Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 612 2015 SOUTHEASTERN NATURALIST 14(4):612–622 Effectiveness of a Stream-Restoration Effort Using Natural Material Instream Structures Sean E. Collins1, Joseph E. Flotemersch2,*, Casey D. Swecker3, and Thomas G. Jones4 Abstract - The substrata of fluvial systems can be altered by human disturbance in watersheds. This disturbance often results in a reduction of habitat diversity and subsequent reductions in species diversity. Restoration efforts in impacted areas require a thorough understanding of the characteristics of exemplary stream habitat in the region as well as habitat requirements of taxa targeted by specific restoration efforts. The Little Coal River, WV, has historically been disturbed by various land-use practices resulting in near homogeneity of the riverbed substratum such that it is composed almost everywhere primarily of fine-particle or sand-substrate classes. Restoration of an 8-km section of the Little Coal River was attempted with the installation of a series of natural material instream-structures. We monitored the riverbed substratum, including sediment size-class data, prior to and after installation of these structures for a 2-year period to evaluate their effectiveness in restoring overall habitat heterogeneity. Our GIS analysis of the data suggested that approximately 80% of the riverbed substratum was composed of fine-particle or sand substrate classes prior to the natural material structure addition. After 2 years, these 2 substrate classes had decreased by nearly 25%, suggesting that restoration efforts reduced the overall percent composition of fine-particle and sand-substrate classes and increased overall habitat heterogeneity. Our analysis of these data indicate that the natural material instream structures installed in the Little Coal River achieved the objective of promoting downstream movement of some fineparticle and sand substrate that characterized the system (i.e., via sediment transport) and increased substratum heterogeneity. Introduction The size and structure of riverbed substrate are determined by a number of factors including current velocity, rock type, and geological history (Benke and Cushing 2005), as well as land use and large instream structures (e.g., large wood). Natural stream and river channels evolve over time toward a balance of deposition and transport. Changes that disrupt the balance result in an unstable system (Leopold et al. 1964). Human interactions with watersheds often impact the balance of deposition and transport and result in changes to substrate composition. Beyond altering the size and structure of riverbed sediments, this interaction may also interfere with the basic structure and function of the affected ecosystem. 1Division of Science and Mathematics, Lees-McRae College, Banner Elk, NC 28604. 2National Exposure Research Laboratory, US Environmental Protection Agency, Ecological Exposure Research Division, 26 West Martin Luther King Drive, Cincinnati, OH 45268. 3Environmental Solutions and Innovations, Inc., 4525 Este Avenue, Cincinnati, OH 45232. 4Department of Integrated Science and Technology, Marshall University, Huntington, WV 25755. *Corresponding author - flotemersch.joseph@epa.gov. Manuscript Editor: Paul M. Stewart Southeastern Naturalist 613 S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 In the US, a 2008–2009 national survey of the condition of streams and rivers detected excessive levels of streambed sediment in about 15% of river and stream lengths (USEPA 2013). The report also stated that poor biological conditions were 60% more likely to exist in rivers and streams where excessive levels of fine streambed sediments were found. On a global scale, UNESCO (1984) listed the evolution of riverbed substrates, their sedimentation, and the influences of direct and indirect drivers as the third of 9 major scientific hydrological challenges that required further study. In general, sediment pollution is a function of disturbance in the watershed (Terrell 2011). Anthropogenic sources of this disturbance include common land-use practices such as logging, mining, agriculture, and urban development. Stream-bank erosion can also contribute considerable amounts of fine sediments when increased discharge exacerbates natural erosional processes (Waters 1995). Deposition occurs when increased input of fine sediments to a channel is greater than the stream flow can transport. As a result, habitats become