Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

57 3.1 Overview and Selection of Treatments for Testing 3.1.1 Overview As noted in Section 2.6.1, a wealth of literature and documentation is available regarding environmentally sensitive stream bank protection measures. However, most of the information regarding biotechnical measures is composed of case studies of particular sites that have, in many cases, limited general applicability. Furthermore, these case studies usually have a shortage of quantitative data regarding the hydraulic, hydrologic, climatic, and geotechnical conditions surrounding the tested sites. Advances in understanding the performance of nonliving bank- protection materials and structures have been based on laboratory flume tests that allow greater control and data acquisition than for field sites, but few of these experiments involved plant materials. Some tests have been conducted with grass-lined channels or with artificial plants made of wooden dowels, plastic strips, or other materials to investigate interactions between plants and the flow field. However, due to the difficulty of conducting scaled tests with real plants in available hydraulic laboratory flumes, only limited work has been done with real plants other than grass. Task 7 of this research addressed this deficiency by conducting well-planned, carefully controlled flume tests in the unique facilities at CSU. During Task 7, prototype-scale laboratory testing of selected biotechnical engineering treat- ments was conducted in the large (20 ft wide by 180 ft long) outdoor flume at CSU. Initially, the treatments were installed in large trays and nurtured in CSU’s climate-controlled greenhouses, which provide customized light, temperature, and humidity conditions for year-round estab- lishment and growth of many different types of vegetation. When vegetation was established to a predetermined condition, the trays were moved into the large outdoor flume for testing under the desired hydraulic conditions. 3.1.2 Proposed Treatments for Laboratory Testing Tests focused on two representative biotechnical measures that were constructed with real plants in large planter boxes (6 ft wide by 20 ft long by either 12 or 18 in. deep). Plant materials were given time to establish on the banks of an experimental trapezoidal channel prior to testing. A variety of environmentally sensitive bank-protection measures were considered as potential candidates for testing at prototype scale at CSU. Due to cost considerations, only two treatments could be accommodated under this research project. Therefore, the various treatments presented in NCHRP Report 544 (McCullah and Gray 2005) were carefully evaluated from the perspective of having wide applicability across the nation, as well as practical issues of: • Constructability, • Physical testing requirements/constraints, C H A P T E R 3 Testing and Appraisal of Testing Results

58 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures • Quantitative measurements of key hydraulic variables, and • Condition monitoring of each component before testing and after each test flow event. The two biotechnical bank-protection treatments suggested to, and approved by, the NCHRP Project 24-39 panel for testing were: 1. Live siltation with live staking and rock toe at a 3H:1V slope (Tray 1). 2. VMSE also referred to as FES lifts in this case at a 2H:1V slope (Tray 2). Figure 3.1 shows CSU’s greenhouse facility and vegetated trays used in the laboratory testing program. The willow components of these treatments were harvested locally and consisted of Salix exigua (also known as sandbar willow, narrowleaf willow or coyote willow), a shrub-type willow that is common to riparian corridors throughout most of the United States. Salix exigua has been specifically designated by the NRCS Plant Materials Program as well suited for erosion control: Sandbar Willow is used for stream bank and lake shore stabilization and riparian area development or restoration. It is recommended for deep wet lowland, overflow areas, wet meadow sites, stream banks, lake shores, and other areas with a high water table. (USDA 2007) Figure 3.1. (Upper) Climate-controlled greenhouses at CSU’s hydraulics laboratory. (Lower) Trays 1 and 2 inside the greenhouse.

Testing and Appraisal of Testing Results 59 Figure 3.2 shows native Salix exigua at a local site in Fort Collins, Colorado, from which the cuttings for the test installations were obtained. The two treatments described above, including planter box (tray) dimensions, are shown in Figures 3.3 and 3.4. The rationale for selecting these two treatments is discussed in detail in the following section. The approach adopted for prototype-scale laboratory testing of the selected bank-protection measures is presented in Section 3.2. 3.1.3 Selection of the Task 7 Testing Treatments All aspects of the proposed testing program were considered in selecting the two biotechnical treatments recommended, including physical restrictions of the greenhouse facilities where the treatments were to be grown, transport to the outdoor flume for testing, and the physical con- straints of the outdoor testing flume. The 44 treatments in NCHRP Report 544 (McCullah and Gray 2005) were reviewed and initially narrowed to 18 candidates. Based on the objectives of NCHRP Project 24-39, treatments that are intended to address geotechnical processes or that are redirective in nature (e.g., some river training structures) were eliminated from consideration, leaving only continuous techniques that protect the bank against fluvial erosion. For the Task 2 survey (see Section 2.3) the initial list of 18 NCHRP Report 544 treatments included live gully repair and joint planting. On closer examination, live gully repair was considered to be somewhat out of the scope of a stream bank protection project, as this treatment represents an application to halt or repair local erosion of the stream bank in a confined area (i.e., a gully) rather than an extended bank-protection treatment. Furthermore, joint planting was considered to be a variation of vegetated riprap, which was on the list for the survey, with the difference being that for joint planting the vegetation would be inserted in the joints of an existing riprap treatment as opposed to be being installed at the time of riprap installation for vegetated riprap. The hydraulic signature of both treatments would be similar once the planted or inserted willows in the stone have reached maturity. Accordingly, vegetated riprap was retained as a candidate but joint planting was eliminated which narrowed the “short list” to 16 treatments. Details on these 16 treatments for stream bank protection are summarized in Appendix A with reference to the McCullah and Gray (2005) findings regarding availability of hydraulic design data (hydraulic loading), common reasons for failure, and research opportunities/needs Figure 3.2. Salix exigua on Fossil Creek, Fort Collins, Colorado.

60 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures LIVE SILTATION 1.5 ft 1 ft Live siltaon willows Live willow stakes through coir mat Live stakes Live siltation willows with stone toe Figure 3.3. Live siltation with live staking and stone toe. identified in NCHRP Report 544. All of these factors as well as the extensive experience of research team members in designing, installing, and monitoring numerous biotechnical treatments in the field, in addition to agricultural engineering and vegetation planting and survivability concerns, were considered in the selection of the two treatments proposed for testing. The planter trays that have previously been constructed at CSU for the USACE levee overtopping testing are 6-ft wide, 20-ft long, and 12-in. deep. To avoid substantial construction costs it was imperative to keep within these basic dimensions for this project’s testing. The requirement to roll the trays out of the greenhouse and move them by crane approximately 250 ft to the outdoor flume also needed to be considered in terms of tray strength, lifting weight, and the 20-ft wide by 180-ft long dimensions of the flume. In developing the testing program, the discharge limit of about 150 ft3/s for the flume also had to be kept in mind. To this end, a hydraulic calculation spreadsheet was developed to provide discharge, velocity, and shear stress ranges for various configurations of the trays in the flume. An additional concern was the practicality of growing willows in a live siltation or staking configuration in a 12-in. deep soil lift. To accommodate a concern that the willows may become “root bound” the tray depth for the stone toe/live siltation configuration (Figure 3.3) was increased to 18 in. The increased soil depth increased the weight of the trays and required reinforcing of the USACE’s configuration. This also created a challenge for lifting and positioning the trays

