The Arctic Coastal Plain of northern Alaska is a vast area of treeless tundra, with low relief and numerous lakes and rivers flowing into the Beaufort Sea. This area, commonly referred to as the North Slope, has supplied a substantial portion of the oil used in the United States since production began in 1977. The Alpine oilfield (70°20’N, 150°56’W) began producing oil in 2000 and was the first field to be developed within the delta of the Colville River, one of the largest rivers on the North Slope (Figure 1). Flooding and erosion are primary environmental concerns that prompted the development of innovative engineering techniques to stabilize the Alpine oilfield infrastructure. As of 2004, infield facilities included approximately 8 miles of gravel sideslopes along roads, pads, and an airstrip. In this article, we describe the combination of physical armoring and biotechnical stabilization methods used to stabilize these slopes. To the best of our knowledge, these biotechnical stabilization techniques have not previously been used on a large scale in the Arctic.
Background
Perennially frozen ground, or permafrost, is a primary factor affecting the design of oilfields in the Arctic. Only the top 1 to 3 feet of soil thaws during the short, cool arctic summer, while the deeper soil remains frozen and may contain a substantial amount of ice. To conduct oilfield activities, it is necessary to insulate the permafrost, to minimize thawing and subsidence (thermokarst) resulting from melting ice and subsequent loss of soil volume. Thus, facilities and vehicle traffic typically operate on roads, pads, and airstrips that are constructed of gravel fill about 6 feet thick. The thick fill minimizes heat transfer to the underlying soil, thereby maintaining a stable working surface. In the Alpine oilfield, however, it was necessary to use gravel fill as thick as 10 feet, to elevate the facilities sufficiently above spring floodwaters. To minimize the footprint of the facilities and reduce the quantity of gravel required, relatively steep slope angles (approximately 2:1) were incorporated into the design (Figure 2).
Placing thick gravel fill structures in the delta was expected to affect water flows under certain conditions. Hydrological modeling indicated that floodwater velocities sufficient to cause some erosion of the steep gravel slopes would likely occur in some years. The largest flood event each year typically occurs during spring breakup in late May and early June (Figure 3). During the spring flood in 1993, which was considered a moderate flood event with a return period of five years, the peak discharge was 379,000 cubic feet per second (cfs), and water covered 69% of the area where the Alpine facilities were to be developed. Simulation modeling indicated that during a 200-year flood event, the peak discharge could be as high as 1,000,000 cfs, resulting in flooding of nearly the entire area.
Concerns about flood erosion were addressed partly by siting most of the facilities on higher portions of the floodplain, where flooding is less frequent and flow velocities are less extreme. For example, the elevation at the toe of most of the gravel sideslopes is at the maximum elevation that floodwaters are expected to reach during a 50-year flood. To protect against the erosion of the sideslopes during flooding, physical armoring and biotechnical stabilization techniques were applied. These techniques were approved in the permit issued by the US Army Corps of Engineers for the development of the Alpine oilfield.
Slope Stabilization Treatments
Based on hydrological modeling results, the 8 miles of sideslopes were classified into areas with low or medium (6 miles) and high (2 miles) flooding risk and treated with corresponding erosion control techniques (Figure 4). In the low-risk areas, we applied a seed mix of native-grass cultivars at 40 pounds per acre, using handheld spreaders. These species are native to northern Alaska and were developed by the Plant Materials Center (PMC, Division of Agriculture, Alaska Department of Natural Resources, Palmer, AK) to provide a commercially available source of grasses for revegetation projects. Once the grasses matured, the rooting system was expected to be sufficient to keep gravel in place during seasonal flooding. The mix used along the work pads and road contained 34% arctared fescue (Festuca rubra), 34% alpine bluegrass (Poa alpina), and 32% tufted hairgrass (Deschampsia caespitosa).
