LINE CREEK LARGE WOODY DEBRIS ENHANCEMENT PROJECT – APPLYING LONG-TERM MONITORING AND RESEARCH TO HABITAT REHABILITATION DESIGN Robinson, M.D.*1, Bransfield, J.1, and M. Gaboury2 1. Lotic Environmental Ltd., Cranbrook B.C. 2193 Mazur Road Cranbrook, B.C. V1C 6V9 2. MN Gaboury and Associates Ltd., Nanaimo B.C. *Corresponding author mike.robinson@lotic.co ABSTRACT Line Creek is an important spawning tributary for Bull Trout (BT; Salvelinus confluentus) inhabiting the Elk River upstream of a migratory barrier at Elko BC. Ongoing fisheries monitoring has shown BT spawning to be associated with the glide (i.e. pool/tail-out) habitat typically created by large woody debris (LWD) jams. In June 2013, a large flood mobilized and removed much of the functional instream LWD from Line Creek. Results from annual monitoring confirmed the loss of LWD and associated pool habitat, as well as the lowest BT spawning counts on record. In 2014, we completed a multi-scale characterization of BT spawning locations in Line Creek to document habitat characteristics at watershed, reach, and channel unit scales. This study successfully described preferred spawning use in certain areas of the watershed, reaches of the creek, and specific habitat units. Hydraulic conditions were also described at observed spawning sites within individual habitat units. At the habitat unit scale, we were able to develop a Habitat Suitability Index (HSI) based on water depth, water velocity, and vertical hydraulic gradient. This information was applied to rehabilitation designs at a variety of scales. At the watershed scale, this study has guided work into portions of the watershed suitable to BT spawning (corroborated by previous assessments) and guided work along reaches that lack the preferred habitat attributes required to support spawning. At the habitat unit or rehabilitation structure scale, these results coupled with habitat assessments are informing the designs of LWD structures that mimic the natural jams displaced in June 2013. Rehabilitation work in Reach 1 of Line Creek was completed in summer 2017. Key Words: Large woody debris, Bull Trout, rehabilitation, spawning, habitat suitability, effectiveness monitoring. INTRODUCTION Line Creek is a high value stream in the upper Elk Valley of southeast British Columbia (Figure 1). Fish population and habitat monitoring of Line Creek dates back to the 1970’s, providing an impressive long-term dataset documenting fish use throughout the system. Most notable in the data is the consistent use of Line Creek for Bull Trout (BT; Salvelinus confluentus) spawning. Line Creek is suspected as being the most heavily utilized BT spawning stream in the upper Elk River system. It also contains a population of Westslope Cutthroat Trout (Oncorhynchus clarkii lewisi; WCT) that are thought to be primarily resident. In June of 2013, a large hydrological event occurred in the Elk River watershed, where approximately 250-300 mm of precipitation in the form of rain coincided with spring freshet (Pomeroy et al. 2015). This hydrological event resulted in widespread flooding, drastically impacting fluvial morphology throughout southeast British Columbia and southwest Alberta. In the case of Line Creek, this hydrological event substantially reduced the abundance and distribution of instream large woody debris (LWD). The removal of LWD resulted in a reduced channel complexity through a loss of pool habitat and increased frequency of riffle and cascade habitat. Furthermore, the habitat types lost were specifically those consistently documented as preferred BT spawning habitat. The morphology of a stream is greatly impacted by the presence and abundance of LWD structures. LWD structures provide channel stability and morphological diversity by manipulating fluvial processes. Scour-pool formation and sediment aggradation on alluvial bars aid in the formation of habitat for aquatic and terrestrial biota. Pool formation in these gravel bed fluvial systems is generally a product of scour effect from channel obstructions such as LWD, boulders or bank projections although LWD is the greatest influence (Montgomery et al. 1995; Abbe and Montgomery 1996). When complex LWD structures are perpendicular to the thalweg of a stream resulting in stream flow converge and turbulent velocities forming scour-pools. Research by Montgomery et al. (1995) demonstrate increased frequency of LWD positively correlated with pool frequency in forest channels of alluvial mountain streams in Washington state and Southeast Alaska. In addition to scour-pool formation Abbe and Montgomery (1996) also observed LWD as a functional component of riparian forest