces.

Growth is sometimes enhanced on or below fill portions of roads because of reduced competition and greater soil depth. Improper fill placement and drainage can cause upslope groundwater to rise, and kill some trees (Boelter and Close 1974, Stoeckeler 1965), but this effect is not common.

General effects of roads and road-associated factors on a wide variety of vertebrate taxa are well documented from a broad range of studies conducted in North America, Europe, and other areas (Bennett 1991, Forman and Alexander 1998, Mader 1984, Trombulak and Frissell 2000, Vestjens 1973). Effects of roads on vertebrate populations act along three lines:

1. Direct effects, such as habitat loss and fragmentation.
2. Road use effects, such as traffic causing vertebrate avoidance or roadkill.
3. Additional facilitation effects, such as overhunting or overtrapping, which can increase with road access.

These factors and their effects on vertebrates in relation to roads are summarized from Wisdom and others (2000) as follows:

• Road construction converts large areas of habitat to nonhabitat (Forman 2000, Hann and others 1997, Reed and others 1996); the resulting motorized traffic facilitates the spread of exotic plants and animals, further reducing quality of habitat for native flora and fauna (Bennett 1991, Hann and others 1997). Roads also create habitat edge (Mader 1984, Reed and others 1996); increased edge habitat favors species that use edges, and harms species that avoid edges or experience increased mortality near or along edges (Marcot and others 1994).
• Species dependent on large trees, snags, or logs, particularly cavity-using birds and mammals, are vulnerable to increased harvest of these materials along roads (Hann and others 1997). Motorized access facilitates firewood cutting, as well as commercial harvest, of these materials.
• Several large mammals are vulnerable to poaching, such as caribou, pronghorn antelope, mountain goats, bighorn sheep, wolves, and grizzly bears (Autenrieth 1978, Bruns, 1977, Chadwick 1973, Dood and others 1986, Greer 1985, Gullison and Hardner 1993, Horejsi 1989, Knight and others 1988, Lloyd and Fleck 1977, Luce and Cundy 1994, Mattson 1990, McLellan 1990, McLellan and Shackleton 1988, Mech 1970, Scott and Servheen 1985, Singer 1978, Thiel 1993, Van Ballenberghe and others 1975, Yoakum 1978). Roads facilitate this poaching (Cole and others 1997).
• Gray wolves and grizzly bears experience chronic, negative interactions with humans, and roads are a key facilitator of such interactions (Mace and others 1996, Mattson and others 1992, Thiel 1985). Repeated, negative interactions of these two species with humans increases mortality of both species and often causes high-quality habitats near roads to function as population sinks (Mattson and others 1996, Mattson and others 1996, Mech 1973).
• Carnivorous mammals such as marten (Martes americana), fisher (M. pennanti), lynx (Lynx canadensis), and wolverine (Gulo luscus) are vulnerable to overtrapping (Bailey and others 1986, Banci 1994, Coulter 1966, Fortin and Cantin 1994, Hodgman and others 1994, Hornocker and Hash 1981, Jones 1991, Parker and others 1983, Thompson 1994, Witmer and others 1998), and overtrapping can be facilitated by road access (Bailey and others 1986, Hodgman and others 1994, Terra-Berns and others 1997, Witmer and others 1998). Movement and dispersal of some of these species also is believed to be inhibited by high rates of traffic on highways (Ruediger 1996), but this has not been validated. Carnivorous mammals such as lynx also are vulnerable to increased mortality from highway encounters with motorized vehicles (as summarized by Terra-Berns and others 1997).
• Reptiles seek roads for thermal cooling and heating, and in doing so, these species experience significant, chronic mortality from motorized vehicles (Vestjens 1973). Highways and other roads with moderate to high rates of motorized traffic may function as population sinks for many species of reptiles, resulting in reduced population size and increased isolation of populations (Bennett 1991). Roads also facilitate human access into habitats for collecting and killing reptiles.
• Many species are sensitive to harassment by humans or simply to human presence. Both often are facilitated by road access. Potential reductions in productivity, increases in energy expenditures, or displacements in population distribution or habitat use can occur (Bennett 1991, Mader 1984). Examples of such road-associated effects are human disturbance of leks (sage grouse [Centrocercus urophasianus] and sharp-tailed grouse [Tympanuchus phasianellus]), ferruginous hawk nests [Buteo regalis], and fox kits [Vulpes macrotis]). Another example is elk, which avoid large areas near roads open to traffic (Lyon 1983, Rowland and others 2000). Elk avoidance of roads increases with increasing rate of traffic (Wisdom and others 2000, Johnson and others 2000).
• Bats are vulnerable to disturbance and displacement caused by human activities in caves, mines, and on rock faces (Hill and Smith 1984, Nagorsen and Brigham 1993). Cave or mine exploration and rock climbing are examples of recreation that could reduce population fitness of bats that roost in these sites (Nagorsen and Brigham 1993, Tuttle 1988). Such activities may be facilitated by human developments and road access (Hill and Smith 1984).
• Ground squirrels often are targets of recreational shooting (plinking), which is facilitated by human developments and road access (Ingles 1965). Many species of ground squirrels are local endemics; these small, isolated populations may be especially vulnerable to recreational shooting.  The potential exists for severe reductions or local extirpations of populations.
• Roads often restrict the movements of small mammals (Mader 1984, Merriam and others 1988, Swihart and Slade 1984), and consequently can function as barriers to population dispersal and movement by some species (Oxley and Fenton 1974).
• Many granivorous birds are attracted to grains and seeds along roadsides and as a result have high mortality from collisions with vehicles (Vestjens 1973). Pine siskens (Carduelis pinus) and white-winged crossbills (Loxia leucoptera), for example, are attracted to road salt; mortality from vehicle collisions can result (Ehrlich and others 1988).
• Terrestrial vertebrates in areas near roads accumulate lead and other toxins that originate from motorized vehicles; the largely undocumented effects may be lethal (Bennett 1991).