For safety reasons, it was necessary to avoid attracting wildlife such as caribou (Rangifer tarandus), muskox (Ovibos moschatus), and greater white-fronted geese (Anser albifrons) to the vicinity of the airstrip. Therefore the sideslopes of the airstrip were seeded with 72% tufted hairgrass and 28% alkali grass (Puccinellia borealis), which are considered less palatable than the other grasses. The alkali grass seed was generously supplied by the PMC, which is currently assessing its palatability to migrating birds at the international airport in Anchorage, AK. On all the sideslopes where grasses were sown, we also applied a granular fertilizer comprising nitrogen, phosphorus, and potassium (20-20-10 NPK) at the rate of 400 pounds per acre using a truck-mounted spreader, to ensure adequate nutrients for seedling establishment and initial growth.
 |
| Figure 5. Aerial view of the swale area and the CD-2 access road showing erosion control treatments (left); Figure 6. Installation of mesh blankets in the trench and articulated concrete revetment covering the sideslope adjacent to a bridge (right) |
The moderate risk area included a 2,000-foot-long section of road that crosses a low-lying swale, where flooding occurs routinely (Figure 5). Due to higher water velocity and the possibility of wind-driven waves, grasses alone were considered insufficient to stabilize the gravel sideslopes on the upstream side (southern) of this portion of the road. The sideslopes along this section of the road were covered with BioD-Mat 40 mesh blankets manufactured by RoLanka Coir Products (Stone Mountain, GA). Each blanket was approximately 10 feet wide, 130 feet long, and 0.35 inch thick, with a density of 13.6 ounces per square yard. Our installation method was designed to maintain the stability of the mesh blanket during periods of high water. The rolls of blanket were cut into sections of appropriate length, and the top of each section was buried in a 1-foot-deep trench excavated 2 feet below the shoulder of the road, to provide a secure anchor that did not interfere with routine road maintenance (Figure 6). The sections extended to the toe of the slope and overlapped by approximately 2 feet at the seams. The orientation of the seam overlaps varied depending on the expected direction of current flow, as determined by the locations of the bridges and culverts. Staples normally used to anchor erosion control blankets were not suitable for the unconsolidated gravel sideslopes. Instead, mesh blankets were anchored to the gravel embankments by driving 2-inch by 2-inch wooden stakes through the mesh at 1-foot intervals along overlapping seams of each blanket and at 1.5-foot intervals in the rest of each panel (Figure 7). We also used UV-resistant cable-ties to secure the mesh blanket to the stakes. Figure 8 shows the completed installation of the mesh blankets.
 |
| Figure 7. Wooden stakes for anchoring mesh blankets to the sideslope (left); Figure 8. Sideslope protected by blankets and riprap around a culvert (right) |
Once the mesh blanket was in place, we sowed grass seed as described previously. We also planted live willow stakes to increase the vertical structure of the vegetation on the sideslopes and provide increased erosion protection. The live stakes were feltleaf willow (Salix alaxensis), a common shrub species in early-successional riparian habitats throughout northern Alaska (Figure 9). We harvested willows from local populations along the Sagavanirktok River, approximately 70 miles east of the Alpine oilfield, under a collection permit from the Alaska Department of Natural Resources.
 |
| Figure 9. Close-up of transplanted feltleaf willow (left); Figure 10. Gravel-filled bags covering sideslopes along the airstrip. Note the layer of gravel over some of the bags (right) |
Dormant willow cuttings usually are planted in spring (May) in Alaska, but spring planting was not feasible at Alpine; the gravel sideslopes are often frozen until early June, because snow removal from the work surface creates a persistent snowpack on the sideslopes. Therefore we tried a new approach; cuttings were harvested and planted in early September 2001, soon after the onset of dormancy. Before planting, we treated the cuttings with a mixture of water and soil from the collection sites to “inoculate” them with soil microbes, including symbiotic fungi (mycorrhizae) that are important to the growth of arctic shrubs. We prepared 0.25- to 0.75-inch-diameter cuttings that were 16 to 18 inches long and planted them through the mesh blanket at a density of nine per square yard. We planted willows only on the lower 6 feet of the slope, where water availability during the summer was expected to be sufficient for successful establishment and growth of the live stakes.