succession. The authors found apex bar jams slowed stream velocity when high flow events inundated alluvial bars, which led to aggradation of fine sediments and colonization of riparian vegetation (Abbe and Montgomery 1996). Aggradation of sediment on gravel bars leading to riparian forest succession can reduce the width of flow and increase the depth, effectively increasing habitat for fish. These morphological changes were most pronounced in Reach 1. In this reach these changes were suspected to be the cause for a decrease in BT spawning use. Prior to 2013, monitoring documented a preference for BT spawning habitat in Reach 1, relative to Reaches 2 and 3 located further upstream. After the June 2013 flood use of Reach 1 decreased by more than half. Reach-level preference of redd sites has fluctuated between Reaches 2 and 3 since 2013 (Robinson and MacPherson 2014; Robinson and Arnett 2015; Robinson and McKay 2016). Robinson and McPherson (2014) attributed this decline in redd frequency throughout Reach 1 to the loss of specific habitat units that would develop in the presence of LWD. The long-term monitoring had therefore identified potential rehabilitation. Restoring LWD to the system was identified as a desirable rehabilitation project to specifically improve BT spawning use. BT typically require a narrow suite of habitat parameters for spawning purposes, thus the integrity of their spawning habitat is highly important to the success of a population. Multiple studies throughout Western Montana have demonstrated that BT frequently select only 10% of the available spawning habitat to support up to 80% of total spawning (MBTSG 1998). In the upper Elk River watershed (i.e., above the waterfall barrier at Elko), Line Creek hosts the highest frequency of BT spawning of any tributary (Wilkinson 2009; Robinson et al. 2017). Its value to the fluvial population of the Elk River BT is therefore of high importance. Typical components of preferred BT spawning habitat include overhead cover, hyporheic flow, suitable substrate size, channel gradient, and water depth and velocity. In gravel bed streams, LWD plays an important role in the function of many of these habitat components. The decision to restore Reach 1 of Line Creek was made by considering multiple factors such as those previously mentioned (i.e., documented loss of LWD and decline in spawning use) but also because of research performed by Robinson et al. (2014). Their study examined the spatial selection of BT redds at the watershed, valley segment, reach, and channel unit scale. The study utilized a stream classification system created by Montgomery and Buffington (1998) to order the spatial scale at which they studied the stream. This report will utilize the same scale for descriptive purposes. Robinson et al. (2014) were able to discuss large-scale factors controlling BT spawning site selection at a watershed, as well as develop a habitat suitability model that used water depth, velocity and vertical hydraulic gradient to predict where, at a habitat unit scale, BT spawning was likely to occur. Pool-tailout (i.e., glide) habitat with adequate depth, velocity and downwelling were the most determinant factors in site selection for BT redds of Line Creek (Robinson et al. 2014). These results agree with previous research completed by Baxter and Hauer (2000). Within the channel unit the transition in bedform creates a pressure differential from the pool to tail out/riffle crest which forces oxygenated water down through the gravel, upwelling on the downstream end of the bedform (Bowerman et al. 2013). These hydraulic conditions form in the presence of LWD jams and were therefore what was targeted for rehabilitation. Figure 1. Site location map. Goals and Objectives In 2016, Teck Coal Ltd contracted Lotic Environmental to design a rehabilitation project for Line Creek – Reach 1 for construction in 2017. The primary goal of the project was to improve channel form and function of Line Creek – Reach 1 by restoring LWD and improving the quality of spawning habitat. It was expected that restoring LWD would also improve rearing and migration habitat. A secondary goal of the project was to accelerate colonization and natural succession of vegetation on in-channel bars by stabilizing unvegetated gravel bars using bar-top LWD jams. To complete these goals the following objectives were derived: 1. Install 13 in-stream LWD structures and: a. Observe return of natural fluvial processes such as scouring, LWD recruitment and increased pool-glide morphological complexity upon ≥1:2 year hydrological events. b. Observe increased use of restored habitat for spawning by BT (increasing observed redds) after a ≥1:2 year hydrological event. 