Due to the expense and time required for paired watershed studies and in-stream evaluations of sediment following BMP applications, few studies have measured the actual efficiency of forest BMPs. Edwards and Williard [91
] evaluated the literature and found only three studies that provided BMP efficiencies with regard to sediment loading reductions and reported BMP efficiencies ranging from 53%–94%. Rather than using in-stream water samples, numerous evaluations of BMPs have examined erosion, either through direct measurements or models. Five recent studies of bladed skid trails [75
], overland skid trails [58
], skidder stream crossing approaches [56
], and truck road stream crossing approaches [16
] in the Piedmont of Virginia have compared direct measures of sediment trap erosion with a variety of models. Overall, they found that the USLE-Forest provided rankings similar to those of direct measures of soil erosion, thus indicating that it is an appropriate tool for use by managers.
Another approach to evaluating the adequacy of BMPs is through examination of BMP compliance levels. Ice et al.
] reported that western states implementing BMPs under the auspices of state forest practices acts were evaluating the effectiveness of the programs by monitoring compliance with the statutes. Sugden et al.
] reported on the BMP progress made in Montana by reporting BMP compliance percentages. Schuler and Briggs [80
] evaluated the effectiveness of New York’s BMP program by reporting BMP compliance rates, and Briggs et al.
] used a similar technique to evaluate the effectiveness of Maine’s BMP program. Ice et al.
] reported on BMP implementation rates across the U.S. and reported that the rates were widely used as surrogates for effectiveness.
Rather than use estimates of erosion or BMP compliance levels individually, we elected to use multiple evaluations regarding the quality of the BMP installations, including two variations of the VDOF BMPs (BMP audit scores and rankings based on BMP guidelines) and the USLE-Forest soil erosion model. Arguments can be made regarding the advantages of any one approach, but we believe the aggregation of three evaluation approaches provides strong evidence of BMP efficiencies.
The prevalent usage of culverts for truck roads and portable bridges for skid trails probably reflect the advantages of culverts for truck stream crossings and portable bridges for skid trail crossings from the loggers’ perspective (Table 2
). For log trucks, culverts are advantageous because they are relatively simple to install, can support heavier loads than many portable bridges, and require less knowledge of engineering compared to permanent bridges. In a survey of Virginia loggers, McKee et al.
] found that portable steel or wooden bridges were commonly used by loggers for temporary skidder crossings (hence the alternative name of skidder bridge) because they can be used multiple times, thus having a reduced cost per crossing compared to culverts. However, culverts also have potential disadvantages. Improperly installed culverts can alter in-stream flow characteristics and potentially hinder migration of stream organisms [3
]. Also, the installation of culverts requires that fill material be placed around the pipe in the streambed, which is why Reeves et al.
] and Aust et al.
] concluded that culverts may be associated with increased sediment levels compared to portable bridges.
Morris et al.
] conducted a rainfall simulation experiment to determine the sediment contribution from a truck road culvert in the Virginia Piedmont that had substandard BMPs (BMP−), standard BMPs (BMP-standard) and enhanced BMPs (BMP+). BMP− produced 3.5 times more sediment than BMP-standard and 4.4 times more sediment than BMP+. Aust et al.
] evaluated sediment above and below four types of operational stream crossings (portable bridges, culverts, fords, and pole crossings) and potential erosion rates for the associated stream crossing approaches in the Virginia Piedmont. Overall, they concluded that BMPs were an important factor for reducing sediment from all crossing types and that poorly designed and maintained approaches with insufficient BMPs could potentially produce up to 95 Mg·ha−1
of erosion. They also found that temporary bridge crossings tended to produce the least sediment, while culverts were associated with higher sediment levels. Permanent crossings tended to produce more sediment than temporary crossings, because temporary crossings were effectively closed with BMPs and temporary roads disturbed smaller areas within SMZs.
Continued reliance on culverts for truck road stream crossings would benefit from enhanced BMPs, and our data emphasize that culverts are still being used and that BMP implementations can be improved for these situations. Our predicted enhancement of BMP− truck road crossings to BMP-standard level would potentially reduce erosion rates by 5.1 times, and we estimated the average cost of these improvements to be approximately $450 (Table 6
). Improving a BMP-standard crossing to BMP+ levels would reduce erosion by 11 times, and predicted erosion rates would drop to less than 1 Mg·ha−1
, which is similar to erosion rates in undisturbed forests [7
]. Costs for these improvements would be between $450 and $500 per crossing. These BMP enhancements provide dramatic evidence of the potential positive effects of BMPs for a relatively small cost per crossing, compared to a shift to permanent bridges.
Truck crossings had higher mean and median BMP audit scores (85.4% and 91.4%, respectively) compared to the mean and median BMP audit scores for the skidder stream crossings (70.8% and 77.8%, respectively) (Table 3
). For the same general period, the VDOF 2014 BMP audit report provided a somewhat similar mean BMP audit score (89.2% overall) for both truck and skid trail stream crossings in the central region. Our BMP audit scores could have been lower than the VDOF audit for several reasons. The VDOF evaluations cover the entire harvest site, whereas we were applying the questions to one particular crossing. Also, we conducted our audit on sites that were harvested from January through April, which was a very wet season and could have affected BMP performance.
In this particular study, we also suggest that our estimates of soil erosion are nearly equivalent to the sediment delivery to the stream. Lang et al.
] examined SMZ failures in the Piedmont and found that stream crossings are their primary cause. In a separate study, Lang et al.