increasingly homogenous, which negatively affects the composition of aquatic communities (Henley et al. 2000, Poff et al. 2007). In 2000, the West Virginia Department of Environmental Protection (WVDEP) identified several sections of the Little Coal River (roughly half the total stream length) as impacted by a variety of activities (e.g., resource extraction, highway development) and pollutants (e.g., pathogens, fecal coliform bacteria, acid mine drainage) (WVDEP 2000). The Little Coal River remained listed as impacted in the subsequent 6 biennial reports (WVDEP 2002, 2004, 2006, 2008, 2010, 2012). Substrate- composition surveys of the Little Coal River revealed it to be homogenous in many places and dominated by sand or fine-particle substrate. Physical habitat assessments in sections of the river detected low substrate-diversity, and bioassessment surveys found depauperate species diversity (T.G. Jones, pers. observ.), a finding consistent with those of Williams (1980), who, in an experiment performed on the east branch of Duffin Creek, ON, Canada, noted that more macroinvertebrate taxa were found where substrate heterogeneity was highest. In 2007, WVDEP supported a large-scale project to install several natural material instream rehabilitation structures (e.g., rock, large wood; as described in Rosgen and Silvey 1996). The project objectives included reducing streambank erosion, facilitating sediment transport, and enhancing fish habitat. We positioned these Rosgen-type structures in an 8-km section of the Little Coal River, WV, to mimic naturally occurring rock and debris formations in streams. We installed the structures to disrupt the prevailing laminar flow patterns and promote downstream movement (i.e., via sediment transport) of some fine-particle substrate and sand that characterized the system. Further, we intended the structures to promote sediment deposition on upstream and downstream banks while narrowing the river channel to promote increases in flow rates that would function to help maintain substrate heterogeneity through time. Herein, we evaluate the effectiveness of the installed structures at increasing substrate heterogeneity in sections previously dominated by sand or fine-particle substrate. Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 614 Figure 1. The Little Coal River. (a) The approximate location (outlined in red) of the Little Coal River within WV; (b) the location (outlined in red) of the study area between Lincoln and Kanawha counties; and (c) aerial imagery surrounding the section of interest, the shapefile outlining the bankful width of the study area in blue, and the approximate location and type of management structures installed in this study. Material and Methods Study area The Little Coal River, a 5th-order stream in the Appalachian Plateau physiographic province of southern West Virginia (Fig. 1), begins at the confluence of Southeastern Naturalist 615 S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 Spruce Fork and Pond Fork in Boone County. It flows northward through Boone, Lincoln, and Kanawha counties with a total drainage area of ~205 km2. It is a major tributary of the Coal River, which has a watershed area of ~2232 km2 and flows into the Kanawha River. The Coal River includes ~141 km of stream, making it the second-longest river completely contained within West Virginia. Based on the first sampling event (described below), riverbed substrate composition on the Little Coal River was fairly homogenous and dominated by sand or fine-particle substrate classes; however gravel, cobble, and boulder substrate classes were also present along the river. Other similar streams in the Appalachian Plateau predominantly contained carbonate (e.g., limestone, dolostone) and clastic (e.g., sandstone, shale) rock (Messinger and Hughes 2000). We installed 10 natural material instream structures within the 8-km zone targeted for restoration (Fig. 1). These included 3 J-hook and 3 cross-vane structures (Rosgen 2001), and 3 boulders and 1 wing-deflector (Fig. 1). J-hook structures, typically placed on the outside of stream-bends, are designed to reduce bank erosion and nearbank stream velocity (Rosgen 2001). Cross-vane structures are designed to reduce near-bank stream velocity while increasing the energy in the main channel (Rosgen 2001). Wing deflectors are used to decrease the width of over-widened streams. Sediment sampling We collected sediment data along the 8-km section of the Little Coal River (Fig. 1) before construction of the instream structures in 2007 (hereafter, 2007a). To ensure consistency in data quality, we collected sediment data using the same methodology after structure addition in summer 2007 (hereafter, 2007b), 2008, and 2009. We collected data using the sounding-rod method (Collins and Flotemersch 2014). This procedure consists of using a capped hollow copper (or aluminum) pipe to determine substrate type at ~100 points per km in a zigzag pattern sampled along a continuum from bank to bank (as recommended by Bevenger and King 1995). We repeatedly struck the substrate with the capped pipe at each sampling point to obtain tactile and auditory cues relating to the substrate composition. This method is reasonably accurate and precise, with >80% of samples allocated to the correct sediment class (Collins and Flotemersch 2014). The substrate survey resulted in a sample of ~800 points per sampling event, and a total sample of ~3200 points across all 4 events. At each point, we recorded substrate as belonging to one of 6 common classes: fine-particle, sand, gravel, cobble, boulder, or bedrock. These substrate classes were based on the Wentworth (1922) scale, which is commonly used for sediment analysis. We recorded the geographical position of each data point using a Garmin eTrex® Venture handheld GPS unit. We detected change in the sediment-size distribution using Kolmogorov- Smirnov tests. We determined the proportional contribution of the 5 sediment-size classes (excluding bedrock) from the ~800 samples for each sampling event. We used the sediment-size distribution from 2007a as the expected distribution and compared it to each of the 3 post-construction samples. We set a significance level of α = 0.05 to detect differences; sample size was set to n = 800 for each sample. Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 616 GIS analysis We managed georeferenced data (using the UTM system) in a Microsoft Access database (Office 2007; Microsoft Corporation, Redmond, WA). We gave each substrate class an integer value, where 1 = fine-particle substrate (fines), 2 = sand, 3 = gravel, 4 = cobble, 5 = boulder, and 6 = bedrock. We used ArcGIS (Version 10.2.1; ESRI, Redlands, CA) to view points and manipulate data. We created a polygon shapefile using ArcCatalog and edited it in ArcMap to outline bankful width of the target section of the stream (Fig. 1) overlaid on aerial imagery of the area (http:// wvgis.wvu.edu/data/data.php). The outline shapefile masked any raster outputs that fell outside the bankful width of the stream within the study area. We employed an inverse distance-weighted (IDW) technique found within Spatial Analyst Tools to interpolate sediment samples from each sampling event (i.e., 2007a, 2007b, 2008, 2009) and create a numerical substrate raster. We generated a reclassified raster for each of the 4 sampling events using 6 substrate designations (Reclassify, Spatial Analyst Tools; ESRI). We reclassified the substrates as fines = 1 (values ranging from 0 to 1.50), sand = 2 (1.51–2.0), gravel = 3 (2.51–3.50), and so forth. From the attribute table of each reclassified raster, we determined the percent composition of each substrate class for the 8-km section of the river . We used the Minus tool from the Math toolset (Spatial Analyst Tools) to quantify change in substrate composition over time. By comparing 2007a data with data from each subsequent sampling event, we were able to calculate and visually portray changes in substrate composition along the Little Coal River over time. In this case, negative values indicated a change from a lower numerical value (e.g., sand) to a higher numerical value (e.g., cobble) and vice versa . Results Sediment-size distributions in each of the 3 post-construction samples were significantly different from the pre-construction sample. The Kolmogorov- Smirnov statistics for 2007b (D = 0.103), 2008 (D = 0.314), and 2009 (D = 0.245) all exceeded the critical statistic (D800, 0.05 = 0.048), indicating that these samples were significantly different from the expected (i.e., original) distribution (2007a). The greatest proportional difference in sediment-size distribution between the pre- and post-construction sampling events was always in the sand substrate class, with the proportional contribution of sand reduced by 10% from 2007a to 2007b, 31% from 2007a to 2008, and 24% from 2007a to 2009 (Table 1). There was also a continual decline in the proportion of fine sediment