Testing and Appraisal of Testing Results 61 in the testing flume. To keep Task 7 testing within the constraints of the planned budget the VMSE treatment was installed in an existing 12-in. deep tray (see Figure 3.4). In addition, the treatments needed to be grown in the greenhouse with the trays at the same slope as they will have in the flume so that the orientation of the willow plantings as they grow is consistent with the orientation that would be expected if they were planted in a field installation. This required some additional framing in the greenhouse facility, as the USACE’s treatments consist of various species of turf grass that are grown with the trays in a horizontal position. Also, the plantings needed to be rooted to a depth of about 2/3 of the height of the plant protruding into the flow to preclude physical pull out of the vegetation by hydraulic forces. As noted by Gray (2002): A conflict appears to exist between engineering requirements to compact soil to a high density to im- prove its engineering properties—such as increased strength and decreased compressibility—and agro- nomic needs to maintain soil in a relatively loose condition to improve its ability to support vegetation. This conflict or contradiction, while real, has been misunderstood and overstated. The objectives of com- paction from an engineering perspective have frequently been obscured in a manner that makes accommo- VEGETATED MECHANICALLY- STABILIZED EARTH VMSE protected stream bank (construction) VMSE protected stream bank (vegetated) 1.0 ft 1.5 ft 8” thick soil lis encapsulated by coir fabric Figure 3.4. VMSE.

62 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures dation for plant-growth needs more difficult to achieve. Furthermore, vegetation can be grown successfully in compacted soil under less-than-ideal conditions provided certain limits and precautions are observed. Based on guidance in the article quoted above and Goldsmith et al. (2001), the following criteria for soils in the treatment testing trays were followed: 1. Plants need soils compacted to less than 90% of the standard Proctor maximum dry density, with best conditions in the 80% to 85% range. 2. To retain hydraulic conductivity, soils should be compacted dry of optimum moisture content. 3. Absolute densities that limit plant growth vary by soil texture. Sandy soils can be denser. In general, growth limiting bulk densities are 87 lb/ft3 for clay soils and 106 lb/ft3 for sandy soils. With the physical constraints of the system identified, the treatments were considered from the perspective of having the widest applicability as prototypes or components of treatments that might be installed nationwide. The final considerations for the two recommended treat- ments involved both a review of the research needs identified in NCHRP Report 544 (see this report’s Appendix A) and the experience of the research team regarding resource agency prefer- ences as they have evolved as well as treatments that are currently being used in the widest variety of applications. The Hydraulic Engineering Circular No. 23 design guidelines for FHWA (Lagasse et al. 2009) and the guidance in NCHRP Report 544 indicate emphatically that, just as with traditional hard engineering treatments, the key to success for any treatment is that the toe must not fail if the treatment is to succeed. Consequently, it was desirable to test at least one treatment with a rock toe. It was also important to have a treatment that was composed of multiple components in an upslope direction. Transitions between components strongly influence the progressive roughness response from the high-velocity flow or splash zone to the less aggressive hydraulic conditions of the upper-bank slope zone. The detailed measurements in the flume provide insight on the hydraulic effects of these transitions, which are applicable to treatments composed of multiple material types. The initial concept for the stone-toe treatment was to transition from the stone-toe to willows and then to turf grass on the upper slope; however, from a species survivability point of view it was noted that as the willows mature they tend to shade and override the turf grass, making this a less than desirable combination. In addition, willows for live siltation and willow staking are becoming much more frequently used across the country (see Table 2.8). Accordingly, for this configuration it was decided to place the willows above the stone toe in a live siltation orientation and replace the turf grass concept for the upper slope with live willow stakes through a coir mat (see Figure 3.3). The hydraulic data from testing this configuration had broad applicability to either multi-component systems or treatments using just individual components (i.e., simply using live siltation or willow staking alone) for stream bank protection. The second proposed treatment recognizes that many resource agencies tend to prefer a treat- ment without any rock component. A treatment without a hard toe incorporating soil lifts and willow plantings is described in NCHRP Report 544 as VMSE as shown in Figure 3.4. The FES lifts can be tested within the confines of a treatment tray on a 1V:2H slope. A 10-degree back slope on the soil lifts was suggested and, to prevent pull out, the coir fabric would normally extend deeper into the embankment than would be possible with the trays; but, fastening the fabric to the tray base would prevent pull out and will not affect the hydraulic response and data acquired for this treatment. Since the resource agencies tend to discourage rock, and, in some cases, actually prohibit rock, testing the hydraulic limits of this second configuration would have wide potential applicability and interest. Appendix A contains a summary of both the hydraulic design data available in, and research opportunities recognized by, McCullah and Gray (2005) for the 16 treatments that were considered.

Testing and Appraisal of Testing Results 63 Several of the research opportunities that were addressed, at least in part, by the two treatments selected for testing under NCHRP Project 24-39 are summarized below: Live Siltation—Research into velocities that this technique can withstand would be helpful. Live Staking—Studies would be valuable regarding the effect live staking has on increasing the ability of other measures to withstand higher velocities and shear stresses. VMSE—Some uncertainty exists at present as to the exact permissible shear stresses and veloci- ties for VMSE interfaces. Additional research would also be helpful on the nature of the inter- action between roots and fabric and root architecture and distribution in VMSE structures. Since Task 7 was restricted by the available budget to testing only two treatments in the labo- ratory, the two treatments selected had the widest applicability to other treatment approaches and addressed a number of research needs identified in NCHRP Report 544. In reviewing the “short list” of 16 treatments from NCHRP Report 544, the Task 7 testing provided information relevant to live brush layering, VMSE, live staking, live siltation, live brush mattresses, live fas- cines, and ECBs. Thus, the treatments selected for Task 7 testing provided currently unavailable hydraulic design data related to seven of the 16 treatments. This maximized the return on the investment in the laboratory testing task of this research. 3.2 Laboratory Testing Plan The steps involved in installing, testing, and evaluating the bank-protection treatments shown in Figures 3.3 and 3.4 are described in this section. 3.2.1 Installation 1. The soil, rock, and coir fabric required for each specific treatment were installed in large planter trays in the CSU climate-controlled greenhouse (see Figure 3.1). 2. Live (dormant) willow cuttings, 3⁄8 to 5⁄8 in. diameter, were delivered to CSU in January, 2014. The cuttings were soaked for 3 weeks in aerated water at 37-degrees Fahrenheit prior to installation in the trays. 3. Soil used for the installation consisted of custom blended 20% compost, 40% sand, and 40% topsoil, measured by volume. A compost sample was delivered to the researchers on December 19, 2013. A compost maturity index of 7 on the Solvita scale was measured on December 26, 2013. Thus, the compost met the requirement of a “maturity index of 6 or greater as measured by the Solvita scale” (Brewer 2001). 4. Sand consisted of fine aggregate as defined by AASHTO standard M6. Grain size distribution (GSD) and Unified Soil Classification System (USCS) classification (SM—silty sand) were determined by an independent geotechnical firm. The median grain size was 0.73 mm. 5. The loamy topsoil component was screened through a 2.3 mm sieve to remove all large particles and deleterious materials. The resulting soil was approximately 68% sand and 32% silt and clay, and was classified as a sandy loam/loamy sand as defined by the USDA soil triangle. 6. Compost, sand, and loamy topsoil were combined at a 1:2:2 ratio by volume in a large container and blended until uniformly mixed. This mixing procedure was repeated until enough soil was obtained. A sample of this material was subjected to washed sieve analysis (ASTM C117 and C136), Atterberg limits (ASTM D4318), and Standard Proctor Density (ASTM D698) tests. The blended soil properties are: • D50 = 0.43 mm where D50 is the median particle size. • Liquid Limit = 28%. • Plastic Limit = 10%. • Plasticity Index = 18%. • Maximum Standard Proctor Density = 116 lb/ft3.

