LIVE BRUSH MATTRESS
bar
 


1. CATEGORY

2.0 – Bank Armor and Protection

2. DESIGN STATUS

Level I

3. ALSO KNOWN AS

Brush Mat

4. DESCRIPTION

A live brush mattress, or brush mat, consists of a thick (15 to 30 cm (6 to 12 in)) blanket of living cuttings and soil fill placed on a streambank or lake shore to simultaneously revegetate and armor the bank.

5. PURPOSE

Brush mattresses are a bank armoring technique for scour control and revegetation that is constructed using live willow branches or other species that root easily from cuttings. The dense layer of brush increases roughness, reduces velocities at the bank face, and protects the bank from scour, while trapping sediment and providing habitat directly along the waters' edge.

6. PLANNING

Useful for Erosion Processes:
Toe erosion with upper bank failure
Scour of middle and upper banks by currents
Local scour
Erosion of local lenses or layers of noncohesive sediment
Erosion by overbank runoff
General bed degradation
Headcutting
Piping
Erosion by navigation waves
Erosion by wind waves
Erosion by ice and debris gouging
General bank instability or susceptibility to mass slope failure

Spatial Application:
Instream
Toe
Midbank
Top of Bank

Hydrologic / Geomorphic Setting
Resistive
  Redirective
Continuous
  Discontinuous
Outer Bend
Inner Bend
  Incision
  Lateral Migration
  Aggradation

Conditions Where Practice Applies:

Brush mattresses are appropriate for eroding streambanks or slopes where immediate protection is needed from flooding stream flows or wave-induced erosion. Willow (Salix spp.) is the most common plant material used for brush mattresses because of its rooting ability. This practice is suitable for streams where willow is naturally occurring and where the soil and moisture conditions are favorable.

Brush mattresses are planted relatively shallow, as compared to brushlayering, therefore it is most successful on streams where the basal ends of the cuttings will be kept moist during most of the growing season, but flows do not exceed the tolerance of the structure (Allen & Fischenich, 2001).

Complexity:

Moderate.

Design Guidelines / Typical Drawings:

The optimal bank slope for brush mattresses is 1V:2H, because stem to soil contact can be maximized at that angle; however, mattresses can successfully be installed at angles of 1V:1.25H or steeper, but sprouting will occur mostly at the basal ends. In some cases, fill will be required to bring the bank to the desired grade. If rock fill is used, at least 45 cm (18 in) of soil should be placed over the rock to ensure proper stem to soil contact for the cuttings (Allen & Fischenich, 2001).

It is important to protect the brush mattress against flanking and undermining; some type of toe protection is necessary, and depending upon the erosivity of the bank, keys or refusals may be necessary at upstream and downstream ends. Rock toe protection is useful with brush mattresses. If there is any overbank runoff occurring, flows should be diverted around the brush mattress and outleted in a stable area. If piping is evident, a granular filter should be installed underneath the brush mattress (Allen & Fischenich, 2001).

Allen & Fischenich (2001) discuss the installation of two or more rows of brush mattress on banks with long slope lengths, in which the terminal ends of the bottom (streamward) row overlap the basal ends of the upper row by at least 30 cm (1 ft). It should be noted, however, that the survival of cuttings that do not have their basal ends near the annual low water level is questionable in arid and semi-arid environments.

Brush Mattress Typical Drawing

7. ENVIRONMENTAL CONSIDERATIONS / BENEFITS

Brush mattresses provide a dense network of branches that quickly stabilize a slope or streambank. As the live branches root and grow, not only do they provide cover, but the soil is reinforced with an underground matrix of spreading roots. If used on streambanks, a brush mattress will trap sediments during high water and eventual plant growth will enhance aquatic habitat. If used on slopes, a brush mattress collects soil, providing germination sites for other plants.

8. HYDRAULIC LOADING

Studies have shown that brush mattresses have stabilized a bank in a test flume against velocities exceeding 7 m/s (20 ft/s) (Gerstgrasser, 1999).

Allen and Fischenich (2001) report the following velocity and shear stress data:

Brush Mattress Type

Velocity

Shear

Mattress without rock toe, initial

<2.7 m/sec
(<4 ft/sec)

2.0 – 14.6 kg/m2
(0.4 – 3.0 lb/ft2)

Mattress without rock toe, grown

<3.4 m/sec
(<5 ft/sec)

19.5 – 34.2 kg/m2
(4.0 – 7.0 lb/ft2)

Mattress with rock toe, initial

3.4 m/sec
(<5 ft/sec)

3.9 – 20.0 kg/m2
(0.8 – 4.1 lb/ft2)

Mattress with rock toe, grown

8.2 m/sec
(<12 ft/sec)

19.5 – 39.0 kg/m2
(4.0 – 8.0 lb/ft2)

Florineth (1982) provided additional shear force tolerances for brush mattresses without a rock toe: 20.5 kg/m2 (4.2 lb/ft2) just after construction, 30.8 kg/m2 (6.3 lb/ft2) after 15 months, and 41.0 kg/m2 (8.4 lb/ft2) after the third year.