In summary, no terrestrial vertebrates seem immune to the myriad of road-associated factors that can degrade habitat or increase mortality. These multifaceted effects have strong management implications for landscapes characterized by moderate to high densities of roads. In such landscapes, habitats are likely underused by many species that are negatively affected by road-associated factors. Moderate or high densities of roads sometimes create population sinks in areas that otherwise would be excellent environments.

See also:  Effects of Roads on Roadkill

Road closures can strongly affect Forest Service timber programs. On Federal timberlands, the timber program and an extensive road network evolved simultaneously. Many roads were built by purchasers or with purchaser credits from timber sales. These roads often served a variety of users. By the late 1980s, about 25,000 National Forest timber sales of more than $300 were recorded per year.  They supplied 14 percent of the U.S. timber harvest. This harvest supported some 125,000 direct jobs in many communities, mostly in the western United States. By 1997, the proportion of total U.S. harvest supplied from Federal land had dropped by half because of efforts to protect various habitats for species at risk of extinction.

Road closures on Federal timberlands have stimulated the development of logging systems that reduce the need for roads. In steep terrain, reducing road densities may require longer cable yarding distances, and because yarding distance is a significant cost factor, especially in thinnings (Hochrein and Kellogg 1988, Kellogg and others 1996, Kellogg and others 1996) timber harvesting costs likely will increase. In addition, greater reliance could be placed on helicopter logging, which can increase logging costs by as much as 2.5 times. Another result could be more wood left behind in the forest because logs must be bucked to their optimum length to maximize the payload of the helicopter.

In gentler terrain, a reduction in road densities could lead to increased use of cut-to-length (harvester-forwarder) systems or more reliance on cable yarding. Movement of logs from stump to landing significantly affects the logging costs. Lanford and Stokes (1996) note that at least with similar primary transportation distances in the Southeast, harvester-forwarder systems have comparable costs per unit harvested to traditional ground-based skidder systems, yet with lower environmental effects. If cable yarding replaced some ground-based systems, costs could increase by 1.4 times or more (Kellogg and others 1996).

More difficult to determine are the long-term effects of focusing future management activities in only the roaded sections of National Forests, where one of the primary management tools is stand manipulation through timber-sale contracts. Some management activities, such as prescribed fire, do not depend heavily on roads, but most of the techniques for stand improvement require some type of vehicle access.

Another issue is how changes in one region relate to changes elsewhere in North America.  Reductions in Federal timber harvesting largely in the West are offset by increases in harvesting elsewhere (mostly Canada and on private timberlands in the South). These offsetting changes moderate effects on consumers.  The largest effects are borne by producers (and their employees) in affected regions. 

The low-cost, low-maintenance intermittent-use road pioneered by Coweeta is strongly recommended by state agencies.  The aim is to reduce sediment, the principal nonpoint source of pollution from forestry activities. Studies at Coweeta have concluded that practical water quality protection can be achieved by:

1. Designing roads with nearly vertical cut banks, broadbased dips, and no inside ditches. 
2. Completing construction and revegetation of cut and fill slopes before winter.
3. Installing brush barriers at the toe of fills if the fills are located within 150 feet of a defined stream channel.
4. Fully graveling roadbeds that drain into stream channels (Swift 1988).

The concepts behind these guidelines are discussed in more detail below.

Road Design

An inexpensive design and field layout procedure can produce a serviceable and environmentally acceptable road. The most effective road system results from a transportation plan developed to serve an entire basin rather than the sum of individual road projects constructed to serve short-term needs (Swift 1988).  Stabilization of cut and fill slopes can reduce sedimentation from newly constructed roads (Grace 1998).

