Design of Forest Riparian Buffer Strips (Part 5)

Buffer strip effectiveness is evaluated in five categories: [1] trapping sediment or nutrients, [2] moderating stream temperatures, [3] providing food and cover, [4] providing large organic debris, and [5] moderating cumulative watershed effects. The first two categories each have several subsections, reflecting the greater availability of information in those two categories. The cost effectiveness of buffer strips is a sixth category addressed in this section.

Effectiveness Trapping Sediment or Nutrients

According to Brown (1985), streamside buffer strips are "of little value in handling erosion from side slopes above the buffer in most of the mountainous West." Erosion in western forests, unlike that from agricultural watersheds where sheet erosion is common, is more likely to occur as channelized flow through the buffer strip. This is due to the relatively high degree of slope dissection by ephemeral channels in upland areas adjacent to the riparian zone. These channels frequently continue through the buffer strip to the channel. Where these channels do not exist, however, sheet flows do move overland.

Effectiveness of buffer or filter strips is expressed in several ways. The more common measure of efficiency is the filter strip width required to contain a given percentage of the number of sediment flows. This can also be stated as the probability that flows will reach a given distance or exceed a given buffer strip width. An alternative expression of efficiency is the percent of sediment actually trapped. This is used when a barrier, such as a hay bale or brush, is placed in the path of the flow. Efficiency is calculated by comparing the quantity of sediment trapped behind the barrier with the quantity of sediment trapped plus that moving through the barrier.

The following sub-sections summarize research on the effectiveness of filter strips below roads, and the effectiveness of riparian vegetation in controlling nutrient and sediment losses from forest harvest sites and agricultural fields.

Trapping or filtering sediment from logging roads. Logging, grazing, fire, road construction, and mass wasting or landslides are common sources of sediment in forested watersheds. Road construction is generally recognized as the largest single source of sediment because removal of vegetation and construction of cut and fill slopes initially exposes large areas of erodible surface. (The exception would be in drainages where mass wasting was extensive.) Packer (1967) studied logging roads in the northern Rocky Mountain region and reported that "most sediment from forest lands that reaches stream channels originates on logging roads." Even after erosion control measures have been implemented, roads continue as sources of sediment for extended periods after logging is completed. Road construction as a sediment source has been well described (Burns 1972, Haupt and Kidd 1965, Megahan et al. 1986) and modeled (Leaf 1974, Burroughs and King 1989, Packer 1967). Mass wasting triggered by road construction is a significant problem. On granitic soils in the Idaho batholith, Megahan et al. (1978) found that almost 66% of landslides occurred on road cuts. In the Oregon coastal range, Beschta (1978) reported severe problems from mass failures caused by roads.

Because of the key role roads play in producing sediment, much attention has been focused on limiting sediment delivery from them. Assuming surface flow, factors controlling the movement of sediment from roads fall generally into two categories: [1] those controlling movement of sediment below the road and within the filter strip, and [2] those influencing sediment production and movement from the road surface. Studies pertaining to each category are described in the following paragraphs.

[1] The key factors controlling sediment movement within the filter strip are slope and the density of obstructions, or surrogate variables for these factors. Trimble and Sartz (1957) identified the average slope of the land below the road as the controlling factor in movement through the filter strip and recommended filter strip widths be increased as the average slope between the road and the stream increased. Swift (1986) compared down-slope sediment movement from roads for various roadway and slope conditions by using as control variables the percent slope and type of surface obstruction, e.g., grass, litter, brush, and downed trees. In Idaho, Haupt (1959a, 1959b) related sediment movement in the filter strip to site conditions and road drainage factors, e.g., aspect, cross-ditch interval, road gradient, fill slope length, and the number and types of flow obstructions along the slope. A similar regional study in the northern Rocky Mountains by Packer (1967) found that travel distances from cross-drain outlets were determined by soil type, age of road, cross-drain spacing, initial distance to slope obstruction, and fill slope cover density.