2. Install 21 LWD Bar Jam structures and: a. Observe return of natural fluvial processes such as fine sediment deposition and natural soil formation on channel bars where Bar Jams were installed. b. Observe colonization and survival of vegetation (willow, alder, cottonwood) on bar tops. PRECONSTRUCTION ASSESSMENT Reach 1 has riffle-pool morphology and minor anthropogenic-caused changes to the physical habitat. The riparian is largely intact with only two stream crossings (one rail bridge and one forestry bridge). Channelization is also essentially non-existent. The stream has adequate floodplain connectivity, as indicated by fine sediment deposition within the riparian vegetation on nearly an annual basis. The hypothesis that BT spawning use was impaired in Reach 1 as a result of a loss of LWD and subsequent available habitat post-June 2013 was based on observations through annual monitoring. A series of preconstruction surveys were completed to support these observations. As well, detailed information on fish habitat was also collected to inform the rehabilitation design and to provide baseline data to assess the project against through time. Specifically the following surveys were completed. Topographic surveys were completed using real-time kinematic (RTK) GPS and traditional total station survey where GPS signal was not available. Topographic surveys described channel cross sections and longitudinal profiles within habitat unit types. Photo points were taken of representative habitat units. Permanent benchmark locations were also established. Fish habitat surveys followed the BC Fish Habitat Level 1 Fish Habitat Assessment Procedure (FHAP) (Johnston and Slaney 1996). FHAP surveys were completed in 2007 (Berdusco and Arnett 2008) and 2015 (McKay and Robinson 2016) to provide a pre- and post-flood comparison. The channel was first stratified into individual habitat units (e.g., pool, riffle, glide and cascade) through a continuous survey by a two- person crew with hip-chain and GPS. The continuous survey provided lengths of each habitat unit to provide an absolute estimate of linear proportions of each habitat unit. Complete sampling was used to sample every habitat unit. Each habitat unit was located using GPS coordinates. Pre- and post-flood channel complexity was assessed by comparing pre-flood data (Berdusco and Arnett 2008) to post-flood data (McKay and Robinson 2016). Percentage channel unit types were compared. These data will also be used for future monitoring of the rehabilitation efforts. Topographic survey and FHAP data were also used to describe channel condition using nine metrics derived through a combination of criteria in Johnston and Slaney (1996), Newbury and Gaboury (1994), and regional channel morphology relationships (Table 1). Qualitative ranks of poor, moderate, and high value were developed for each criterion through consultation with the Elk Valley Fish and Fish Habitat Committee1. Table 1. Qualitative scoring table of fish habitat and morphology of a stream reach, adapted from Johnston and Slaney 1996. Metric Poor Fair Good rank Bankfull width:depth ≥25:1 16-25:1 ≤15:1 Sinuosity <1.2 1.2 >1.2 Channel complexity <2 mesohabitat units/10xWbf 2-3 mesohabitat units/10xWbf ≥3mesohabitat units/10xWbf % Pool (by area) <15 15-40 40-60 Pool frequency (mean pool spacing) >10 channel widths/pool >8-10 channel widths/pool <8 channel widths/pool Holding pools/km (adult migration) >1 m deep with good cover (30% of pool area) <1 1-2 >2 LWD pieces per bankfull width <1 1-2 >2 Percent wood cover in pools Pools in reach average 0-5% cover Pools in reach average 6-20% cover Pools in reach average >20% cover FHAP surveys of Line Creek – Reach 1 confirmed the suspected change in distribution of habitat units (Table 2). Specifically of note was the conversion of glide habitat to riffle and cascade. Annual monitoring programs in Line Creek routinely report glides as the preferred habitat unit for BT redds to be constructed in as it provides the correct set of hydraulic conditions (Robinson et al. 2014). 