Outsloping and Broad-Based Dips

Roadbeds are typically insloped and crowned to force storm water into a ditch on the inside edge of the roadbed. However, inside ditches are major sources of eroded soil. Adding large rocks to the ditch base can reduce erosion. Where possible, however, storm water should be removed from the road at frequent intervals and in small amounts by outsloping using broad-based dips, rather than by consolidation runoff into ditches and culverts. The outsloped road with broadbased dips disperses sediment-laden stormwater onto the forest floor rather than into a stream. The technique is reasonable if the forest floor is protected by a root mat and litter and the soil has a high infiltration rate typical of Appalachian mountain soils. Current management guidelines call for a filter strip of 84 inches on a 60 percent slope with moderately erosive soil.  Contour roads and gentle grades require less maintenance and produce less sediment than climbing roads (Swift 1988).

Road Surfaces

Gravel surfacing is best, but a grassed roadbed is satisfactory where traffic is light and can be controlled to exclude use in wet weather. Gravel surfacing is the largest single cost item for forest roads; consequently, lower standard, intermittent-use roads often receive only thin coatings of gravel, spot treatments, or no gravel at all. Soils with high coarse fragment content, such as occur in places in the central Appalachians can develop a natural gravel surfacing after an initial loss of finer soil particles (Kochenderfer and others 1984). If only a small quantity of gravel is available, it should be applied on climbing grades, on soils with poor trafficability, in dips, and near stream crossings (Swift 1988).

Berms

Small volumes of water are easier to control and have lower sediment transport capacity than large volumes. Where a road is close to a stream, runoff should be kept on the road until it can be released onto a filtering site. Berms, low mounds of soil or gravel built along the edge of a road, keep storm water temporarily on roadways until a suitable infiltration or sediment trap site is reached. Berms also are used over culvert crossings to keep storm runoff from going directly into a channel. Natural berms may develop along roads from improper road grading or entrenchment of the road.  These berms, which trap runoff on the road, may be removed or channels can be cut through them at sites selected to release runoff away from a stream (Swift and Burns 1999).

Sediment Traps

Sediment traps, or settling basins, placed in ditches or at the ends of turnouts will reduce the velocity of stormwater and deposit suspended sediments. The greater the capacity of the trap relative to the volume of stormflow, the finer the sediment particles that will be trapped. Basin size depends on the volume of water to be treated. Ideally, traps should be designed not to overflow during storms, thus keeping all the sediment out of streams. To prepare for large storms, traps should be cleared of sediment when they are half full (Swift and Burns 1999).

Stream Crossings

A study at Coweeta showed that all sediment collected in a weir originated from roads during the first year after road construction, and that most of it came from stream crossings. Road sections near crossings were often constructed with ditches that drained directly into the channels. Additional ditch outlets or relief culverts installed outside the stream-crossing zone intercept ditch runoff and minimize sediment carried to the stream. Roads often have a low point, a grade sag, at the channel crossing that causes storm runoff to flow from both directions into the stream. Raising the road surface over the crossing will direct road runoff away from the stream and reduce sediment loading. Where there is no grade sag, berms can be built to pass the road runoff across the channel area to a suitable disposal site. Effort should be made to protect and vegetate fill slopes and divert storm water on the road away from streams. Filter strips and brush barriers prevent sediment from reaching streams (Swift and Burns 1999).

Ditches

Ditches can also be constructed to minimize the amount of ditch water flowing directly into streams. Long ditch lines should be broken into shorter sections by constructing additional drains that will reduce both the volume of flow and the sediment load. Ditch outlets should be constructed where they will not empty directly into a stream. Instead, they should empty into sediment traps or onto infiltration areas. Relief culverts may be needed to pass water from the upslope side of the road to infiltration areas or sediment traps on the downslope side (Swift and Burns 1999).  The long-term efficacy of using the forest floor as a buffer zone, or filter strip, is unknown.  Future work is needed to determine factors affecting sediment travel distances and minimum filter strips (Grace 1998).

Road Age

Soil exposed by construction should be revegetated quickly. Typically, newly constructed roads lose the most soil, during the short period before grass becomes well established and the roadbed is graveled or compacted (Swift 1988).

Road Maintenance

The broad-based dip design reduces: (1) the maintenance costs during periods of heavy use, (2) the long-term costs for standby maintenance, and (3) the expenses of reopening a closed road. Experience has shown that grassed roadbeds carrying less than 20-30 vehicle trips a month require a very low level of maintenance. Primary needs are annual mowing of the roadbed and periodic trimming of encroaching vegetation. The outlet edges of broadbased dips need to be cleaned of trapped sediment to eliminate mudholes and prevent the bypass of storm water. Small bulldozers or front-end loaders appear to be more suitable than large bulldozers for periodic maintenance of intermittent-use forest roads (Swift 1988).

Roads provide access to and increase the opportunity for applying a variety of chemicals in National Forests. Some applications, such as road surface treatment, are designed to improve the road. Other chemicals are intended for adjacent ecosystems to control pests and fertilize vegetation. In addition, asbestos from brake linings, oil leakage, and accidental spills are associated with roads. Some portion of applied and spilled chemicals eventually reaches streams by drift, runoff, leaching, or adsorption on soil particles. Roads also increase the nutrient delivery to streams by removing vegetation, rerouting water flow paths, and increasing sediment delivery. Finally, roads increase the likelihood of toxic spills associated with accidents along streamside corridors.