[2] Several studies focused on mitigation measures to control sediment leaving the road surface and fill slope. Swift (1986) found that brush barriers and hay bales used in windrows are effective sediment traps when placed at the base of the fill slope. On 47% slopes without barriers, the maximum sediment travel distance was 314 feet and the average travel distance 81 feet. When brush barriers were used, these distances were halved. Cook and King (1983) examined the effectiveness of filter windrows on road fill slopes adjacent to streams. Windrows constructed from slash and cull logs obtained from the road right-of-way were 75-85% efficient in trapping sediment before it moved into the filter strip. Similarly, Burroughs and King (1985) compared sediment yields from treated fill slopes to the yields from fill slopes with a loose soil surface. Dense grass planted on a section of fill slope at a 67% slope reduced sediment yield by 97%, a wood fiber mulch reduced sediment yield by 91%, and a slash windrow reduced sediment yield by 87%. The effectiveness of road surface treatments in reducing sediment yields in comparison to unsurfaced roads was also examined. Gravel, dust, oil and bituminous surface treatments reduced yields by a factor of 4.3, 7.7, and 91 respectively.

Reported sediment travel distances and filter strip efficiencies showed considerable variation from study to study. The following studies highlight the difference in travel distance between sediment moving off road fill onto a vegetated filter strip and sediment moving from the road and into a channel formed below a drain. Filter strips on the order of 200-300 feet are generally effective in controlling sediment that is not channelized. Assuming an adequate water flow, sediment from drains can move several thousand feet or more. In New Hampshire, to trap 90% of the number of flows, Trimble and Sartz (1957) recommended filter strips ranging from 25 feet at zero percent slope to 165 feet at 70% slope. For areas where the "highest possible water quality standard" was to be maintained, presumably near 100% efficiency, they recommended doubling the distance. Swift (1986) measured travel distances through forest litter on 47% slopes. The maximum travel distance was 314 feet and the average distance was 65 feet. On burned forest floor at a 42% slope, the maximum travel distance was 198 feet and the average was 96 feet. Working in granitic soils in Idaho, Haupt (1959a, 1959b) reported minimum protective strip widths for a range of road and site conditions. For a road with a 10% gradient on a south slope where the side-slope gradient is greater than 56%, the required filter strip width would be 185 feet to dissipate 83.5% of the number of flows. An additional 45 feet would be needed to contain 97.5% of the flows. The maximum protective strip width recommended in this study was 200 feet for cross-ditch intervals of 130 feet. Packer (1967) reported protective strip widths needed to contain 83.5% of the number of flows on comparatively stable basalt soils ranged from 35 to 127 feet depending on the type of obstruction--e.g., slash or herbaceous vegetation--and spacing between obstructions. Efficiency of the protective strip could be increased to 97.5% by adding an additional 60 feet to the strip widths. In the Idaho batholith, Ketcheson and Megahan (1990) observed sediment deposition on slopes below roads and concluded that sediment originating from cross-drains where sediment can accumulate and water supply was relatively large could reach streams up to 4,500 feet down-slope. However, the probability of sediment from cross-drains traveling in excess of 300 feet is only 15%. Sediment discharged from other road sources--e.g., fill slopes, berm drains and rock drains--traveled no more than 200 feet, with a near-zero probability of exceeding 200 feet. In another Idaho study on steep slopes with soils derived from gneiss and schist parent materials, Burroughs and King (1989) examined sediment travel distances below road fill slopes. They found that 90% of the sediment flows below fill slopes traveled less than 88 feet. Where fill slope flows were influenced by flows from drains, 90% of the flows traveled 200 feet or less. In southwestern Washington, Bilby et al. (1989) documented the export of sediment from road surfaces and found that about 34% of the road drainage points studied entered first and second order streams via small channels. They observed that retention of sediment in these channels increased with particle size, and that the small channels became temporary storage repositories for sediment.