1 A technical committee formed of representatives from Teck Coal Ltd, Fisheries and Oceans Canada (DFO), the Ministry of Forests, Lands, Natural Resource Operations and Rural Development, and the Ktunaxa Nation Council to review and discuss fish and fish habitat topics as they relate to Teck Coal’s Operations in the Elk Valley. Table 2. Habitat unit linear coverage throughout Reach 1 of Line Creek in 2007 (Berdusco and Arnett 2008) and 2015 (McKay and Robinson 2016). Habitat Type Reach 1 2007 2015 Difference (%) % Pool 1 1 0 % Riffle 58 61 3 % Glide 27 13 -14 % Cascade 14 25 11 Data analysis in 2007 was slightly modified from 2015 and 2017. As such, not all of the nine metrics were available for the pre-flood year. Bankfull width:depth, channel complexity (mesohabitat units/10xWb), % pool (by area), and LWD pieces per bankfull width were four of the nine metrics available for all three years of data (Table 3). Percentage pool was the only of these that remained unchanged since 2007 (poor for all three years). The other three metrics showed reduction in habitat value from 2007 to 2015, and remained low in 2017. Preconstruction surveys confirmed habitat changes suspected to have occurred as a result of the 2013 flood. A reduction in channel complexity and increased bankfull width:depth ratio were shown to coincide with reduced LWD frequency. This supports the concept that LWD is required in Reach 1 to increase habitat complexity by converting shallow, fast habitat units (i.e., riffles and cascades) to pool and glide habitats. Furthermore, through a combination of long-term monitoring and detailed description of the hydraulic conditions required for BT spawning (Robinson et al. 2014), it is suspected that these habitat changes are also adversely affecting use of Reach 1 by spawning BT. Combining long-term monitoring with detailed study supports the concept that BT spawning habitat in Line Creek – Reach 1 can be rehabilitated through the addition of instream LWD. Table 3. Qualitative scoring table of preconstruction fish habitat and channel morphology in Reach 1 of Line Creek. Metric 2007 ranking (value) 2015 ranking (value) 2017 ranking (value) Bankfull width:depth Good (14.9) Fair (18.3) Fair (15.9) Sinuosity - Good (1.4) Fair (1.3) Channel complexity (mesohabitat units/10xWb) Good (4.19) Fair (2.2) Poor (1.5) % Pool (by area) Poor (1%) Poor (1%) Poor (1%) Pool frequency (mean pool spacing, pool/Wb) - Poor (38.5) Poor (17.6) Holding pools (adult migration) - Moderate (1.7) Poor (0) LWD pieces per bankfull width Fair (1.42) Poor (0.1) Poor (0.4) % LWD cover in pools - Fair (10.0%) Poor (%) REHABILITATION DESIGN AND CONSTRUCTION The rehabilitation objective in Reach 1 of Line Creek was to re-establish a stable riffle-pool morphology within a slightly sinuous channel by constructing LWD Instream structures at meanders, and LWD Bar Jams situated at the upstream ends and mid-sections of existing gravel bars. A total of 13 LWD Instream structures and 20 LWD Bar Jams were proposed for this 1,200 m section of Line Creek. Field fitting relative to the number and siting of the structures at the time of construction was minimal as a post-freshet field reconnaissance was conducted before final ‘issued for construction’ drawings were prepared. All LWD was either western larch or Douglas fir. No cabling was completed on any structure. Construction was completed in August 2017. The LWD Instream structures were intended to cause pool scour and provide instream cover. Each LWD Instream structure was comprised of approximately 12 logs with rootwads attached. Preferentially, the logs were to be whole trees with rootwads and branches attached. The logs had an average diameter at breast height (dbh) of greater than or equal to 0.4 m and were approximately 12 m to 15 m in length. For some LWD handling and transporting limited overall length to 10-14 m. LWD Instream structures were constructed by interlacing the logs between live trees on the streambank (Figure 2, Figure 3, Figure 4). Figure 2. Typical LWD Instream structure design cross-section and plan view drawings. Figure 3. Upstream photo of completed LWD structure #1. Figure 4. Downstream photo of completed LWD structure #12. A total of 21 Bar Jam structures were constructed at two locations. LWD pieces were all between 10-12 m long by 0.4-0.6 m dbh. Each piece had the rootwad attached. Logs were placed parallel with the rootwad upstream and the upper portion of the tree downstream. The upper portion (~5-6 m) of each foundation log was embedded below the gravel bar surface a minimum of 1 m. The LWD Bar Jams situated at the upstream ends of the gravel bars were expected to help reduce channel braiding by deflecting low and moderate flows toward the proposed channel alignment. The LWD Bar Jams on the mid-sections of the gravel bars created roughness and complexity on the bar surface. Over the long term, this should reduce the likelihood of channel braiding and promote fine sediment aggradation on the bar surface to allow for greater survival of vegetation (e.g., willow, dogwood, cottonwood) plantings. Bar jams were installed on two separate bars. They were constructed as either single or multiple log