Chemicals

Chemicals applied on and adjacent to roads can enter streams by various pathways. The likelihood of water-quality deterioration from ground applications is a function of how much chemical is applied, the proximity of the road to a stream, and the rainfall, snowmelt, and wind events that drive chemical and sediment movement. The risk is a function of the likelihood of water-quality deterioration, the likelihood of exposure of organisms, including people, and the susceptibility of the organisms to the pollutant or pollutants. A large proportion of Forest Service roads are low standard and few if any chemicals are applied, so the risk of chemical contamination for most Forest Service roads is relatively low.

Chemicals are applied directly to roads and adjacent rights-of-way for various purposes, including dust abatement, stabilizing the road surface, deicing, fertilizing to stimulate plant growth on road cuts and fills, and controlling plants on the roadway (Furniss and others 1991, Norris and others 1991, Rhodes and others 1994). Applied chemicals can enter streams directly when they are applied, but little is known about the effects of these chemicals on stream biota (Furniss and others 1991). Norris and others (1991) provide a comprehensive review of the types and amounts of fertilizers, pesticides, and fire retardants applied to forests in the United States, but little information is given to distinguish road-related from other applications. They report: (1) that most herbicides are applied by ground-based equipment, presumably using roads for access; (2) that ground-based applications in or near aquatic zones can result in chemicals entering streams by drift or direct application; and (3) that these problems are more serious when the chemicals are applied from the air. Movement of sediment containing adsorbed chemicals is possible, and the risk increases with increasing persistence (Norris and others 1991). The amount of input by this pathway is thought to be small, however.

Hazardous chemical spills from vehicle accidents can pose a direct, acute threat of contamination to streams. The risk of hazardous chemical spills resulting from vehicle accidents adjacent to waterways is recognized and documented by the National Forest System and by state transportation departments (IDT 1996). Risk-analysis models of accident-related chemical spills are available, but they are designed for paved roads in nonmountainous terrain. Models take into account risk to human health, traffic frequency, vehicle type, and proximity to water. Possible contaminants include any substance being transported, such as fuel, pesticides, chemicals used in mining, fertilizers, and fire retardants.

The degree to which aquatic organisms are affected by applied and routinely spilled chemicals is poorly known or not understood in most places. Better information on effects is needed to make decisions about chemical application, road drainage control, and road location. Also, better models of chemical spill risks on forested roads are needed.

Increased Nutrient Supply

Increased nutrient supply to streams from roads is proportional to the area disturbed and maintained free of vegetation and the amount of sediment delivered. These increases rarely have detrimental effects on stream water quality, but they may modify the composition of aquatic biota (Hawkins and others, in press). Few studies examining watershed responses to logging separate the effect of road building from those of the broader disturbance associated with removing timber. In one such study, Swank (1988) monitored stream chemical composition during the pretreatment, road building, logging, and posttreatment phases in a cable-logged watershed in the southern Appalachian Mountains. No stream chemical response was found to result from the road-building phase of the watershed treatment.

Nutrient movement to streams often increases significantly after timber harvests (Frederiksen and others 1973, Hornbeck and others 1973, Likens and others 1970, Pierce and others 1972, Swank and Waide 1988). The primary intent of these studies was to assess onsite nutrient losses, with changes in water quality a secondary concern. All cited studies report increases in nitrogen cation and phosphorus concentrations in streams after treatment. In general, nutrient loss to streams is roughly proportional to how much vegetation was removed. For example, three studies at Hubbard Brook in New Hampshire compared three treatments: clearcutting with a herbicide treatment to suppress vegetation regrowth (Likens and others 1970), clearcutting without suppressing regrowth (Pierce and others 1972), and strip cutting of one-third of the forest (Hornbeck and others 1973). The three studies found nitrogen concentrations in streams reduced, most by the first treatment, less by the second, and least by the third. These findings suggest that residual or reestablished vegetation immobilizes released nutrients, thus diminishing the disturbance effect. Although roads might not respond in the same way because of drainage rerouting, we expect that nutrient mobility is proportional to the area maintained in a disturbed, unvegetated state.

Problem roads include those built in poor locations and those constructed and maintained using designs not acceptable by todays standards. Often, these older roads were built along river and stream bottoms. Other problem roads include abandoned "orphan" roads and isolated road sections that are inaccessible. These unused roads may have negative impacts on both terrestrial and aquatic ecosystems (Swift and Burns 1999).

Some problem roads have been redesigned to meet established BMPs. However, forest road systems built before BMP implementation may be prohibitively expensive to move to better locations, even though their impacts on ecosystem health are unacceptable. Also, historic-use rights block the legal closure of some roads, many with unclear ownership, even when they are causing unacceptable damage (Swift and Burns 1999).