Results from the road filter strip studies summarized above have important implications for designing SPZs in Idaho. First, the Idaho study by Haupt (1959a, 1959b) and the regional study by Packer (1967) provided reasonable estimates of needed filter strip widths where the sediment source is a logging road and that road is located near a stream, a common situation in Idaho. Similarly, the erosion control work by Cook and King (1983) and Burroughs and King (1985) is applicable in the same context. Second, although results from the studies cited in this section are not directly applicable to situations where the sediment source is other than roads, they do provide useful general information about riparian buffer strip effectiveness. These studies specifically indicate that given a sediment source, non-channelized transport distance increases with slope and decreases with the number of obstructions within the filter strip. The studies also suggest that for non-channelized flow, sediment rarely travels more than 300 feet. Channelized flows through filter strips, however, can move thousands of feet and are limited primarily by the amount and frequency of flow. A survey of forest practice compliance by the Idaho Water Quality Bureau (1988) found that "...existing roads near stream channels is [sic] the most important factor currently contributing to water quality degradation."

These findings suggest four things about buffer strip design: [1] riparian buffer strip widths should be greater where slopes within the zone are steep, [2] riparian buffers are not effective in controlling channelized flows originating outside the buffer, [3] sediment flow through a buffer can travel up to 300 feet in a worst-case scenario, and [4] removal of natural obstructions to flow--vegetation, woody debris, rocks, etc.--within the buffer increases the distance sediment can flow.

Filtering nutrients and sediment from forest lands. The impacts of forest practices on nutrient cycling and the loss of nutrients through streamflow have received considerable attention in the literature (see Martin and Harr 1989, Tiedemann et al. 1988, Hornbeck et al. 1986, Clayton and Kennedy 1985, Martin et al. 1984, and Aubertin and Patric 1974 and are summarized well in the textbook by Brooks et al. (1991). However, the influence of riparian filter strips on sediment and nutrient discharge, with the exception of the previously discussed road-side filter strips, has not been examined extensively. In northern Idaho, Snyder et al. (1975) found that following clearcutting and burning of slash, buffer strips reduced the loss of certain nutrients and filterable solids, i.e., organic matter and sediment. Effectiveness of the buffer strips as filters was not determined in that study. In northern Idaho, Skille (1990) monitored the effects of fall slash burning in or near SPZs and noted substantial increases in nitrogen and phosphorus loading of streams following late fall rainfall. Although the effectiveness of buffer strips as filters was not evaluated, Skille noted that early fall burning tended to reduce increases in stream concentrations of N and P because early rainfall was adequate to move the nutrients into the soil, but inadequate to carry them to the stream.

These studies suggest that filter strips reduce the amount of nutrient loading following harvest and slash burning, but they do not provide a basis for determining the size or effectiveness of buffer strips. The studies also suggest that where nutrient loading is a problem, burning slash within the buffer is likely to increase the loading and the problem.

Trapping nutrients and sediment from agricultural lands. The utility of forest riparian zones as buffers for sediment and nutrients from agricultural lands is of interest because forested riparian lands are commonly used for containment of wastes. Statistical models were developed by Omernik et al. (1981) to relate nutrient levels in streamflow to the extent and proximity of forested and agricultural lands to streams. These models were unable to show that the proximity of forest lands impacted stream nutrient levels. Cooper et al. (1987) estimated that more than 50% of the sediment lost from cultivated fields was deposited in the channels within 100 meters of the fields and that only 25% reached a riparian swamp two kilometers distant. In the Southwest, Kuenzler (1988) found that freshwater forested wetlands were effective filters in removing suspended sediment and nutrients. In Georgia, Lowrance et al. (1984) examined nutrient cycling in a forested riparian ecosystem and reported it was potentially an excellent sink to store nutrient and chemical releases from agroecosystems. To maintain the capacity of the riparian ecosystem as a buffer, they suggested that "proper streamside forest management requires both periodic harvest of trees to maintain nutrient uptake and minimum disturbance of soil and drainage conditions." Lee et al. (1989) modeled phosphorus transport through grass buffer strips and found the transport process, via dissolved solids and sediment, to be largely controlled by buffer strip width and length, infiltration rates, grass spacing, the buffer strip slope, and Manning's roughness coefficient (see the Glossary). Transport of dissolved P was also sensitive to the amount of above-ground biomass.