over the 4 sampling events (Table 1). Our GIS analysis showed that approximately 80% of the riverbed substratum was composed of fine-particle or sand substrate classes prior to the Rosgen-type structure additions. By the final sampling event, these 2 substrate classes combined had been reduced by nearly 25% (Table 1), suggesting that restoration efforts reduced the overall percent composition of fine-particle- and sand-substrate classes and increased overall habitat heterogeneity (Fig. 2). Southeastern Naturalist 617 S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 We compared maps from 2007b, 2008, and 2009 to data collected in 2007a to show the immediate change in substrate composition as well as the total change within a 1- and 2-year period (Fig. 3). We used the resulting maps to visually assess changes in substrate. Large sections of the Little Coal River were transformed Table 1. Percent composition of each substrate class within the 8-km study area of the Little Coal River. 2007a indicates substrate sample prior to addition of Rosgen-type structures, and 2007b indicates substrate sample after this addition. A Komolgorov-Smirnov test was used to determine if the substrate distribution for each sampling event varied from the expected distribution (i.e., 2007a) with α = 0.05. * denotes a significant change in distribution when compared to 2007a. The overall change (Δ) in substrate composition between 2007a and 2009 is shown in the final column. Substrate Size (mm) 2007a 2007b* 2008* 2009* Δ Fines less than 0.06 8.49 4.45 1.05 0.83 -7.66 Sand 0.06–2 68.88 62.67 44.93 52.08 -16.80 Gravel 2–64 11.77 18.51 35.22 31.23 19.45 Cobble 65–256 7.84 11.36 16.69 14.62 6.77 Boulder 257–500 3.02 3.01 2.11 1.25 -1.77 Figure 2. Rasters showing overall substrate composition of an 8-km section of the Little Coal River were created using an inverse distance-weighted (IDW) technique and reclassified to include common substrate classes. (a) 2007 before construction of Rosgen-type structures, (b) 2007 after construction of Rosgen-type structur es, (c) 2008, and (d) 2009. Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 618 from smaller-size class sediment (e.g., sand) to larger-size class sediment (e.g., cobble), presumably resulting from increased downstream transport of fine-particle substrate and sand, indicating the efficacy of the structures designed and installed Figure 3. Rasters showing change over time in substrate composition of an 8-km section of the Little Coal River were created using the Minus tool. Comparisons were made between (a) pre-construction and post-construction in 2007, (b) pre-construction and 2008, and (c) pre-construction and 2009. In each panel, gray indicates no change in substrate composition, green indicates a change from a smaller substrate class to a larger substrate class, and red indicates a change from a lar ger substrate class to a smaller substrate class. Southeastern Naturalist 619 S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 for this purpose. Sediment-size class decreased in some areas. This change further highlighted that increased substrate heterogeneity occurred in some sections of the river that previously supported a rather homogenous substrate. Discussion Stream-restoration efforts often have the goal of emulating pre-disturbance conditions of stream substrata while also mitigating the degradation of streams and rivers. Evaluating such efforts requires monitoring over space and time (Mueller et al. 2014). Such assessments also require collaborations among diverse stakeholders by sharing images, maps, and analyses in a non-technical, inclusive manner (Soomai et al. 2013). Stream-restoration projects are often not monitored, or when they are, they are monitored poorly (Bernhardt et al. 2005). In many cases, researchers assess the implementation of the restoration effort, not the ecosystem response. This approach has led to a focus on the restoration process that comes at the expense of ecological concepts or ecosystem services (Lake et al. 2007). Ward et al. (2001) recognized the need for particular configurations of physical habitat and appro priate flow regimes for successful restoration at an ecosystem level. In some cases, restoration efforts have qualitatively improved physical-habitat condition (e.g., Jähnig and Lorenz 2008, Moerke et al. 2004), but a quantitative approach for determining substrate composition as well as the spatial arrangement of various habitat patches facilitates