Figure 1. Studies conducted by Dr. Christoph Gerstgraser, Universitat fur Bodenkultur, Vienna, Austria, demonstrated that brush mattresses (on right bank) stabilized the bank in a test flume against velocities exceeding 7 mps (20 f/s), while other techniques, even rock riprap, failed.

Figure 2. Dr. Christoph Gerstgrasser describes the shear forces and velocity that the brush mattress survived. Photo by J. McCullah

9. COMBINATION OPPORTUNITIES

Brush mattresses are well suited for combined installation with many other streambank or slope stabilization techniques. They are often combined with Vegetated Riprap, Live Stakes, Live Fascines, Rootwad Revetment, Live Siltation, and Coconut Fiber Rolls.

10. ADVANTAGES

11. LIMITATIONS

Brush mattresses do not show high success on streams where basal ends cannot be kept wet for the duration of the growing season. In addition, they should be installed during the dormant season for woody vegetation. Installation is labor intensive.

12. MATERIALS AND EQUIPMENT

Brushy cuttings (stems having leaves and twigs) of tree and shrub species capable of propagating from cuttings, typically willow (Salix spp) species (see Special Topic: Harvesting and Handling for more information on harvesting techniques and storage methods). 10 to 50 branches per m (3 to 15 per ft) of bank to be protected should be harvested. The cuttings should be long (1.5 to 3 m (5 to 10 ft)), straight, brushy, 2 to 3 year old branches up to 4 cm (1 ½ in) in diameter. For optimum success, the fascines should be soaked for 24 hours or installed on the same day they are harvested and prepared (McCullah, 2004).

Wooden construction stakes and/or live stakes will be needed. The length of stakes will vary based on soil conditions. Biodegradable natural fiber or polypropylene rope is usually preferable to wire in most situations. In addition, a sledgehammer will be needed for driving in wooden stakes, or a dead-blow mallet and pilot bar (rebar) for live stakes (McCullah, 2004).

13. CONSTRUCTION / INSTALLATION

Prepare the slope or streambank by clearing away large debris and grading the slope so the branches will lay flat on the bank. If bank is not graded evenly, air pockets will form during backfilling, causing poor stem to soil contact, and ultimately resulting in poor sprouting success. Do not compact the slope over 85%, as it will inhibit rooting. Excavate a horizontal trench, 20 to 30 cm (8 to 12 in) deep at the toe of the streambank (Allen & Fischenich, 2001; McCullah, 2004).

Lay the cuttings flat against the graded slope, slightly crisscrossed, with the basal ends placed as deeply into the trench as possible, and just below any toe protection to be installed. Continue to lay the cuttings along the face of the bank or slope until 80% groundcover is achieved (the mattress will be about 6-30 cm (2 to 12 in) thick). It will take 10 to 50 branches per m (3 to 15 branches per ft) of mattress (Allen & Fischenich, 2001; McCullah, 2004).

Pound a grid of stakes, 60 to 90 cm (24 to 36 in) long, into the mattress at 0.9 to 1.2 m (3 to 4 ft) centers. Do not pound the stakes completely in, as this will be done after tying. Longer stakes can be used in sandy soil and shorter stakes in heavy soils. Secure the brush mattress by using cord, rope (a machine-spun bristle coir, 5 to 6 mm (.2 to .25 in) thick, with a breaking strength of 32 to 45 kg (70 to 100 lb) is suitable), or 10 to 12 gauge, galvanized annealed wire, tied with clove hitches in a diamond pattern between each row of stakes (Allen & Fischenich, 2001; Johnson & Stypula, 1993; McCullah, 2004).

After securing the mattress with cord or wire, drive the stakes in further to compress the mattress tightly against the slope. Secure the toe of the mattress using a suitable technique such as Vegetated Riprap, Live Fascines, Rootwad Revetments or Coconut Fiber Rolls (Allen & Fischenich, 2001; Johnson & Stypula, 1993; McCullah, 2004).

Backfill around and in between the branches of the mattress by using material excavated from the trench, and additional soil if needed, working the soil in well around the branches. Tamp soil by walking on it, and lightly water the soil with buckets or a hose to wash it down into the stems, and ensure good stem to soil contact. It is necessary for the thicker, basal ends of the mattress to get good soil cover for rooting; at least 1/4 of the depth of the mattress is recommended. Leaving some branches exposed above the soil will facilitate sprouting (Allen & Fischenich, 2001; McCullah, 2004).

14. COST

Depending upon slope or streambank characteristics and amount of cuttings required, 1 to 5 work hours/m2 (0.1 to 0.5 work hours/ft2) will be needed for installation.