structures. In total, 13 single-log and 8 multiple-log structures were constructed. Bar #1 included five double-log structures and three single-log structures (Figure 5). Bar #2 comprised of one large, six-log apex structure at the bar’s upstream end, and one double-log structure and seven single-log structures downstream (Figure 6). Figure 5: Completed bar jam structures, Bar #1. Figure 6: Completed bar jam structures, Bar #2. CONCLUSION The Line Creek Reach 1 Rehabilitation Project was successfully completed as per the project design. The addition of logs placed during this rehabilitation project has resulted in a 33% increase in instream LWD pieces in comparison to the pre-construction FHAP data. These structures are expected to increase in size through the natural recruitment of drifting LWD from upstream reaches. The LWD Instream structures are expected to provide numerous benefits, including increases in: habitat complexity, pool frequency, overhead cover in pools, holding pools (migration), and quality of rearing habitat. Habitat will require multiple high flow events in order for the instream LWD to create the scour needed to modify the instream habitat. Of particular interest will be the development of glides at the tail-out of scour pools. This habitat was specifically described as being key BT spawning areas and was the habitat unit that had the greatest decrease post-2013 flooding. The bar top jams are expected to result in deposition of fine sediment; building soil, establishing seedlings, and placing the gravel bar vegetation on a trajectory towards a climax community. Ultimately stabilizing bars is expected to promote a decrease in channel width:depth ratio, as well as provide valuable cover and nutrient inputs. Effectiveness monitoring will be completed 3-5 times over a minimum of 10 years post-construction, depending on flow events experienced. Monitoring will include physical, morphological and biological components. A drone flight should be completed during low flows in August – October, prior to snowfall. The flight can be used to effectively collect high resolution ortho-rectified imagery. Air photo analysis will allow for: counts of individual LWD, spatial distribution of LWD, and changes in channel pattern. The structural stability of each LWD Instream structure will be assessed following the data collection protocol provided in the Guidelines for In-Stream and Off-Channel Routine Effectiveness Evaluation (REE) (FIA 2003). Bar Jam LWD structures will be assessed through air photo analysis. The guidelines will be used to qualitatively assess the stability of each structure and their function, as well as additional habitat features such as scour pools and spawning gravel aggradation. The assessment will include establishment of ground level photo points as described in the REE document (FIA 2003). Two channel sections of approximately 20 channel bankfull widths in length will be identified for detailed monitoring to complement the morphologic data obtained from the drone flight. Detailed habitat surveys will include a topographic survey and FHAP-based measurements. Both surveys will be used to assess changes in channel morphology relative to as-built conditions and compared to the nine channel metrics listed during the pre-construction surveys. Cross-sections established during the as-built topographic surveys will be monitored. Substrate size distributions will be determined at representative riffles, pools, glides and runs to examine changes in channel geomorphology and spawning gravels over time. The annual Line Creek Aquatic Monitoring Program (Teck Coal Ltd) will continue to provide necessary biological monitoring components for the rehabilitation work completed. Specifically, the Program will complete spawning and adult rearing surveys on an annual basis. For BT spawning surveys, redd counts will be completed during BT spawning (late September) to index spawning effort through redd counts. Spawning surveys will provide a measure of adult BT use before and after rehabilitation. To estimate adult numbers, snorkel surveys will be conducted in late-August/early-September to specifically target Westslope Cutthroat Trout in Line Creek, as they are the primary resident fish present over summer rearing months. Both spawning and snorkel data will be compared to before and after treatment in Reach 1, as well as to non-rehabilitated reaches within Line Creek. Habitat rehabilitation sought to restore an abiotic component (LWD structures) of the ecosystem that is integral to fluvial morphological processes within Line Creek – Reach 1. To do so, this project employed a unique approach that