the successful monitoring of stream restoration projects. From an ecosystem perspective, increasing substrate diversity in restored sections may result in higher beta-diversity when compared to un-restored sections (Jähnig and Lorenz 2008). Simply increasing the diversity should not be the sole aim of restoration; rather, the goal should include restoration of stream sections to a physical structure capable of supporting populations and communities of organisms that existed prior to disturbance or the best regional expectations for such. Achieving this outcome requires a thorough understanding of the unique habitat (Southwood 1977) and spatiotemporal (Schneider 1994) requirements of the desired biota; an especially challenging goal when dealing with organisms requiring very specific substrate-condition requirements such as lithophilic fishes (e.g., Mueller et al. 2014) and unionid mussels (e.g., Niraula et al. 2015, Watters et al. 2009). The mapping technique we present in this paper offers a simple visual portrayal of existing substrate conditions, graphically imposing desired conditions, and monitoring post-restoration success. In the present study, results suggested that the installation of natural material instream structures are an effective method for promoting downstream movement of accumulated fine-particle substrate and sand. This finding was evidenced by a nearly 25% decrease in fine-particle or sand substrate, which in most cases occurred directly downstream of the installed structures. Additionally, substrate size decreased at some locations in the study area that were rather homogeneous prior to the installation of structures (Fig. 3). This observed substrate change, apparently Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 620 resulting from the installation of the structures, further contributed to the stated goal of increasing the heterogeneity of the substrate in the st udy area. Several questions remain that warrant further investigation. First, to what extent has the increase in substrate heterogeneity in the study area resulted in an increase in biotic heterogeneity? Second, has the instream-structure installation, and subsequent downstream movement of accumulated fine-particles and sand resolved a problem, or simply translocated the problem to a new location? Lastly, how have the structures and consequent increase in sediment heterogeneity that they apparently encouraged persisted through time? Acknowledgments Funding sources for this project included a grant to Marshall University from the West Virginia Department of Environmental Protection and the Coal River Group, a Wieman Wendel Benedict Research Award from the University of Cincinnati to S.E. Collins, and collaborative development support from the USEPA. We thank Jo Garofalo, Alex Hall, Roger Wolfe, Dennis Stottlemyer, Randy Huffman, and anonymous reviewers for comments on an earlier manuscript draft. The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the US Environmental Protection Agency. Literature Cited Benke, A.C., and C.E. Cushing. 2005. Background and approach. Pp. 1–16, In A.C. Benke. and C.E. Cushing (Eds.). Rivers of North America. Elsevier Academic Press, San Diego, CA. 553 pp. Bernhardt, E., M.A. Palmer, J.D. Allan, G. Alexander, K. Barnas, S. Brooks, J. Carr, S. Clayton, C. Dahm, J. Follstad-Shah, D. Galat, S. Gloss, P. Goodwin, D. Hart, B. Hassett, R. Jenkinson, S. Katz, G.M. Kondolf, P.S. Lake, R. Lave, J.L. Meyer, T.K. O’Donnell, L. Pagano, B. Powell, and E. Sudduth. 2005. Synthesizing US river restoration efforts. Science 308:636–637. Bevenger, G.S., and R.M. King. 1995. A pebble-count procedure for assessing watershed cumulative effects. Research Paper RM-RP-319. US Department of Agriculture, Forest Service, Fort Collins, CO. Collins, S.E., and J.E. Flotemersch. 2014. Evaluating the sounding-rod method for sampling coarse river-bed materials in non-wadeable streams and rivers. River Research and Applications 30:1065–1069. Henley, W.F., M.A. Patterson, R.J. Neves, and A.D. Lemly. 2000. Effects of sedimentation and turbidity on lotic food webs: A concise review for natural-resource managers. Reviews in Fisheries Science 8:125–139. Jähnig, S.C., and A.W. Lorenz. 2008. Substrate-specific macroinvertebrate-diversity patterns following stream restoration. Aquatic Sciences 70:292–303. Lake, P.S., N. Bond, and P. Reich. 2007. Linking ecological theory with stream restoration. Freshwater Biology 52:597–615. Leopold, L.B., M.G. Wolman, and J.P. Miller. 