These studies suggest the utility of forest vegetation and wetlands as filters for sediment, nutrients, and other chemicals; but provide no definitive means of estimating the dimensions of the required buffer strip.

Effectiveness Moderating Stream Temperatures

Stream temperature elevation and control following harvesting. Summer stream temperature increases from the removal of riparian vegetation have been well documented. Increases in June to August maximum stream temperatures of 2C to 10C are common in the Pacific Northwest (Beschta et al. 1987). These studies generally support the findings of Brown and Krygier (1970) that for summer periods when streamflow is normally low and air temperatures are high, loss of riparian vegetation results in larger diurnal temperature variations and elevated monthly and annual temperatures. Measurements by Hewlett and Fortson (1983) under winter conditions also indicate that removal of riparian vegetation can reduce temperatures by about 10F. These studies have been summarized by Beschta et al. (1987).

The effectiveness of buffer strips in moderating stream temperature has also been studied. In West Virginia, Aubertin and Patric (1974) reported negligible changes in stream temperature after clearcutting and attributed this to a buffer strip and fast regrowth after harvest. In Pennsylvania, maximum monthly stream temperatures on a clearcut area where a buffer zone was left along a perennial stream showed only a slight change of less than 1C in comparison to control watershed measurements (Rishel et al. 1982). Similarly, in North Carolina, a narrow buffer strip left in clearcut areas would moderate stream temperatures caused by harvesting (Swift and Baker 1973). Although these studies demonstrated the utility of riparian buffer strips, they provided limited information as to the dimensions or vegetative characteristics of the buffer strips that are required to make them effective.

Shade from riparian vegetation and stream temperature. Several studies of the heat energy exchange between a partially shaded stream and its environment have shown that solar radiation is the dominant source of energy, whereas evaporation and conduction to the channel bottom are the principal energy sinks (Brown 1969, Sullivan et al. 1990). Little can be done about the sinks in practice, so the major opportunity to control stream temperature is to moderate the input source--solar radiation from the sun. Buffer strips provide the opportunity.

The presence of shade-producing vegetation in buffer strips is a key factor determining the amount of radiant energy that reaches a stream. Other important determinants are local topography, stream reach orientation to the sun, and stream width and depth (Brown 1985). The volume of timber in a buffer strip is not well correlated with shade (Brazier and Brown 1973); however, statistically significant relationships have been found between buffer strip width and shade expressed as angular canopy density, or ACD (Steinblums et al. 1984, Brazier and Brown 1973). ACD effectively integrates spatial factors--e.g., stream width, tree height, and canopy density--for a given site. A series of ACD readings at intervals along a stream reach provides an average value of ACD for the stream reach.

Buffer strip width and stream temperature. Two studies in Oregon have demonstrated that buffer strip width is not a good measure of buffer strip effectiveness in protecting stream temperature. Steinblums et al. (1984) measured 40 riparian buffer strips. For 28 buffer strips with widths from 25 to 145 feet, ACD measurements ranged from 15% to 87%, with an average of 51%. On 12 other buffer strips of essentially infinite width, ACD measurements ranged from 26% to 83%, with an average of 62%. The relatively small and unreported statistical differences in these data and in their means illustrate the importance of factors other than buffer strip width in determining ACD, including species, tree height, stream width, and stream orientation. Also in Oregon, Brazier and Brown (1973) defined buffer strip effectiveness as net radiation or heat blocked by the canopy, and developed two statistical relationships: [1] between heat and ACD, and [2] between buffer strip width and ACD. They concluded that ACD is the most appropriate single measure of canopy effectiveness, and that buffer strip width alone is not a significant variable for predicting stream temperature. For the streams included in that study, maximum ACD occurred with an 80-foot buffer strip, and 90% of the maximum ACD could be obtained with a 55-foot strip.