64 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures 7. Both Trays 1 and 2 were outfitted with substantial drainage ports in the bottom to allow for proper soil drainage. One inch of pea gravel was placed in the bottom of the box as an under- drain. A permeable geotextile was placed over the gravel layer to prevent migration of soil into the underdrain. The custom blended soil was then placed in 6-in. lifts and compacted to a target density of 83% ± 3% of Standard Proctor Density. 8. The willow cuttings were installed in the trays in early February 2014 and allowed to establish over a 7-month period. The soil in the trays was kept moist (but not saturated) using both drip and mist-type irrigation. The temperature was maintained at 80 to 90-degrees Fahrenheit, 75% or greater relative humidity, under grow lights that remained on approximately 16 hours per day for the first 8 weeks to ensure good establishment. 9. Tray 1: Willow spacing in Tray 1 consisted of two rows of live siltation willows. The first was placed immediately above the rock toe at a 45-degree angle to the plane of the slope, with cuttings spaced 2 in. apart. The second row was placed 1 ft upslope from the first. Live willow stakes were installed by piercing the ECB and hammering a metal stake (approximately 0.5-in. diameter) through the soil to create a pilot hole to the bottom of the box. Willow stakes were then inserted into the pilot holes. On the upper part of the slope, cuttings were placed in rows with 2-ft longitudinal spacing and 1-ft spacing up the slope in a staggered pattern. On this part of the slope, 10-in. long soil staples were added to hold the ECB in place. The staples were hand driven throughout the upper area of live staking at the same spacing as the live stakes (2-ft horizontal, rows offset 1-ft vertical) with the staples located midway between the willow cuttings. Cuttings were embedded the full 18-inch depth of Tray 1. Figure 3.5 shows the completed installation in Tray 1. 10. Tray 2: Willow spacing in Tray 2 consisted of rows of cuttings spaced 2 in. apart. Each row was sandwiched between successive lifts of fabric-wrapped soil 8 in. thick, having a back slope of 10 degrees into the slope. No soil staples were used in Tray 2. Cuttings were embedded the full 12-in. depth of Tray 2. No cuttings were placed into the top of the uppermost lift, nor into the vertical face of any lift. Figure 3.6 shows the completed installation in Tray 2. 11. Measurements of plant growth and root establishment for the vegetation were performed after 3 months and again after 7.5 months to ensure that a representative canopy and root system had established prior to testing. Figure 3.7 provides photographs of willow growth Tray 1 willow installation. Note rock toe in tray. Tray 1 willow installation showing growth after 4 weeks. Figure 3.5. Tray 1 installation.

Testing and Appraisal of Testing Results 65 determination and Table 3.1 presents the results of oven-dried weights of the biomass of the various above-ground and below-ground components. 12. After full willow establishment, the trays were moved by crane and placed in the outdoor River Engineering Flume at CSU with associated upstream and downstream transition sections, each 20-ft long, which provided artificial roughness elements to condition the flow entering and leaving the test section. The longitudinal slope along the test section and exit transition was 4.0%. 13. Between the toe of the tray and the flume wall, a layer of 6 in. D50 riprap was placed and grouted to establish a “stream bed.” The riprap was approximately 1 ft wide for the Tray 1 installation, and 1.5 ft wide for the Tray 2 installation. Tray 2 willow installation. Note no rock toe and no planting into top of uppermost lift. Tray 2 willow installation showing growth after 4 weeks. Figure 3.6. Tray 2 installation. Growth after 3 months. Growth after 7.5 months. Figure 3.7. Willow growth determination.

66 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures 14. Between the top edge of each tray and the opposite flume wall, metal flashing was used to seal the gap between the tray and the wall. The flashing was covered with a strip of coir fabric to give it the same roughness as the rest of the bed surface. Figure 3.8 shows both Trays 1 and 2 installed in the flume, prior to testing. 3.2.2 Testing Protocols and Data Collection 1. Cross-section surveys were performed prior to testing at pre-established transects located 4, 8, 12, and 16 ft downstream from the upstream edge of the test tray. 2. The flume flow and stoplog tailgate were adjusted (when necessary) to achieve the desired flow conditions. Three discharges were examined: a) A relatively low “mean annual” discharge of 50 cfs which only partially submerged the stream bank; Component Weight After 3 Months of Growth (grams) Weight After 7.5 Months of Growth (grams) Total biomass 506.3 1052.4 Above-ground biomass: 252.4 412.3 1.Leaves/shoots (new growth) 123.9 199.1 2.Original cuttings 128.5 213.2 Below-ground biomass 253.9 640.1 1.Roots (new growth) 94.8 277.1 2.Original cuttings 159.1 363.0 Ratio, new growth to original cuttings 0.76 1.28 Ratio, roots to shoots 0.77 1.39 Note: Weights after oven drying at 160 degrees Fahrenheit. Table 3.1. Willow growth after 3 months and after 7-½ months. Tray 1 (looking upstream). Tray 2 (looking downstream). Figure 3.8. Trays 1 and 2 prior to testing.