Options for treating problem roads include relocation, reconstruction, and closure. The road class-- arterial and collector, local, or orphan,--  directly influences options for treating a problem road. Arterial and collector roads are seldom closed; they are usually recontructed due to the high costs for relocation and new construction. Local roads may be permanently or seasonally closed, if such action is legally feasible. Water and sediment control practices used on local or lower-class roads are similar to those used on arterial or collector roads. However, because local roads typically have less traffic, the road surface on a lower-class road may be rougher and grade changes more abrupt. Designing a replacement in a better location is preferable to attempting to repair a poorly designed road on an inappropriate site. Orphan roads should be permanently closed, if legally possible, unless a responsible individual or organization is willing to reconstruct and maintain them (Swift and Burns 1999).

Problem Road Reconstruction

Problem roads may be improved by reconstructing certain portions, relocating sections, and closing abandoned sections. The treatment selected will depend on road type, cost of treatment, and funding. Closing an abandoned road section without restoration to a more natural hydrologic condition may not arrest the adverse impacts. Light reconstruction may have fewer negative impacts if less soil is disturbed (Swift and Burns 1999).

The landowner or manager must consider several factors when selecting treatments for roads that must be kept open. Experience and studies have established that early application of grass and gravel on forest access roads greatly reduces sediment output (Swift 1984). The types of vehicles and the traffic volume using the road will determine the road width and grade needed to reduce erosion and provide safe passage (Swift and Burns 1999).

Problem Road Closure and Restoration

Primary objectives of road closure and restoration are to eliminate surface erosion and create a more natural site hydrology. Although only portions of a road are causing problems, the entire road should be treated before closure while its full length is accessible. Restoration of an orphan road by closure assumes that the landowner can legally and permanently close the road to further use (Swift and Burns 1999).

Since revegetation increases soil infiltration rates, exposed bare soil must be revegetated to protect it from accelerated surface erosion. Occasionally, streamflow is "captured" by a road crossing, converting the roadway into an eroding channel. Restoration includes placing the flow back in its natural channel. When increased infiltration and runoff dispersion reduce storm runoff and lengthen the time rain takes to reach the channel, peak streamflow rates decline and the watersheds hydrologic function is restored (Swift and Burns 1999).

Roads affect geomorphic processes by four primary mechanisms:

• Accelerating erosion from the road surface by both mass and surface erosion processes.
• Directly affecting channel structure and geometry.
• Altering surface flow paths, leading to diversion or extension of channels onto previously unchannelized portions of the landscape.
• Causing interactions among water, sediment, and woody debris at engineered road-stream crossings.

These mechanisms involve different physical processes, have various effects on erosion rates, and are not uniformly distributed either within or among landscapes.

Mass Erosion

On steep forest land prone to landsliding, the greatest effect of roads on erosion rates is from increased rates of mass soil movement after road building. Mass soil movements affected by roads include shallow (three to several feet deep) debris slides, deep-seated (depths of tens of yards) slumps and earth flows, and debris flows (rapid channelized and fluidized movements of water, sediment, and wood). Of these, effects of roads on debris slides and flows have been the most extensively studied. Typically landslides have been inventoried using some combination of sequential aerial photography and ground verification. Accelerated erosion rates from roads because of debris slides range from 30 to 300 times the normal rate. The magnitude of road-related mass erosion differs with climate, geology, road age, construction practices, and storm history. Several studies in the Eastern United States show that landslides are driven more by storm magnitude and geology than by land use. A threshold of 5 inches of rain per day (Eschner and Patric 1982) and meta-sedimentary geology are associated with large debris slides in the Appalachians. Road drainage can cause small slides in road fills, but some major landslides originate on undisturbed forest land (Neary and Swift 1987, Neary and others 1986).

Road-related mass failures result from various causes: improper placement and construction of road fills and stream crossings; inadequate culvert sizes for water, sediment, and wood during floods; poor road siting; modification of surface or subsurface drainage by the road surface; and diversion of water into unstable parts of the landscape (Burroughs and others 1976, Clayton 1983, Furniss and others 1991, Hammond and others 1988, Larsen and Parks 1997, Larsen and Simon 1993). Effects of roads on deep-seated mass movements have been much less extensively studied, but cases of road building apparently accelerating earth-flow movement have been documented. Such movement can be caused by destabilizing the toe area or diverting water onto the earth-flow complex (Hicks 1982). Little is documented about the potential for increased mass failures from roads resulting from decay of buried organic material that has been incorporated into road fills or landings during road building. Anecdotal evidence, however, shows that failures occur after decay of the organic material.

Although mass erosion rates from roads typically are one to several orders of magnitude higher than from other land uses based on unit area, roads usually occupy a relatively small fraction of the landscape. Hence, their effect on erosion may be comparable to other activities, such as logging. Studies by Swanson and others (1981) in the Oregon Coast Range, for example, showed that although the unit-area increase in erosion from roads was 30 times greater than the increase from clearcutting, road-related landslide erosion accounted for just three times as much accelerated slide erosion in the watershed when the areas in roads and clearcuts were taken into account. Road and clearcut erosion were nearly equal in a study in the west side of the Cascade Range in Oregon (Swanson and Dyrness 1975). In the Klamath Mountains of southwest Oregon, erosion rates on roads and landings were 100 times those on undisturbed areas, but erosion on harvested areas was 7 times that of undisturbed areas (Amaranthus and others 1985). A related point is that only a few sites can be responsible for a large percentage of the total erosion. For example, major erosional features occupied only 0.6 percent of the length of roads studied by Rice and Lewis (1986).