] compared USLE erosion estimates from forest road stream crossing approaches in the Piedmont of Virginia and found that sediment delivery rates from forest approaches to streams were nearly 100%. Rivenbark and Jackson [44
] examined causes of SMZ failures in the Piedmont and concluded that roads and stream crossings were directly associated with 25% of all SMZ failures. Thus, our USLE-Forest estimates potentially provide the approximate quantity of sediment delivery that could be expected at these crossings. This higher sediment delivery at stream crossings is markedly different than most forest operations, where sediment delivery would be far less than potential erosion rates. Ward and Jackson [97
] compared soil erosion below timber-harvested and site-prepared sites in the Piedmont with the USLE and sediment traps at the SMZs, and estimated a sediment delivery ratio of 25% to the edge of the SMZ. They found that the SMZ trapped 71%–99% of the erosion before delivery to the stream. Lakel et al.
] did a comparable study below harvested sites without site preparation and found a sediment delivery ratio between 3% and 14%, but Lakel also emphasized that BMP failures and resultant sediment problems were potentially much higher at stream crossings. It is important to understand that roads may provide a more direct conduit to the stream.
In our study, the differences in erosion between the truck roads and skid trail stream crossings were not unanticipated, as skid trails have lower BMP standards than truck roads. For example, truck road stream approaches have a recommended slope of ≤10%, whereas bladed and dispersed overland skid trails may have slopes of ≤25% and ≤35%, respectively, and still comply with the VDOF BMP guidelines [71
]. The steeper approaches allowed on the skid trail stream crossings are a contributing factor to the twofold potential erosion rate of the skid vs.
truck crossing (Table 3
), and another future BMP enhancement for skid trail stream crossings could be to generally keep skid trail approaches ≤15%, although this may not always be feasible.
The third surrogate for road crossings and quality of BMPs involved evaluation of the road template, road drainage structure, road cover BMPs, and stream crossing structure in order to develop a general ranking of the crossing and approach (Table 4
). Subsequently, we recommended BMP upgrades that would allow the BMP− and BMP-standard categories to move to the next higher BMP category and estimated the potential erosion for each crossing using the USLE-Forest (Table 6
). Such hypothesized improvements and costs have been used regarding forestry BMPs by several researchers [89
] in order to estimate costs.
For BMP− situations, inadequate water control features were the most common BMP implementation problem for both skid trails and truck roads. The most prevalent water control on BMP− skid trails related to water bars with inadequate spacing based on the slope of the skid trail or ineffective construction. Reasons for water control inadequacies are uncertain, but equipment operators who visually estimate requirements with little measurement often are responsible for placement of water control structures. Poorly designed turnouts that were too small or led the water directly towards the stream were common reasons for turnout failures on BMP− truck roads. Coverage problems for BMP− roads typically involved either inadequate or no gravel, and poor establishment of seeded grass. BMP− skid trails that had problems with slash coverage were often related to the use of slash that was too large to provide good soil contact or was too sparse. For BMP− skid trails using seed, the problems for cover typically related to lack of grass establishment.
The two most common recommendations were to add additional water control structures or to increase surface cover. Such simple BMP improvements were estimated to reduce potential erosion from BMP− truck and skid trail crossings by five times (Table 6
). The improvements for the BMP-standard crossinigs suggested upgrading these to the BMP+ category with similar positive effects. The reduction for the truck roads would potentially reduce erosion by 11-fold, and the skid trail enhanced BMPs would potentially result in a 6-fold decrease in erosion. These reduction levels compare favorably to the reduction levels measured by Wade et al.
], who found that enhanced ground cover on bladed skid trails could reduce erosion by 30 times, and the road crossing studies by Brown et al.
], who reported that increased levels of BMPs could reduce road erosion rates by 7.5 to10 times. As expected, the average BMP audit scores improved for both truck and skid trail crossings as they progressed from BMP− to BMP-standard to BMP+. Morris et al.
] used a similar methodology for evaluating the sediment contributions of different stream crossings in the Piedmont (ford, culvert, bridge) having three levels of BMP implementation. They evaluated total sediment loading for one year, and overall contributions were 98.5 Mg for BMP−, 28.5 Mg for BMP, and 22.5 Mg for BMP+. Our findings also emphasize the importance BMPs at stream crossings and the potential benefits of using more than the standard level of BMP for such critical areas. The Virginia Department of Forestry has estimated that approximately 1000 new stream crossings are constructed annually, and improved BMP compliance at these crossings provides obvious opportunities for sediment reduction.
After completing the Virginia BMP guideline rankings, we evaluated the specific BMPs needed for enhancement of existing BMPs and estimated the cost of the recommended BMP improvements using the Virginia Tech Road and Skid Trail Cost Method [85
]. Costs of applying hypothesized treatments to upgrade BMP levels were substantially more for truck roads than for skid trails, and the costs were greater to improve the BMP− roads and skid trails (Table 4
). Average road improvement costs were between $450 and $580 per crossing, and skid trail improvements ranged between approximately $50 and $150 per crossing. McKee et al.
] conducted interviews with 70 Virginia loggers and found that the estimated costs for stream crossing closure BMPs in the Piedmont was $445 in 2009. Our costs for improvement are greater than the estimated costs in 2009, but our evaluations entailed a careful assessment of each site rather than relying on loggers’ recollections. Our estimates were based on remedial BMPs, which would have been less cost-efficient than if completed during the original BMP installations. Furthermore, our hypothesized recommendations would not include any substandard BMPs that might have been included in previous estimates.