Testing and Appraisal of Testing Results 67 b) An intermediate flow rate of 100 cfs which just reached the top of the stream bank; and c) A “design” discharge of 150 cfs that fully submerged the entire bank slope. 3. During each flow, 1-D point velocity measurements were taken at each transect at 20%, 60%, and 80% of the total flow depth, and also as close to the bed as possible. The measurements were made at approximately 1-ft intervals across each cross section. 4. During each flow, 3-D acoustic Doppler velocimeter measurements were taken near the bed at selected locations where the probe could be positioned within the submerged willow vegetation. 5. After each flow, each treatment was examined for any damage to its various components (e.g., loss of vegetation, movement of rock, or soil loss) and cross-section surveys were repeated at each of the four transects. 3.3 Tray 1 Testing—Live Siltation and Live Staking with Riprap Toe Testing of the prototype-scale willow treatments was conducted at CSU’s hydraulics labora- tory during August and September 2014. Tray 1 was craned into place in early August 2014 and tested August 11–13. The water temperature during all tests remained at 45 to 46 degrees Fahrenheit. A summary of the three tests (Test Numbers 1 through 3) is provided below, and Figure 3.9 provides photographs of the tests in progress. Discussion of the data analyses and results is provided in Section 3.5. Test 1 was conducted on August 11, 2014 with the 3H:1V test section installed in the flume. The test discharge of 50 cfs ran for approximately 4.5 hours. The water surface came just to the top of the test tray but did not reach the flume wall on the upslope side of the test section. The live siltation willows along the lower portion of the slope were inundated and pronated in the flow, but the live staking willows on the upper portion of the slope remained upright for the duration of the test. Test 2 was conducted on August 12, 2014. Discharge through the test channel was 100 cfs and was sustained for 4 hours. Water reached the flume wall at the top of the slope, which inundated the top of the test tray to a depth of about 0.3 to 0.5 ft. Willows were pronated approximately two-thirds of the way across the flume. Test 3 was conducted on August 13, 2014, and was run for 4 hours at 150 cfs. Water reached the flume wall at the top of the slope, which inundated the top of the test tray to a depth of about 1.0 to 1.2 feet. Virtually all of the willows were pronated into the water during Test 3 (two stems at the top of the slope near cross section XS16 did not pronate). Erosion during Tray 1 tests: Overall, very little erosion occurred during Tests 1, 2, and 3 on Tray 1. On average, the elevations measured at the predetermined transects after Test 1 were 0.01-ft higher than the pre-test survey, indicating that overall, little or no soil material was lost from the test section. Some migration and rearrangement of soil beneath the coir/jute ECB was observed after each test, with local areas exhibiting, at most, 0.13 ft of degradation after Test 3; other local areas exhibited as much as 0.18 ft of aggradation. No damage to the coir/jute fabric was observed after completion of the three tests. 3.4 Tray 2 Testing—VMSE Without Hard Toe After the third and final test on Tray 1, it was removed from the flume and Tray 2 was installed. Tray 2 was tested during the period September 4–11, 2014. The water temperature during all tests remained at 45 to 46 degrees Fahrenheit. A summary of the three tests (Test Numbers 4 through 6)

68 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures is provided below, and Figure 3.10 provides photographs of the tests in progress. Discussion of the data analyses and results is provided in Section 3.5. Test 4 was performed on September 4, 2014 for a duration of 4 hours. A free fall occurred at the downstream end of the exit section. The discharge was approximately 50 cfs, with water reaching midway up the vertical face of the uppermost soil lift. The lowermost two rows of willows were pronated, while the uppermost two rows remained upright. A tear in the coir fabric, along with localized depressions indicating soil loss beneath the fabric at other locations, began forming on the test section during Test 4. The tear occurred in the face of the lowermost soil lift and was located 8 ft downstream of the upstream edge of Tray 2. The tear resulted in a 3-in. by 5-in. hole with 3 in. of soil loss measured back into the face of this lift, but was not apparent on the top surface of the lift. Tray 1, Test 1 (50 cfs) looking upstream Tray 1, Test 2 (100 cfs) looking upstream Tray 1, Test 3 (150 cfs) looking downstream Tray 1, post-test inspection after Test 3 Figure 3.9. Tray 1 during testing.

Testing and Appraisal of Testing Results 69 Test 5 was conducted on September 9, 2014 with a discharge of 100 cfs. Duration of the test was 4 hours. After Test 4, stoplogs were installed to a height of 17.5 in. above the flume floor to raise water levels on the exit section and the test section. Water completely inundated the test section and steel flashing, reaching the flume wall at the top of the slope and inundating the uppermost soil lift to a depth of about 0.8 to 1.0 ft. All the willows were pronated during this test except for a small group in the vicinity of cross section XS16 near the downstream end of the test section. A new, 3-in.-deep hole in the coir fabric formed during Test 5 in the top surface of the fourth soil lift approximately 6.5 ft from the upstream edge of the test section. The hole that was present in the face of the bottom soil lift after Test 4 was enlarged during Test 5, growing to 15 in. in length with a depth of about 4 in. Both large and fine willow roots were exposed in this hole, and the top surface of the soil lift began showing a depression. During Test 5, a significant amount soil within the uppermost soil lift was washed out, although no tears in the fabric were observed. Tray 2, Test 4 (50 cfs) looking downstream Tray 2, Test 5 (100 cfs) looking downstream Tray 2, Test 6 (150 cfs) looking downstream Tray 2, post-test inspection after Test 6 Figure 3.10. Tray 2 during testing.