1964. Fluvial Processes in Geomorphology. Dover Publications, Mineola, NY. 544 pp. Messinger, T., and C.A. Hughes. 2000. Environmental setting and its relation to water quality in the Kanawha River Basin. Report 00-4020. Water-Resources Investigations, US Geological Survey, Charleston, WV. 66 pp. Southeastern Naturalist 621 S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 Moerke, A.H., K.J. Gerrard, J.A. Latimore, R.A. Hellenthal, and G.A. Lamberti. 2004. Restoration of an Indiana, USA, stream: Bridging the gap between basic and applied lotic ecology. Journal of the North American Benthological Society 23:647–660. Mueller, M., J. Pander, and J. Geist. 2014. The ecological value of stream-restoration measures: An evaluation on ecosystem and target-species scales. Ecological Engineering 62:129–139. Niraula, B.B, J.M. Hyde, J.M. Miller, P.D. Johnson, and P.M. Stewart. 2015. Microhabitat associations among three federally threatened and a common freshwater mussel species. American Malacological Bulletin 33:1–9. Poff, N.L, J.D. Olden, D.M. Merritt, and D.M. Pepin. 2007. Homogenization of regional river dynamics by dams and global biodiversity implications. Proceedings of the National Academy of Sciences 104:5732–5737. Rosgen, D.L. 2001. The cross-vane, w-weir, and j-hook-vane structures: Their description, design, and application for stream stabilization and river restoration. American Society of Civil Engineers, Restoration Proceedings, Reno, NV. doi: http://dx.doi. org/10.1061/40581(2001)72). Rosgen, D.L., and H.L. Silvey 1996. Applied River Morphology. Wildland Hydrology Books, Pagosa Springs, CO. 256 pp. Schneider, D.C. 1994. Quantitative Ecology: Spatial and Temporal Scaling. Academic Press, San Diego, CA. 395 pp. Soomai, S.S., B.H. MacDonald, and P.G. Wells. 2013. Communicating environmental information to the stakeholders in coastal and marine policy-making: Case studies from Nova Scotia and the Gulf of Maine/Bay of Fundy region. Marine P olicy 40:176–186. Southwood, T.R.E. 1977. Habitat, the templet for ecological strategies? Journal of Animal Ecology 46:337–365. Terrell, J.H. 2011. Evaluating substrate metrics for monitoring sediment impairment of East Tennessee streams. M.Sc. Thesis. University of Tennessee, Knoxville, TN. 89 pp. United Nations Educational, Scientific, and Cultural Organization (UNESCO). 1984. International Hydrological Challenges. Reference NS/188, Annex II, Paris, France. US Environmental Protection Agency (USEPA). 2013. National Rivers and Streams Assessment 2008–2009: A collaborative survey. EPA/841/D-13/001. US Environmental Protection Agency Office of Wetlands, Oceans and Watersheds and Office of Research and Development, Washington, DC. Ward, J.V., K. Tockner, U. Uehlinger, and F. Malard. 2001. Understanding natural patterns and processes in river corridors as the basis for effective restoration. Regulated Rivers: Research and Management 17:311–323. Waters, T.F. 1995. Sediment in streams: Sources, biological effects, and control. American Fisheries Society Monograph 7. Bethesda, MD. Watters, G.T., M.A. Hoggarth, and D.H. Stansbery. 2009. The Freshwater Mussels of Ohio. Ohio State University Press, Columbus, OH. 98 pp. Wentworth, C.K. 1922. A scale of grade and class terms for clastic sediments. The Journal of Geology 30:377–392. West Virginia Department of Environmental Protection (WVDEP). 2000. Water quality asessment: 305(b) report. Division of Waste and Water Management, Charleston, WV. WVDEP. 2002. Water quality assessment: Section 303(d) List. Division of Waste and Water Management, Charleston, WV. WVDEP. 2004. Water quality assessment: Section 303(d) List. Division of Waste and Water Management, Charleston, WV. Southeastern Naturalist S.E. Collins, J.E. Flotemersch, C.D. Swecker, and T.G. Jones 2015 Vol. 14, No. 4 622 WVDEP. 2006. Water quality assessment: Section 303(d) List. Division of Waste and Water Management, Charleston, WV. WVDEP. 2008. West Virginia integrated water quality monitoring and assessment report. Division of Waste and Water Management, Charleston, WV. WVDEP. 2010. West Virginia integrated water quality monitoring and assessment report. Division of Waste and Water Management, Charleston, WV. WVDEP. 2012. West Virginia integrated water quality monitoring and assessment report. Division of Waste and Water Management, Charleston, WV. Williams, D.D. 1980. Some relationships between stream benthos and substrate heterogeneity. Limnology and Oceanography 25:166–172.