coupled long-term monitoring and directed research to inform designs. Long-term monitoring allowed Habitat Biologists the unique opportunity to first identify reach-specific fish uses over decades. From this Habitat Biologists were able to identify likely factors limiting fish productivity, and furthermore, identify those with opportunity for rehabilitation. Lastly, the long-term dataset allowed Habitat Biologists to demonstrate that the system had historically functioned with higher productivity, meaning that the proposed rehabilitation was correctly addressing impaired ecological components. Direct research allowed the prescription to be taken further towards designing rehabilitation features that specifically addressed the suspected limiting factor. Long-term monitoring produced the conclusion that BT spawning had become compromised in Reach 1 due to a lack of glide habitat was an important observation. However, specifically identifying how BT redd selection was related to the habitat that formed in the presence of LWD provided strong support to the designs selected. The structures are expected to improve habitat complexity and fish use over time as they are exposed to successive hydrological events. Effectiveness monitoring will continue to document the function of these structures as well as the success of this rehabilitation project overall. LITERATURE CITED Abbe, T.B. and Montgomery, D.R., 1996. Large woody debris jams, channel hydraulics and habitat formation in large rivers. Regulated Rivers Research & Management, 12(23): 201-221pp. Baxter, C.V., and F.R. Hauer. 2000. Geomorphology, hyporheic exchange, and selection of spawning habitat by bull trout (Salvelinus confluentus). Canadian Journal of Fisheries and Aquatic Sciences. 57(7): 1470-1481 pp. http://www.env.gov.bc.ca/wat/wq/BCguidelines/turbidity/turbidity.html Berdusco, J. and T. Arnett. 2008. Line Creek Operations 2007 Aquatic Health Monitoring Program. Consultant report prepared for Elk Valley Coal Corporation – Line Creek Operations, Sparwood, BC. Prepared by Interior Reforestation Co. Ltd., Cranbrook, BC. 57 pp + 8 appds. Bowerman, T., Neilson, B.T. and Budy, P., 2014. Effects of fine sediment, hyporheic flow, and spawning site characteristics on survival and development of bull trout embryos. Canadian journal of fisheries and aquatic sciences, 71(7): 1059-1071 pp. FIA. 2003. Guidelines for In-Stream and Off-Channel Routine Effectiveness Evaluation. Prepared for the Forest Investment Account (FIA). 22 pp + append. Johnson, N.T., and P.A. Slaney. 1996. Fish habitat assessment procedures. Watershed Restoration Technical Circular No. 8. Watershed Restoration Program. Ministry of Environment, Lands and Parks. Province of British Columbia, Canada. McKay, A. and M.D. Robinson 2016. Line Creek Aquatic Monitoring Program (2015). Prepared for Teck Coal Ltd. – Line Creek Operations. Prepared by Lotic Environmental Ltd. 21 pp. and 6 appds. Montgomery. D. R., and J. M. Buffington. 1998. Channel processes, classification, and response in R. J. Naiman and R. E. Bilby (eds.), River Ecology and Management: Lessons from the Pacific Coastal Ecoregion. Springer, New York, NY. 13-42 pp. Montgomery, D.R., Buffington, J.M., Smith, R.D., Schmidt, K.M. and Pess, G., 1995. Pool spacing in forest channels. Water Resources Research, 31(4): 1097-1105 pp. Newbury, R.W. and M.N. Gaboury. 1994. Stream Analysis and Fish Habitat Design: A Field Manual. Second edition. Newbury Hydraulics, Gibsons, BC. 262 p. Pomeroy, J. W., Stewart, R. E., & Whitfield, P. H. 2016. The 2013 flood event in the South Saskatchewan and Elk River basins: causes, assessment and damages. Canadian Water Resources Journal/Revue canadienne des ressources hydriques, 41(1-2): 105-117 pp. Robinson, M.D., M. Gaboury, and E. Ellis. 2017. Line Creek Reach 1 Large Woody Debris Placement Conceptual Design Report. Prepared for Teck Coal Ltd. Lotic Environmental. Cranbrook BC. 10 pp + appends. Robinson, M.D., and T. Arnett. 2015. Line Creek Aquatic Monitoring Program (2014). Prepared for Teck Coal Ltd. – Line Creek Operations. Prepared by Lotic Environmental Ltd. 19 pp. and 3 appds. Robinson, M.D., Goodbrand, A., and R.J. MacDonald. 2014. Multi-scale Characterization of Bull Trout Spawning Locations in Line Creek. Prepared for Teck Coal Ltd by Lotic Environmental Ltd. 25 pp. Robinson, M.D., and S. McPherson. 2014. Line Creek Aquatic Monitoring Program (2013). Prepared for Teck Coal Ltd. – Line Creek Operations. Prepared by Lotic Environmental Ltd. 17 pp. and 2 appds. Wilkinson, C. 2009. Sportfish population dynamics in an intensively managed river system. MSc Thesis. University of British Columbia. 168 pp.