In an Ontario trout stream study, Barton et al. (1985) demonstrated the relative insensitivity of stream temperature to buffer strip width. They found stream temperature declined an average of .015C per meter of buffer strip width, or about .5C for a 100- foot buffer strip.

Effectiveness Providing Large Organic Debris (LOD)

Large organic debris (LOD) enters streams on an irregular basis due to natural mortality, severe storms, and fire. Harvesting or other management practices that influence stand characteristics, such as tree species and stocking levels, also influence the timing and quantity of LOD contribution. Site characteristics, such as depth of water table and orientation to dominant winds, affect

windthrow and hence LOD contribution (Steinblums et al. 1984). A Canadian study

reported by Toews and Moore (1982) compared LOD recruitment from three clearcut areas. One was logged intensively, a second carefully, and a third was left with buffer strips 5 meters (16 feet) or less in width. The intensively and carefully logged areas--those without buffer strips--provided large amounts of LOD that resulted in reduced stability of LOD already in the channel as well as bank instability. LOD contributions from the clearcut area with a buffer strip were similar to natural LOD recruitment levels.

Source distances for coarse woody debris--the distance from rooting site to stream bank--were studied at 39 sites in western Oregon and Washington by McDade et al. (1990). Their analysis for old-growth conifer forests suggested that a 30-meter (98-foot) wide buffer strip would provide 85% and a 10-meter (33- foot) strip would supply less than half the amount of naturally occurring debris. Source distance and debris size was less in old-growth stands than in mature stands with shorter trees, indicating tree height was a factor in LOD recruitment. They also found that the number of debris pieces and the source distance increased with bank slope. Using effective tree height as a measure, Robison and Beschta (1990) determined the conditional probability that a tree would provide LOD to a stream. Effective tree height was defined as the height at which a minimum acceptable diameter size for LOD occurred. When the distance from a tree to stream was more than one effective tree height, the probability of the tree contributing LOD approached zero. This suggests that buffer strips with widths at least equal to the effective tree height would provide maximum amounts of LOD. Unfortunately, research data currently are inadequate to provide general guidelines as to how much LOD should be available or how much is required at a given stream reach.

Effectiveness Controlling Cumulative Effects

Cumulative effects are impacts on water quality or beneficial uses which result from the incremental impact of two or more forest practices (Idaho Legislature 1991). Cumulative effects can result from individually minor, but collectively significant, actions taking place over time or space. Although numerous studies of cumulative effects--e.g., water yield increase (Belt 1980, King 1989) and temperature increases (Beschta and Taylor 1988)--appear in the literature, research relating buffer strips to control of cumulative effects is limited. However, a few interesting examples were discovered. A Canadian study of the suitability of streams for trout (Barton et al. 1985) showed that the maximum 3-week average stream temperature was determined by the upstream length of forested buffer strip. In a study of 11 sites, the cumulative effect of the removal of upstream riparian vegetation was to increase the maximum on-site temperature. A simple mixing ratio equation (US EPA 1980) allows estimation of the cumulative effects of upstream temperature increases at downstream locations. Lowrance et al. (1984) noted that buffer strips are simultaneously sinks that retain and sources that release the cumulative effects of agricultural and forestry activities in the form of sediment, nutrients, and chemicals. This suggests that buffer strips and wetlands should be managed to enhance their storage capability. Buffer strips and adjacent wetlands can moderate flooding caused by the cumulative effects of timber harvest by adding the hydraulic resistance from riparian vegetation and additional storage capacity at flood stage.

Effectiveness Providing Food and Cover

Buffer strip impacts on the aquatic food chain are documented reasonably well by studies contrasting the effects of timber harvests with and without buffer strips. However, there are only a few studies where the characteristics of the buffer strip were related to food production so that buffer strip effectiveness could be evaluated.