Allen & Fischenich (2001) report a cost of approximately ½ to 1½ labor hours per m2 (4.6 to 13.9 work hours per 100 ft2) for plant harvest, collection, transport, fabrication and installation. Costs for other necessary materials, such as stakes and twine, were an additional $3 to $5 per m2($28 to $46.5 per 100 ft2). They reported further that the average cost of brush mattress projects they were involved in was about $18 per m2 ($1.67 per ft2), although projects going to contractor bid typically cost between $25 and $55 per m2 ($2.3 to $5.1 per ft2).

15. MAINTENANCE / MONITORING

During the first growing and flood season, periodic maintenance is necessary to make sure the stakes and cord/wire are still securing the mattress to the streambank, and to verify that flows are not getting behind the mattress.

16. COMMON REASONS / CIRCUMSTANCES FOR FAILURE

Flanking or undermining of the revetment due to a lack of toe protection, or lack of upstream and downstream keys.

17. CASE STUDIES AND EXAMPLES

Johnson Creek

The Oregon Department of Transportation realigned a section of Johnson Creek in order to widen a bridge and complete highway construction. The realignment shortened the creek by approximately 20%, thereby increasing its gradient. Riprap was installed to the ordinary high water elevation in the outside bends, and the banks above were stabilized with vegetated mechanically stabilized earth. Live siltation was installed on the lowest floodplain berm adjacent to the sub-channel to provide cover for waterfowl and overhanging cover for fish. The upper bank was protected with a brush mattress.

Belle Island, Detroit, MI

In the fall of 2003 brush mattresses were installed on a lake shore on Belle Island as part of a three-day training course.

Belle Isle training site before the training. The slope was regraded to 1V:2H.

Willow branches were laid with their basal ends into the water and anchored with construction stakes. Photo by D. Gray

Brush mattress nearing completion. Mattress was covered with soil and straw mulch. October 2003. Photo by J. McCullah

Willow sprouting through straw 7 months after construction. Photo by D. Gray

Pit River, Likely and Canby, Modoc County, CA

The Pit River in Modoc County in North Eastern California, is situated in a high elevation desert characterized by cold winters, low precipitation with very heavy clay soils near the river. See Restoration in the Pit River Watershed, Hoyer, et.al. (2003)

Willow brush mattress and live siltation installed on a restored bank of the So. Fork Pit River. Flournoy Ranch-Reach A. February 2001. Photo by J. McCullah

Flournoy Ranch-Reach A after 7 months growth. September 2001. Photo by J. McCullah

Flournoy Ranch-Reach A after 18 months July 2002.
J. McCullah

Bushey Ranch-Reach A. This 190 m (623 ft) reach received brush mattress with rock toe treatment. November 2000. Photo by J. McCullah

Bushey Ranch-Reach A. The stakes were difficult to drive into the bank (heavy clay) and the basal ends of the willows were too high above AHW.

Bushey Ranch-Reach A after 3 years is stable but the willow growth is sparse. November 2003. Photo by J. McCullah

Bushey Ranch-Reach E. Along this reach the willow mattress was constructed below AHW at ALW. November 2000

Bushey Ranch-Reach E. The Pit River remained low the first winter. However the branches were “watered in” and soon received some snow cover.

Bushey Ranch-Reach E after 3 years. The willow vegetation covers over 60% of the reach. July 2003. Photo by J. McCullah

It was conjectured that better handling procedures, because of the semi-arid conditions, may have resulted in better establishment.

Please visit the Photo Gallery for more pictures.

18. RESEARCH OPPORTUNITIES

Additional information regarding the required soil moisture and depth of soil fill for successful vegetation establishment would be valuable to obtain. A study of the effects of slope steepness on brush mattress growth and establishment would be useful as well.

19. REFERENCES
Allen, H. H. & Fischenich, J. C. (2001). Brush Mattresses for Streambank Erosion Control. EMRRP Technical Notes Collection (ERDC TN EMRRP- SR-23), U.S. Army Engineer Research and Development Center, Vicksburg, MS. (pdf)

Florineth, F. (1982). Experiences with bioengineered measures for watercourses in mountains. Landschaftswasserbau, TU Wien, 3, 243-262.

Gerstgraser, C. (1999). The effect and resistance of soil bioengineering methods for streambank protection. Proceedings of Conference 30, International Erosion Control Association.

Gray, D. H. & Leiser, A. (1982).   Biotechnical Slope Protection and Erosion Control. Van Nostrand Reinhold, New York, N. Y.

Johnson, A.W. and J.M. Stypula. eds. (1993). Guidelines for Bank Stabilization Projects in the Riverine Environments of King County. King County Department of Public Works, Surface Water Management Division, Seattle, Wash.

McCullah, J. A. (2004).  Bio Draw 3.0.  Salix Applied Earthcare, Redding, CA