70 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures Test 6 was conducted on September 11, 2014 at a discharge of 150 cfs for a duration of 4 hours. The stoplogs used in Test 5 were left in place for Test 6. The water surface completely inundated the test section to a depth of 1.2 to 1.4 ft above the uppermost soil lift and all willows were completely pronated during the test. Soil fill within the uppermost tier (Tier 5) was almost com- pletely washed away after Test 6, but no tears in the coir fabric were observed. On the fourth soil lift, the hole at 6.5 ft from the upstream edge was enlarged from the top surface into the vertical face. Depressions in the coir fabric up to 4 in. deep indicated soil loss at local areas, either on the top surface, the vertical face, or both. Erosion along the bottom soil lift included the major hole at 8 ft, with 6 in. of erosion into the face of the soil lift and further depression of the top surface. Erosion During Tray 2 Tests. During Test 4 on Tray 2 (50 cfs), little erosion occurred. The average erosion depth when compared with the pre-run survey was 0.02 ft. During the test, willows were pronated up to the second tier. However, it was noted that local tears and holes in the fabric, as well as other areas of soil loss depressions, began forming on the test section during Test 4, the largest of which was a 3-in. by 5-in. hole in the ECB with approximately 3 in. of soil loss. This hole was located on the bottom tier, 8 ft downstream of the upstream edge of the tray. Holes present after Test 4 were enlarged during Test 5 and grew even larger during Test 6. Both Tests 5 and 6 exhibited significant soil loss, and additional areas of damage to the fabric were observed. The photo in the lower right of Figure 3.10 shows a hole that began forming during Test 5 in the top of the fourth tier approximately 6.5 ft from the upstream edge of the test section. That hole grew to about 3 in. deep during Test 6 (shown in Figure 3.10) and the soil was eroded behind the ECB forming the vertical face of the soil lift. The hole in the face of the bottom soil lift grew to approximately 15 in. long after Test 6, with a depth of as much as 6 in. in places. This hole was located approximately 6 to 8 ft downstream of the upstream edge of the tray. Both fine and large willow roots were exposed in the hole. It is important to note that topmost (fifth) soil lift was entirely submerged during Tests 5 and 6. No vegetation was planted into the top surface of this lift, and most of the soil fill within this lift was washed out after Test 6. In Section 3.5, erosion contour maps for each test, created from post-test surveys at predetermined transects, clearly show the overall erosion patterns as well as localized areas of soil loss whether due to tears/holes in the coir fabric or soil washout beneath the fabric. 3.5 Testing and Data Summary 3.5.1 Calibration of Manning n Values Measured water surface elevations in the vegetated test sections of Trays 1 and 2 were used to calibrate HEC-RAS models of each test. Fixed flume components (flume walls and rock riprap stream bed) were modeled using estimated Manning n values of 0.015 and 0.035, respectively. The artificial roughness of the upstream approach section was modeled assuming a Manning n ranging from 0.055 to 0.08 to match the upstream water surface elevation measured by CSU. The vegetated slope of the test section was modeled using a Manning n value that was adjusted by trial and error to achieve the best fit to the observed water surface for all of the tests. Figures 3.11 and 3.12 show the results of the calibration efforts for Trays 1 and 2, respectively. The distance between HEC-RAS computational cross sections ranged from 0.5 to 1.0 ft. In general, the Manning n resistance coefficient was found to be significantly higher for Tray 2 compared to Tray 1. This is presumably due both to the overall density of the cuttings used between

Testing and Appraisal of Testing Results 71 93 94 95 96 97 98 99 100 0 5 10 15 20 25 30 35 40 El ev a on ,  Distance,  Obs WSEL, 50 cfs Obs WSEL, 100 cfs Obs WSEL, 150 cfs Pred. WSEL 50, n = 0.035 Pred WSEL 100, n = 0.035 Pred WSEL 150, n = 0.030 Bed Approach (arficial roughness) Vegetave treatment Figure 3.11. HEC-RAS calibration of Manning n values, Tray 1. 92 93 94 95 96 97 98 99 100 0 5 10 15 20 25 30 35 40 El ev ati on , ft Distance, ft Obs WSEL, 50 cfs Obs WSEL, 100 cfs Obs WSEL, 150 cfs Pred. WSEL 50, n = 0.050 Pred WSEL 100, n = 0.050 Pred WSEL 150, n = 0.045 Bed Approach (artificial roughness) Vegetative treatment Figure 3.12. HEC-RAS calibration of Manning n values, Tray 2.

72 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures the soil lifts in Tray 2, as well as the irregular slope geometry due to the stair-step configuration of the lifts. Table 3.2 provides a summary of the HEC-RAS calibration runs on the vegetative treatments in Trays 1 and 2. The Froude numbers in Table 3.2 indicate that all three tests of Tray 1 were conducted with supercritical flow conditions, whereas the Tray 2 tests were conducted at subcritical to near- critical conditions. The range of cross-section average velocities in Tray 1 is much greater than that in Tray 2 because of the continued acceleration of flow in the downstream direction in Tray 1. The slopes of the energy grade lines in the table also indicate that the Tray 2 tests resulted in near- uniform flow because the energy slopes are very near the nominal bed slope of 0.04 ft/ft. The highest discharge (150 cfs, Tests 3 and 6) resulted in lower energy slopes compared to the smaller flows, and therefore lower Manning n values, which presumably is due to the total pronation of the willow stems. Note that the longitudinal bed profile in Figure 3.11 reflects the riprap “stream bed” surface because it was the lowest component in the surveyed elevations, whereas the lowest component in Figure 3.12 was the edge of the test tray. 3.5.2 Point Velocity Measurements and Velocity Distributions During each test, three streamwise point velocity measurements were taken (20%, 60%, and 80% of the total depth of flow) with a Marsh-McBirney 1-D electromagnetic flowmeter. In addition, at most locations it was possible to obtain a fourth point velocity reading within about 2 in. above the bed surface. The meter was mounted on a point gage and suspended from a data collection cart which traversed the length of the test section on rails mounted on top of the horizontal flume walls. The velocity data were collected at predetermined cross sections located approximately 4, 8, 12, and 16 ft downstream of the leading (upstream) edge of each test tray. Typically, a set of data was taken above the rock riprap “stream bed,” and proceeded at about 1-ft intervals up the vegetated slope at each cross section, adjusted as necessary based on vegetation clusters. The following series of figures provides the contours showing velocity distributions correspond- ing to cross sections 4, 8, 12, and 16, respectively, as Figures 3.13 through 3.16. In these figures, the velocity distributions for each flow (50, 100, and 150 cfs) are shown, with Tray 1 results presented in the left-hand column, and the corresponding plots for Tray 2 shown in the right-hand column. During the low-flow tests (50 cfs, Tests 1 and 4) on both Trays 1 and 2, the velocity contour plots show the effectiveness of the willows in pushing the higher velocity flow away from the bed, particularly near the toe of the slopes. This effect is somewhat less pronounced at the higher Discharge, cfs Manning n of Vegetated Bank Range of Energy Grade Slope, ft/ft Range of Velocity1, ft/s Range of Froude Number TRAY 1 50 0.035 0.021 – 0.037 6.6 – 8.0 1.08 – 1.41 100 0.035 0.017 – 0.032 7.6 – 9.8 1.05 – 1.41 150 0.030 0.013 – 0.025 9.2 – 11.6 1.06 – 1.47 TRAY 2 50 0.050 0.042 – 0.046 6.1 – 6.3 0.89 – 0.91 100 0.050 0.041 – 0.044 7.1 – 7.3 0.92 – 0.94 150 0.045 0.037 – 0.041 8.6 – 8.8 0.99 – 1.03 1Cross-sectional average velocity from HEC-RAS (V = Q/A). Table 3.2. Results of HEC-RAS calibration runs.

Testing and Appraisal of Testing Results 73 CROSS SECTION 4 TRAY 1 TRAY 2 TEST 1: 50 ft3/s TEST 4: 50 ft3/s TEST 2: 100 ft3/s TEST 5: 100 ft3/s TEST 3: 150 ft3/s TEST 6: 150 ft3/s 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 50 ft3/s50 ft3/s 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 150 ft3/s 100 ft3/s100 ft3/s 150 ft3/s Figure 3.13. Velocity contours at Cross Section 4, Trays 1 and 2.