Although road location, design, construction, and engineering practices have improved markedly in the past three decades, few studies have systematically and quantitatively evaluated whether these newer practices result in lower mass erosion rates (McCashion and Rice 1983).

Surface Erosion

Erosion from road surfaces, cut banks, and ditches represents a significant and, in some landscapes, the dominant source of road-related sediment input to streams. Increased sediment delivery to streams after road building has been well documented (Kochenderfer and others 1997, Swift 1985, Swift 1988). Rates of sediment delivery from unpaved roads are highest in the first years after building (Megahan and Kidd 1972) and are closely correlated with traffic volume on unpaved roads (Reid and Dunne 1984, Sullivan and Duncan 1981). Surface-erosion problems are worst in highly erodible terrain, particularly landscapes underlain by granite or highly fractured rocks (Megahan 1974, Megahan and Ketcheson 1996). In the eastern United States, poorly designed and managed forest access and county roads are major sources of sediment input to streams (Hansen 1971, Patric 1976, Van Lear and others 1995). Roads were identified as the major source of sediment in the Chattooga River Basin, where 80 percent of the road sources are unpaved, multipurpose roads (forest and county) (Van Lear and others 1995).  Sediment losses were largest during road building and before exposed soils were protected by revegetation, surfacing, or erosion control materials (Swift 1985, Swift 1988, Thompson and others 1996, Vowell 1985). Soil loss from skid roads in West Virginia ranged from 40 tons/acre during logging, to 4 tons/acre the first year after logging, to 0.1 ton/acre 1 year after logging was completed (Hornbeck and Reinhart 1964). Raw ditch-lines and roadbeds are continuing sources of sediment (Miller and others 1985), usually because of lack of maintenance, inadequate maintenance for the amount of road use, excessive ditch-line disturbance, or poorly timed maintenance relative to storm patterns (Swift 1984, Swift 1988).

Extensive research has demonstrated that improved design, building, and maintenance of roads can reduce road-related surface erosion. Key factors are road location, particularly layout relative to stream systems (Swift 1988, USDA Forest Service 1999), road drainage (Haupt 1959), road surfacing (Burroughs and King 1989, Kochenderfer and Helvey 1987, Swift 1984), and cut slope and fill slope treatments (Burroughs and King 1989, Swift 1988). Many studies show that surfacing materials and vegetation measures can reduce the yield of fine sediment from road surfaces (Beschta 1978, Burroughs and others 1984, Kochenderfer and Helvey 1987, Swift 1984).

Interaction of Roads with Stream Channels

Roads interact directly with stream channels in several ways, depending on their orientation to streams (parallel, orthogonal) and their landscape position (valley bottom, midslope, ridge). The consequences of these interactions, particularly during storms, include increased erosion rates and direct and off-site effects on channel morphology and drainage network structure, but these effects are complex and often poorly understood. Encroachment of forest roads along the main stream channel or floodplain may be the most direct effect in many watersheds. Poorly designed channel crossings also may affect the morphology of small tributary streams, as well as limit or eliminate fish passage. Indirect effects of roads on channel morphology include the contributions of sediment and altered streamflow that can alter channel width, depth, local gradient, and habitat features (pools, riffles) for aquatic organisms (Harr and Nichols 1993).

Roads in midslope and ridgetop positions may affect the drainage network by initiating new channels or extending the existing drainage network. By concentrating runoff along an impervious surface, roads may decrease the critical source area for headwater streams (Montgomery 1994). In addition, concentrated road runoff channeled to roadside ditches may extend the channel network by eroding gullies or intermittent channels on hillslopes and by linking road segments to small tributary streams (Weaver and others 1995, Wemple and others 1996). These effects of roads on the channel network have implications for slope stability, sedimentation, and streamflow regimes.

Dust emitted into the atmosphere by vehicles moving on unpaved roads reduces visibility, and suspended airborne particulates can pose health hazards. Roads built into or surfaced with serpentinic rock may contain asbestos-type minerals that could pose a hazard to people exposed to dust from the road surface.  Soils in the Southwest are often very fine textured, and once dust is stirred up by vehicles, it can remain suspended for a long time and be transported long distances by the wind. 

Dust emissions also raise issues of human health. Where National Forests are close to urban areas, dust from National Forest roads can contribute to the burden of airborne particulate matter. Airborne particles with diameters less than 2 microns have been found to contribute to human health problems and increased mortality, especially in young children, old people, and people with lung problems such as asthma and emphysema. Particles of this size and smaller cannot be effectively cleared by human lungs and therefore accumulate. How much dust from forest roads contributes to the fine particulates in urban atmospheres is unknown for most cities because the EPA is just beginning widespread monitoring of fine particulates, and reliable results will take at least 3 years to gather.