Planting holes for individual cuttings were prepared using a locally manufactured willow planter, which was essentially a larger, sturdier version of a container seedling dibble bar. The willow planters were 6-foot-long sections of steel pipe with an inner diameter of 1.5 inches. The pipe on one end of a planter was cut in half lengthwise and sharpened into a pointed tip. Each planter had a 14-inch-wide T-handle at the top and an 18-inch-wide footstep just above the section of sharpened half-pipe. One person would drive the planter into the gravel and then remove it to leave a planting hole. Before the loose gravel collapsed, a second person would drop in a 21-gram fertilizer tablet (20-10-5 NPK), to ensure an adequate nutrient supply for the developing root system, and then insert one willow cutting. We left 2 to 4 inches of each cutting aboveground and tamped the gravel around the stem to ensure good soil contact and minimize voids in the rooting zone.
Two bridges (450 and 65 feet long) and six culverts (4 feet in diameter) were constructed in the swale area to allow floodwaters to flow through the road (Figures 3 through 5). Physical armoring was used to protect the sideslopes next to the bridge abutments and the culverts, where turbulence, currents, and the risk of erosion were expected to be high. Concrete revetment was used to cover 50 feet of the sideslopes adjacent to each side of the bridges (Figure 6). Rock riprap was used to cover 25 feet of the sideslopes adjacent to the entry and exit of each culvert (Figure 8).
 |
| Figure 11. Growth of non-palatable grasses along the sideslope of the airstrip three years after seeding (left); Figure 12. Growth of palatable grasses along the western embankment of the main camp pad (middle); Figure 13. View of the sideslope in swale area in 2004, three years after treatment with mesh blankets, transplanted willows, and seeded grasses (right) |
The other gravel sideslopes in other high-risk areas outside the swale area (Figure 4) were covered with rock riprap or with gravel-filled bags. Approximately 1,100 bags were used to cover 1.4 miles of gravel sideslopes. Each bag (Super Sacks, manufactured by B.A.G. Corp. in Dallas, TX) was filled with approximately 3 cubic yards of gravel and installed side by side to cover the full height of sideslopes (Figure 10).
In addition to treatments applied directly to the gravel sideslopes, we applied a granular fertilizer (20-20-10) at the rate of 400 pounds per acre to a 25-foot-wide buffer strip of the tundra surrounding all of the facilities. This treatment was intended to increase height and biomass of the existing vegetation and thereby provide additional erosion protection.
Assessing Treatment Performance
We selected a target level of 15% cover of the seeded grasses on the sideslopes, to be achieved within two years after seeding. This conservative goal was based on our experience at other sites with vegetation growing on nutrient-poor, gravelly soils, as well as the limitations on plant growth in the harsh environment of the Arctic. We considered that the attainment of this goal within two years would indicate that the grasses had established successfully and were developing at an acceptable rate. Other vascular plant species that naturally colonized the sideslopes were considered to provide the same protection as the grasses. Target levels for the transplanted willows after two growing seasons included a total density of seven (live or dead) stems per square yard and 30% survival.
Results from a bench-scale experiment in 1999 indicated that plants were less productive on the higher portions of the sideslopes, in large part because of drier conditions. Thick gravel provides a relatively poor growing medium by preventing plant roots from obtaining water from the surrounding tundra and by exposing plants to desiccating winds. Thus, to measure plant cover along a moisture gradient, we subdivided the sideslopes equally from shoulder to toe into upper, middle, and lower slope zones. We established one permanent sampling quadrat in each zone at 10 randomly selected locations along the airstrip, the swale area covered by erosion control blankets, and the remaining sideslopes, for a total of 30 quadrats. We recorded the species present at 20-centimeter intervals below the cross hairs of a 1-square-meter point-frame, for a total of 50 points per quadrat. All plant species were recorded at each point, including multiple layers (canopies); thus, total plant cover could exceed 100%.