A study by Erman and Mahoney (1983) in California measured buffer strip effectiveness relative to food production in terms of the rate of recovery of post-harvest macroinvertebrate diversity to preharvest levels. Diversity in streams at logged sites without buffer strips and with 30-meter (98-foot) buffer strips were compared to diversity in streams where no logging had taken place. The streams without buffer strips showed an increase in diversity but incomplete recovery after a 6-year period, while the streams with buffer strips maintained diversity at a constant level. In a similar New Zealand study, Graynoth (1979) compared impacts on clearcut watersheds, with and without buffer strips, with those on a third uncut catchment used as a control. After harvest, in the stream without a buffer strip water temperature and sediment increased while benthic invertebrate fauna and the number of fish declined. The stream with the buffer showed little or no impact, except for increased sediment. Culp (1987) reported that 10-meter (33-foot) buffer strips reduced fine sediment from bank erosion but did not prevent decreases in macroinvertebrate density. In southwest Alaska, Duncan and Brusven (1985, 1986) developed a series of energy-flow models that related the percent coverage of riparian canopy to three biological production variables--invertebrate, potential salmonid, and usable allochthonous production. These models also included the percentage of deciduous tree species and relative stream nutrient levels as variables. The model for invertebrate production showed that reducing the canopy from 100% to 50% caused a 28% decrease in invertebrate production. Conceptually these models could also be used to estimate the effectiveness of buffer strips in providing food and cover based on changes in riparian canopy density. The authors, however, cautioned that the models should be used to evaluate trends in production rather than absolute values.

Cost Effectiveness of Buffer Strips

Consistent with most analyses of the costs and benefits of natural resources management alternatives, the costs of buffer strips are relatively easy to quantify, but the benefits are not. Establishment of buffer strips normally results in additional costs to the landowner, public or private. Costs incurred include the loss of stumpage, higher costs of logging and road construction, and additional administrative costs (Streeby 1970). Benefits from buffer strips accrue largely to the public and include improved bank stability and water quality, enhanced fish and wildlife habitat, and greater aesthetic value.

Bollman (1984) noted that the costs of specific buffer strip prescriptions vary with market conditions, the type of stand, and other variables, but were relatively easy to evaluate. Conversely, benefits from the prescriptions were frequently non-market values--e.g., fish habitat, species diversity, and water quality--that were much more difficult to evaluate. The question of equity arises when private land owners or logging firms must bear the costs of operating in or around buffer strips that benefit sport fishermen, other industries such as commercial fishing, or the general public (Gillick and Scott 1975). In a detailed benefit-cost analysis of buffer strips in Washington, Gillick and Scott (1975) addressed these problems, and suggested the use of a "financially optimal buffer" with an optimal width where the harvest costs would be offset by the environmental gains. The optimal width is by their definition the most cost-effective width. Considering only the values of fish and logs, they found the "zero foot" buffer strip--i.e., no buffer strip at all--to provide the greatest net economic value.

In Puerto Rico, Scatena (1990) identified an "economically optimal buffer width...[where] ...marginal gain in buffer area equals the marginal increase in commercial basal area included in the buffer." In this case, buffer strip area was used as a surrogate for benefit, and basal area included in the buffer as a surrogate for cost. Based on studies in several tropical catchments, this optimal buffer width was 22 meters (73 feet) for perennial streams and less than 10 meters (33 feet) for intermittent streams. This study, although not directly applicable to Idaho, illustrates an alternative approach to evaluation of buffers based on financial criteria.

These studies suggest potential difficulties in establishing buffer strip areas or widths based on economic criteria such as a benefit-cost ratio. First, although costs are relatively easy to determine, important non-market benefits are difficult to evaluate. Second, the value society places on non-market riparian benefits such as biological diversity is subject to not only measurement difficulties but also considerable changes in public perception and relative scarcity, all of which are likely to be substantially greater in the future. Consequently, establishing buffer strip areas or widths based solely on the economically optimal or most cost-effective methods illustrated by the two studies cited in this section could be short-sighted, as the public perception of benefits might be expected to increase over time faster than the costs.

Previous Part (4)

Next Part (6)

Moscow, ID 83844

Send suggestions to webmaster@uidaho.edu