74 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures CROSS SECTION 8 TRAY 1 TRAY 2 TEST 1: 50 ft3/s TEST 4: 50 ft3/s TEST 2: 100 ft3/s TEST 5: 100 ft3/s TEST 3: 150 ft3/s TEST 6: 150 ft3/s Distance across cross section, ft El ev at io n, ft NO DATA 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 50 ft3/s 150 ft3/s 100 ft3/s 50 ft3/s 150 ft3/s 100 ft3/s Figure 3.14. Velocity contours at Cross Section 8, Trays 1 and 2.

Testing and Appraisal of Testing Results 75 CROSS SECTION 12 TRAY 1 TRAY 2 TEST 1: 50 ft3/s TEST 4: 50 ft3/s TEST 2: 100 ft3/s TEST 5: 100 ft3/s TEST 3: 150 ft3/s TEST 6: 150 ft3/s Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft Distance across cross section, ft El ev at io n, ft Distance across cross section, ft El ev at io n, ft Distance across cross section, ft El ev at io n, ft 50 ft3/s 150 ft3/s 100 ft3/s 50 ft3/s 150 ft3/s 100 ft3/s Figure 3.15. Velocity contours at Cross Section 12, Trays 1 and 2.

76 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures CROSS SECTION 16 TRAY 1 TRAY 2 TEST 1: 50 ft3/s TEST 4: 50 ft3/s TEST 2: 100 ft3/s TEST 4: 100 ft3/s TEST 3: 150 ft3/s TEST 6: 150 ft3/s 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 Distance across cross section, ft El ev at io n, ft 0 1 2 3 4 5 6 7 8 2 3 4 5 6 7 Distance across cross section, ft El ev at io n, ft 50 ft3/s 150 ft3/s 100 ft3/s 50 ft3/s 150 ft3/s 100 ft3/s Figure 3.16. Velocity contours at Cross Section 16, Trays 1 and 2.

Testing and Appraisal of Testing Results 77 flows (100 and 150 cfs), although the live siltation willows near the toe of the slope in Tray 1 were more effective at producing this condition. This can be seen by comparing the velocity contour pattern in Tray 1 at Cross Section 4 vs. the patterns at Cross Sections 8, 12, and 16. This appears to be a result of the flow transitioning further downstream into the vegetative treatment. The point velocity at 60% depth (V60) is generally considered to represent of the depth-average velocity in the water column at that location. The measured V60 values from the Tray 1 and Tray 2 datasets were used with the Manning equation to estimate bed shear stress at each of the measure- ment locations for all tests, as described in Section 3.5.3. 3.5.3 Bed Shear Stress Local shear stress at the bed was estimated at all point velocity measurement locations for all tests by using a rearranged form of the Manning equation: nV 1.486 (3.1)0 60 2 1 3 τ =     γ y where: t0 = Bed shear stress, lb/ft 2 n = Manning n resistance coefficient V60 = Depth-averaged velocity, taken as the point velocity at 60% depth g = Unit weight of water, 62.4 lb/ft3 y = Depth of flow above bed, ft Using the point velocities at 60% depth, the bed shear was calculated using Equation 3.1 and the corresponding values were contoured as a function of the X-Y location within the vegetated test section, where X is the longitudinal (streamwise) direction and Y is the lateral direction across the flume. Figures 3.17 through 3.19 provide the contours showing bed shear stress distributions cor- responding to each of the test flows; they present the shear stress contour plots for each flow (50, 100, and 150 ft3/s). In each figure, Tray 1 results are presented as the upper portion of the figure and Tray 2 results are shown as the lower portion of the figure. The riprap area(s) near the toe/streambed are indicated by dashed lines, as are the areas of metal flashing at the top of the slope. The flow direction is from left to right. As seen in Figures 3.17 through 3.19, the calculated bed shear stress is significantly different between Tray 1 and Tray 2 for the three flow rates. The color intensity is consistent for all of the contour plots shown in these figures. Table 3.3 provides a comparison of typical shear stress ranges on the vegetated portions of the test sections in Trays 1 and 2. 3.5.4 Erosion Using the measured bed elevations at each of the four predetermined transects after each flow, the cumulative erosion was calculated and the corresponding values were contoured as a func- tion of the X-Y location within the vegetated test section, where X is the longitudinal (stream- wise) direction and Y is the lateral direction across the flume. Figures 3.20 through 3.22 provide the contours showing measured erosion corresponding to each of the test flows; they present the erosion contour plots for each flow (50, 100, and 150 ft3/s). In each figure, Tray 1 results are presented as the upper figure and Tray 2 results shown as the lower figure. The riprap area(s) near the toe/streambed are indicated by stippled regions,

78 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures Tray 1, Test 1 (50 ft3/s) Tray 2, Test 4 (50 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl u m e , ft Distance along flume, ft Flow Figure 3.17. Shear stress contours, Trays 1 and 2 at 50 ft3/s. Figure 3.18. Shear stress contours, Trays 1 and 2 at 100 ft3/s. Tray 1, Test 2 (100 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow

Testing and Appraisal of Testing Results 79 Tray 2, Test 5 (100 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl u m e , ft Distance along flume, ft Flow Figure 3.18. (Continued). Tray 1, Test 3 (150 ft3/s) Tray 2, Test 6 (150 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl u m e , ft Distance along flume, ft Flow Figure 3.19. Shear stress contours, Trays 1 and 2 at 150 ft3/s.

80 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures Discharge, ft3/s Typical Range of Bed Shear Stress, lb/ft2 Minimum Bed Shear Stress, lb/ft2 Maximum Bed Shear Stress, lb/ft2 TRAY 1 50 0.50 to 0.75 ~ 0.25 ~ 1.00 100 1.50 to 2.00 ~ 1.00 ~ 2.25 150 1.50 to 2.00 ~ 1.25 ~ 2.25 TRAY 2 50 1.25 to 2.00 ~ 1.00 ~ 2.25 100 1.75 to 3.00 ~ 1.00 ~ 3.50 150 2.50 to 4.00 ~ 1.75 ~ 4.50 Table 3.3. Results of bed shear stress analyses. Tray 1, Test 1 (50 ft3/s) Tray 2, Test 4 (50 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl u m e , ft Distance along flume, ft Flow Figure 3.20. Erosion contours, Trays 1 and 2 at 50 ft3/s.