Stabilization of Low-Risk Areas
Monitoring at the end of the second growing season (2003) indicated that the target level for plant cover had been achieved in most areas. Vegetation cover was even higher at the end of the third growing season (2004); the average grass cover along the airstrip was 101% in the lower and middle slope zones and 118% in the upper zone (Figure 11). This heavy cover of grasses, however, did not appear to attract wildlife that would pose a hazard to vehicles or aircraft. In the other low-risk areas, total live plant cover was 90% in the lower slope zone, 96% in the middle zone, and 48% in the upper zone (Figure 12). Whereas the vegetation in the middle and upper zones consisted entirely of the seeded grasses, 20% of the plant cover in the lower slope zone comprised native sedges and grasses, which appeared to have colonized by vegetative spread from the adjacent tundra. The relatively poor grass growth on the upper slope may have been caused by lower soil moisture, as well as the placement of a thin layer of gravel over some of the grasses by heavy equipment performing routine maintenance of the road and pad surface.
Stabilization of Moderate-Risk Area (Swale)
The mesh blanket has remained in place for three years without requiring maintenance, indicating that the installation was successful in withstanding the effects of spring flooding (Figure 13). Similar to the low-risk areas, the target level for cover was achieved after two years, and plant cover was even higher after three growing seasons. Average total plant cover in 2004 was 68% in the lower slope zone, 134% in the middle zone, and 84% in the upper zone.
Harvesting and transplanting dormant willow cuttings in fall proved to be an acceptable approach for areas where frozen ground prevents spring planting. After two growing seasons, the average density of willow stems (live and dead combined) was seven stems per square yard, and survival was 34%. Thus, the target levels for density and survival of the willow cuttings were achieved. After three growing seasons (2004), we did not observe any remaining dead willow stems, indicating that dead stems had been washed away. The average density of live willow stems in 2004 was 3.4 stems per square yard, and three-year survivorship was 37%. Although willows contributed only 3% to 5% of the ground cover after three growing seasons, we expect that shrub cover will increase with time as the rooting systems develop, contributing to the ability of the plant community to protect the gravel sideslopes from flood erosion.
The fertilization of the tundra vegetation adjacent to the gravel sideslopes resulted in vegetation that was visibly greener and taller compared with untreated tundra beyond the 25-foot buffer zone, in both 2003 and 2004 (Figure 13). This increased biomass of the natural tundra vegetation at the toe of the gravel sideslopes should help reduce water velocity and the severity of erosion.
Stabilization of High-Risk Areas
The rock riprap and underlying gravel installed along the approximately 0.6 mile of sideslopes have remained stable and have not required periodic maintenance. Similarly, the articulated concrete mats that cover the embankments at the two bridges have remained stable and required no maintenance, despite exposure to high floodwaters in 2000 and 2004. The gravel-filled bags also have proven effective at stabilizing the gravel sideslopes. Some bags had to be replaced, however, due to damage by heavy equipment during snow removal.
Summary
A combination of physical armoring and biotechnical stabilization has successfully protected the gravel sideslopes in the Alpine oilfield from erosion. Thus, biotechnical stabilization to control erosion of gravel sideslopes in the Arctic may provide a cost-effective alternative to physical armor, in areas where erosion risk is low to moderate. These techniques have the added benefit of extending the life of material sites since less gravel or rock is required for armoring.
Although the rooting system and aboveground structure of the seeded grasses stabilized gravel in low-risk areas during flooding, grasses alone did not prevent erosion because of surface runoff from precipitation on the flat gravel working surfaces of the pads, road, and airstrip. Surface runoff resulted in erosion “rills” that degraded the integrity of the sideslope and allowed gravel to slough onto the adjacent tundra. These areas have required periodic maintenance by heavy equipment operating on the gravel working surface to remove the sloughed gravel from the tundra and fill in the erosion rills from bottom to top. Unfortunately, most of the vegetation on the sideslope in the reworked sections is destroyed. An integrated strategy to control both flood erosion and runoff erosion is currently being developed.
In contrast, the combination of erosion control blankets, willow cuttings, and grass seeding successfully stabilized the gravel sideslopes by controlling erosion from flooding and from surface runoff without periodic maintenance. Results over the next few years will provide additional information on the relationship between plant performance and erosion protection, allowing us to refine our selection of biotechnical stabilization techniques for erosion control in the Arctic.