During commercial use of unsurfaced roads, watering or the addition of lignin sulfonate or calcium chloride is often required by the Forest Service or other road manager to reduce dust emissions and conserve the fine fraction of the road surface.

The EPA has proposed a regional haze rule calling for more regions to do the kind of analysis done by the Grand Canyon Commission. Such analyses are likely to find emission problems from unpaved roads elsewhere. EPA's recent tightening of the National Ambient Air Quality Standard on the effects of fine particles on human health are likely to require similar analyses of particle emissions, especially as they affect urban air quality. Analyzing the entire transportation system, including National Forest roads, would be a logical approach to finding the most efficient means of controlling this air pollution. Under emission-trading scenarios, treatments like paving or closure to reduce emissions of particles from National Forest roads might qualify for highway funds.

The basic models of dust emission and transport downwind are generally reliable and widely used by the EPA in regulatory decisions. Much of the basic data to make these calculations for National Forest roads have not been collected; thus, most calculations of the emissions are based on very coarse estimates of the conditions that produce dust emissions. Effects of the amount of road maintenance on emissions also are not well understood. The effects of road closures on dust emissions are difficult to predict because they depend on the details of how traffic is rerouted from closed sections and what emissions are created by the rerouted traffic pattern.

The effects of roads on aquatic habitat are believed to be widespread and profound. Mechanisms for these effects include: (1) movement of fine sediment, (2) changes in streamflow, (3) changes in water temperature caused by loss of shade cover or conversion of groundwater to surface water, (4) migration barriers, (5) vectors of disease, (6) introduction of exotic fishes, (7) changes in channel configuration from encroachment, and (8) increased fishing pressure. At the landscape scale, correlative evidence suggests that roads are likely to influence the frequency, timing, and magnitude of disturbance, which are likely to influence the structure of aquatic communities.

Effects of Fine Sediment

Increases in fine sediment in stream gravel has been linked to decreased fry emergence, decreased juvenile densities, loss of winter carrying capacity, and increased predation of fishes. Increased fine sediment can reduce benthic organism populations and algal production. Survival of incubating salmonids from embryos to emergent fry has been negatively related to the proportion of fine sediment in spawning gravels (Chapman 1988, Everest and others 1987, Scrivener and Brownlee 1989, Weaver and Fraley 1993, Young and others 1991). Increased fine sediment in stream gravel can reduce intragravel water exchange, thereby reducing oxygen concentrations, increasing metabolic-waste concentrations, and restricting movements of alevins (Bjornn and Reiser 1991, Coble 1961, Cordone and Kelley 1960). Survival of embryos relates to dissolved oxygen and apparent velocity of intragravel water, and to gravel permeability and gravel size (Chapman 1988, Everest and others 1987). Consequently, juvenile salmonid densities decline as fine sediment concentrations increase in rearing areas (Alexander and Hansen 1986, Bjornn and others 1977, Chapman and McLeod 1987, Everest and others 1987, Shepard and others 1984).

Increases in fine sediment also can reduce winter carrying capacity of streams by destroying concealment cover (Bjornn and others 1977, Chapman and McLeod 1987, Thurow 1997) increasing the likelihood of predation (Chapman and McLeod 1987). Pools function as resting habitats for migrating adults, rearing habitats for juveniles (Bjornn and Reiser 1991), and refuges from natural disturbances (Sedell and others 1990). Pools that lose volume from sediment (Jackson and Beschta 1984, Lisle 1982) support fewer fish (Bjornn and others 1977), and fish that reside in them may suffer higher mortality (Alexander and Hansen 1986). Similarly, populations of tailed frogs can be severely reduced or eliminated by increased sedimentation (Corn and Bury 1989, Welsh 1990), presumably because of their dependence on unembedded interstitial areas in the stream substrate where they hide and overwinter (Brown 1990, Daugherty and Sheldon 1982). Increased sediment reduces populations of benthic organisms by reducing interstitial spaces and flow and by reducing algal production, the primary food source of many invertebrates (Chutter 1969, Hynes 1970).

See also: Geomorphic Effects.

Effects of Barriers to Migration

Improper culvert placement at roadstream crossings can reduce or eliminate fish passage (Belford and Gould 1989), and road crossings are a common migration barrier to fish (Clancy and Reichmuth 1990, Evans and Johnston 1980, Furniss and others 1991). In a large river basin in Washington State, 13 percent of the historical coho habitat was lost as a result of improper culvert barriers (Beechie and others 1994). Roads built adjacent to stream channels pose additional effects.

Effects of Temperature Change

Changes in temperature and light regimes resulting from removing the riparian canopy can have both positive and negative effects on fish populations. Sometimes increased food availability can mitigate negative effects of increased summer water temperatures (Bisson and others 1988). Beschta and others (1987) and Hicks and others (1991) document negative effects, including elevation of stream temperatures beyond the range of preferred rearing, inhibition of upstream migrations, increased disease susceptibility, reduced metabolic efficiency, and shifts in species assemblages.