Testing and Appraisal of Testing Results 81 Tray 1, Test 2 (100 ft3/s) Tray 2, Test 5 (100 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow Figure 3.21. Erosion contours, Trays 1 and 2 at 100 ft3/s. Tray 1, Test 3 (150 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Riprap toe in tray Metal flashing D is ta nc e a cr o ss fl um e, ft Distance along flume, ft Flow Figure 3.22. Erosion contours, Trays 1 and 2 at 150 ft3/s. (continued on next page)

82 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures as are the areas of metal flashing at the top of the slope. Because no erosion occurred on areas of riprap or metal flashing, the contoured portion of the figure is blanked out in these areas. Negative values on the contour lines indicate erosion, while positive values indicate aggra- dation due to rearrangement of soil beneath the coir/jute fabric. The flow direction is from left to right. In Figures 3.20 through 3.22, the change in bed elevation is significantly different between Tray 1 and Tray 2 for the three flow rates. The color intensity is consistent for all of the contour plots shown in these figures. Table 3.4 provides a comparison of measured erosion depths in Trays 1 and 2 corresponding to the different discharges. The contour maps in Figures 3.21 and 3.22 clearly indicate the severe washout of soil from within the uppermost soil lift of Tray 2 during the 100 and 150 ft3/s flows. The soil loss in the uppermost lift was evident along nearly the entire length of the test tray, as described previously in Section 3.4. A photo of the uppermost soil lift after Test 6 (150 ft3/s) is shown in Figure 3.23. Note that no willow staking was done in the top surface of this lift. Figures 3.21 and 3.22 also indicate the development and growth of the tear in the coir/jute fabric forming the face of the lowermost soil lift in Tray 2. The tear was initiated during Test 4 and began sagging the top surface of this lift due to soil loss during Tests 5 and 6. Tray 2, Test 6 (150 ft3/s) 0 2 4 6 8 10 12 14 16 18 20 -8 -6 -4 -2 0 Riprap stream bed Metal flashing D is ta nc e a cr o ss fl u m e , ft Distance along flume, ft Flow Figure 3.22. (Continued). Discharge, ft3/s Typical Range of Bed Erosion, ft Minimum Bed Erosion, ft Maximum Bed Erosion, ft TRAY 1 50 -0.025 to +0.025 ~ +0.050 ~ -0.025 100 -0.050 to -0.000 ~ +0.050 ~ -0.150 150 -0.150 to -0.025 ~ +0.075 ~ -0.300 TRAY 2 50 -0.050 to -0.000 ~ +0.100 ~ -0.075 100 -0.150 to -0.000 ~ +0.100 ~ -0.300 150 -0.400 to -0.100 ~ 0.000 ~ -0.550 Note: Positive values indicate aggradation (deposition) beneath fabric; negative values indicate erosion. Table 3.4. Results of bed erosion analyses.

Testing and Appraisal of Testing Results 83 3.6 Appraisal of Testing Results The quantitative analyses of velocity, shear stress, and erosion data presented in the previous sections support the qualitative observations made during the course of the testing. Tray 1 (3H:1V bank slope with live staking, live siltation willows, and riprap toe within the testing tray) exhibited significantly less erosion, lower bed shear stress, and lower Manning n values compared to Tray 2 (VMSE soil lifts with no hard toe). 3.6.1 Erosion vs. Shear Stress Although the calculated bed shear stress and the measured erosion data were highly variable over the vegetated treatment areas of both trays as evidenced by the contour plots presented pre- viously, trends are apparent in the datasets. The average erosion depths and corresponding shear stresses for Trays 1 and 2 after each test are presented in Table 3.5 and Figure 3.24. The average values of these variables are computed from all V60 point velocity and bed surface measurements taken over the entire test section. Figure 3.24 suggests that a threshold shear stress of about 0.9 lb/ft2 is required to initiate erosion. Once erosion begins, the erosion depth appears to be a linear function of shear stress. Further trend analysis was conducted by collating the shear stress and erosion data at all of the measurement points by partitioning each cross section into four separate stream tubes (measured from left flume wall looking downstream, adjacent to the “stream bed”), as follows: • Stream Tube 1: 0.0 to 2.0 ft • Stream Tube 2: 2.0 to 4.0 ft Figure 3.23. Tray 2 at the end of Test 6 (150 ft3/s). Note soil washout and collapse of the uppermost soil lift.

84 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures Test No. Discharge, ft3/s Average Shear Stress, lb/ft2 Average Erosion Depth, ft Tray 1 1 50 0.75 0.01 2 100 1.50 0.06 3 150 1.70 0.10 Tray 2 4 50 1.61 0.03 5 100 2.43 0.13 6 150 3.24 0.24 Table 3.5. Average values of shear stress and erosion, all tests. y = 0.094x - 0.080 R² = 0.910 0.00 0.10 0.20 0.30 0.0 1.0 2.0 3.0 4.0 A ve ra ge E ro si on D ep th , ft Average Shear Stress, lb/ft2 50 cfs 100 cfs 150 cfs Figure 3.24. Average erosion depth vs. average shear stress, all tests. • Stream Tube 3: 4.0 to 6.0 ft • Stream Tube 4: 6.0 to 8.0 ft This partitioning is necessary because the velocity data at a particular cross section were not collected at the same lateral distance (y-coordinate) as the erosion measurements. In most cases, the number of velocity measurements in any given stream tube was not the same as the number of erosion measurements. Thus, in order to pair the data, the maximum erosion rate within each stream tube was paired with the maximum shear stress in the same stream tube. Erosion rate in inches per hour was determined by taking the erosion depth and dividing by the duration of each test (4 hours, with the exception of Test 1 which was run for 4.5 hours). The results of this partitioning approach are shown in Figures 3.25 and 3.26 for Trays 1 and 2, respectively. In these figures, the x- and y-axes scales are consistent to facilitate comparison. The partitioning better shows the variability of erosion rate and shear stress and more clearly shows the clustering of data for the different discharges used in the full-scale laboratory testing program. Comparing Figures 3.25 and 3.26, it can be seen that the stream-tube maximum shear stress in Tray 1 did not exceed a value of 2.6 lb/ft2 for any of the tests, whereas it reached a maximum of about 5.4 lb/ft2 in Tray 2.

Testing and Appraisal of Testing Results 85 E = 0.144τ 0.0 0.5 1.0 1.5 2.0 2.5 0.0 1.0 2.0 3.0 4.0 5.0 6.0 M ax im um S tr ea m T ub e Er os io n Ra te E , i n/ hr Maximum Stream Tube Shear Stress τ, lb/ft2 50 cfs 100 cfs 150 cfs Figure 3.25. Erosion rate vs. shear stress, Tray 1. E = 0.232τ 0.0 0.5 1.0 1.5 2.0 2.5 0.0 1.0 2.0 3.0 4.0 5.0 6.0 M ax im um S tr ea m T ub e Er os io n Ra te E , i n/ hr Maximum Stream Tube Shear Stress τ, lb/ft2 50 cfs 100 cfs 150 cfs Figure 3.26. Erosion rate vs. shear stress, Tray 2.

86 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures Likewise, the maximum erosion rate in any stream tube in Tray 1 was 1.0 in. per hour, while in Tray 2 the maximum erosion rate was 2.0 in. per hour. Therefore, in general, doubling the shear stress results in a doubling of erosion rate. This again suggests a linear relationship between erosion and shear stress. While correlation between maximum stream tube erosion rates and shear stresses is poor, trends are apparent in both Trays 1 and 2. The data indicate that erosion begins to occur when a threshold shear stress of about 1.0 lb/ft2 is exceeded. This is very similar to the averaged data of Figure 3.24; however, no predictive relationship is implied by presenting the data in this fashion. Shear stresses and erosion were least at the lowest flow rate (50 ft3/s) in both trays. In Tray 1, the cluster of erosion vs. shear stress points is similar for the 100 and 150 ft3/s flows. However, in Tray 2, the data clusters are different for the two higher flows. A possible explanation for this may be that the construction of the vertical lifts of the VMSE in Tray 2 created preferential path- ways for higher velocity flow to occur against the face of each lift. Figure 3.27 shows a close-up of preferential flow pathways along the faces of soil lifts in Tray 2 before, during, and after testing. 3.6.2 Pronation of Willows As discussed in Sections 3.3 and 3.4, willow stems were bent in the direction of flow during the tests to different degrees based on depth of inundation as well as local velocity impinging on the vegetation. During each test, the lateral distance from the left flume wall to the point where the willows remained upright was noted (at the highest flow, 150 ft3/s, all the vegetation was pronated and lying beneath the water surface in both trays). The data from all tests were segregated into two groups: willows standing upright and willows laid over. The depth was then plotted vs. velocity at 60% depth (V60) for every data point as shown in Figure 3.28. The product of depth and velocity is unit discharge, and is often used as a mea- sure of flood hazard to humans in inundated areas. A unit discharge of 5.3 ft3/s/ft was found to distinguish between pronation (solid symbols) and non-pronation (open symbols) of the Salix exigua (sandbar willow) used in this testing program. Sandbar willow is extremely resilient and pliable, and resisted damage at all ranges of flow conditions tested. Characteristics of the willows at the time of testing included the biomass measurements presented in Table 3.1, typical stem diameters of 3⁄8 to 5⁄8 in., and typical above- ground stem heights of 3 to 5 feet. Figure 3.29 is a close-up photograph of the only stem that actually broke during the testing program (Test 6 at 150 ft3/s, near the toe of the slope at Cross Section 16 in Tray 2). The point velocity V60 and shear stress at this location during Test 6 were 11.4 ft/s and 3.5 lb/ft2, respectively. 3.6.3 Summary of Laboratory Testing Program From the observations, data collected, and subsequent data analyses of the full-scale tests, the following observations are made: 1. The vegetative treatment of Tray 1 (live siltation willow and live staking with stone toe) exhibited significantly lower Manning n resistance coefficients, bed shear stresses, and ero- sion compared to Tray 2 (VMSE soil lifts with soft toe) at the same discharge. 2. A higher density of willows per square yard of surface area does not necessarily result in bet- ter performance. Rather, the overall geometry of the bank slope and planting configuration appears to be the more important factor in the overall performance of the vegetative com- ponent. For example, the stair-step configuration of the VMSE creates preferential pathways for high-velocity flow, which results in the potential for damage to the fabric of the soil lifts.

Testing and Appraisal of Testing Results 87 a. Prior to testing (photo taken in greenhouse). b. During Test 6. Note pronation of willows and exposed faces of soil lifts. c. After Test 6. Note exposed faces of soil lifts. d. After Test 6. Note deformed face of soil lift (originally vertical). Figure 3.27. Preferential flow pathways (yellow arrows) along the faces of soil lifts in Tray 2.

88 Evaluation and Assessment of Environmentally Sensitive Stream Bank Protection Measures 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 De pt h,  Velocity at 60% depth, /s Tray 1, 50 cfs laid over Tray 2, 50 cfs laid over Tray 1, 100 cfs laid over Tray 2, 100 cfs laid over Tray 1, 150 cfs laid over Tray 2, 150 cfs laid over Tray 1, 50 cfs standing up Tray 2, 50 cfs standing up Tray 1, 100 cfs standing up Tray 2, 100 cfs standing up Unit discharge = 5.3 cfs/ Willows laid over Willows standing up Figure 3.28. Depth vs. velocity curve distinguishing pronated vs. non-pronated willow stems. Figure 3.29. Photograph showing broken stem at the end of Test 6 (150 ft3/s).

Testing and Appraisal of Testing Results 89 3. The live siltation willows at the toe of the bank slope in Tray 1 were significantly more effective at shifting the region of high-velocity flow away from the stream bank and toward the main channel compared to the stair-step configuration of soil lifts and vegetation of the VMSE treatment in Tray 2. 4. Vulnerable areas of the VMSE treatment are the faces of the individual soil lifts. In particular, the lowermost lift at the toe of the bank slope is subject to damage if left unprotected by other means, such as a stone (i.e., riprap), toe armor, or other “hard” engineered treatment. Qualitative observations suggest that preferential pathways for high-velocity flow along the faces of the soil lifts appear to be a shortcoming of this bank-protection treatment. 5. The uppermost soil lift of VMSE is also vulnerable to soil loss if it becomes fully submerged and is not vegetated on its top surface. In contrast, the top surfaces of soil lifts that are lower down on the slope were seen to be protected by the pronated willows of the lift above, which provided a “shielding” effect. 6. For both bank-protection treatments, the threshold shear stress at which erosion begins is approximately 1.0 lb/ft2. 7. The shear stress corresponding to local areas of significant soil loss (up to 0.25 ft) was approximately 2.5 lb/ft2 in both trays. This threshold was reached in Tray 1 during Test 3 (150 ft3/s) and in Tray 2 during Test 5 (100 ft3/s). 8. Areas of excessive soil loss ranging from 0.3 to 0.5 ft were observed in Tray 2 during Test 6 (150 ft3/s). The corresponding shear stresses for this condition were 3 to 4.5 lb/ft2. 9. Because of the difference in Manning n roughness between the two treatments, limiting values of permissible velocity were found to be 7 to 9 ft/s for Tray 1, and 5 to 7 ft/s in Tray 2. 10. For both treatments tested, there is a large amount of variability in bed shear stress and erosion from point to point along the bank slope. This was seen in both the longitudinal (downstream) flow direction as well as the lateral (upslope) direction.