Effects of Altered Streamflow

The size, timing, duration, and frequency of streamflow changes also strongly influence salmonid  reproductive success and overwintering survival (McFadden 1969). For example, high flows after spawning can wash out eggs or displace fry (Latta 1962, Mortensen 1977, Shetter 1961). The effect of roads on peak flows is relatively modest, and the issues of changing stability and predictability because of roads may be of little importance to aquatic habitat suitability.

See also: Hydrologic Effects

Effects of Stream Crossings

Road-stream crossings have effects on stream invertebrates. Hawkins and others (in press) found that the aquatic invertebrate species assemblages (observed versus expected, based on reference sites) were related to the number of stream crossings above a site. Total taxa richness of aquatic insect larvae including mayflies, (Ephmeroptera), stoneflies, (Plecoptera), and caddisflies (Trichoptera) were negatively related to the number of stream crossings. Another study (Newbold and others 1980) found significant differences between macroinvertebrate assemblages above and below road-stream crossings.

Effects of Road Density

Several studies at broad scales document aquatic habitat or fish density changes associated with road density or indices of road density. Eaglin and Hubert (1993) show a positive correlation with numbers of culverts and stream crossings and amount of fine sediment in stream channels, and a negative correlation with fish density and numbers of culverts in the Medicine Bow National Forest. Macroinvertebrate diversity negatively correlates with an index of road density (McGurk and Fong 1995). Increasing road densities are associated with decreased likelihood of spawning and rearing of nonanadromous salmonids in the upper Columbia River Basin, and populations are negatively correlated with road density (Lee and others 1997).

Roads can have major adverse effects on biodiversity (Forman and Collinge 1996). A recent review by Forman and Hersperger (1996) distinguishes these aspects of the road-biodiversity interaction:

• Road density: As road density increases, thresholds may be passed that cause some species to go locally extinct. The probability of extinction depends, in part, on body size, with larger animals requiring larger residual populations to prevent their extinction.
• Road-effect zone: The effects of roads can extend over some distance from their centers, such that their "effective widths" can be many times their actual widths.

Two critical uncertainties must be resolved to understand how roads affect fragmentation and population viability. First, in the mechanistic analysis of the effects of roads and roadlike entities, such as power lines, on landscape fragmentation and species viability, the question of the "effective width" of roads is open. Kiester and Slatkin (1974) predict that, for species using conspecific cuing for movement strategies and habitat selection (likely most vertebrates), a spatially localized source of mortality in an area of otherwise suitable habitat can act as a sink, drawing individuals in as residents die, and making it likely that the new individuals will die as well. Consider a road traversing the habitat of a territorial or conspecific-cuing species. Individuals whose home range overlaps a road have some probability of being hit each time they cross the road. Eventually they may be killed, and their neighbors, in the process of constantly testing the boundaries of their home ranges, may move into the vacated area next to the road and themselves run the risk of road mortality. The question is: How far from a road does this probability of mortality spread?

Second, at the landscape scale, the relation between patterns of dispersal of individual species and measurements of fragmentation must be clarified. Schumaker (1996) suggests that most of the commonly used measures of fragmentation do not predict habitat connectivity for individual endangered species; rather, a model of fragmentation must be derived from species-specific dispersal characteristics. This kind of analysis has been done for only a few species.

See also: Passive-Use Value

A widely cited generalization about biological invasion is that it is promoted by disturbance. Building roads and subsequently maintaining them (including ditch clearing, road grading, and vegetation clearing) in the interior of a forest are disturbances that create and maintain new edge habitat. These roadside habitats can be invaded by an array of exotic plant species, which may be dispersed by "natural" agents such as wind and water as well as by vehicles and other agents related to human activity. Roads may be the first point of entry for exotic species into a new landscape, and the road can serve as a corridor along which plants move farther into the landscape (Greenberg and others 1997, Lonsdale and Lane 1994). Some exotic plants may then be able to move away from the roadside into adjacent patches of suitable habitat.

Invasion by exotic plants may have significant biological and ecological effects if the species are able to disrupt the structure or function of an ecosystem. Cowbirds (Molothrus ater), for example, can be introduced into forested environments by roads and subsequently affect populations of neotropical migratory birds through nest parasitism. Roads can also act as vectors for the spread of pathogens (see Forest Diseases).

Invasion may be of concern to land managers, if the exotic species disrupt management goals and present costly eradication problems. However, few environmentally benign approaches to exotic plant control or eradication have been tested.

Although few habitats are immune to at least some invasion by exotic plants, predicting which species will become pests usually is difficult. Assessing the scale of a biological invasion problem is complicated by the lag between when an exotic is introduced and when it begins to expand its distribution and population size in a new area. Also, observations in different settings suggest that the exotic species that successfully invade and the scale of invasion problems differ regionally. A less-than-ideal science base exists for identifying which exotic species pose the greatest threat and what preventive or remedial measures are appropriate.