Long-term response of salmonid populations to habitat restoration in a boreal forest stream

Ecological Engineering 91 (2016) 148–157

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Long-term response of salmonid populations to habitat restoration in a boreal forest stream Michael van Zyll de Jong a,∗ , Ian G. Cowx b a b

School of the Environment, Memorial University Grenfell Campus, Corner Brook, NL A2H 5G4, Canada Hull International Fisheries Institute, University of Hull, HU6 7RX, UK

a r t i c l e

i n f o

Article history: Received 10 October 2015 Received in revised form 20 January 2016 Accepted 27 February 2016 Keywords: Atlantic salmon Brook trout Habitat Restoration Boreal

a b s t r a c t Assessing the sustainability of restoration measures for salmonid populations and their habitat is limited due to a lack of long-term evaluations. In this paper we report on a study to assess the effect of boulder clusters, V-dams and half-log covers on stream habitat and population abundance of Atlantic salmon and brook trout two decades after installation. Structures were installed in Joe Farrell’s Brook, Newfoundland Canada in 1993 and fish population and habitat parameters were initially measured annually from 1993 to 1995. All stream sites were re-sampled in 2014. Boulder clusters or V-dams remained intact, stable and functional. By contrast, only 60% of the half-logs were in place and those remaining were in relatively poor shape with limited functionality. Boulder clusters increased the percentage area of pool habitat (p = 0.05) and the availability of instream cover (p = 0.04). V-dams did not significantly alter any habitat component after 20 years. Half-log covers were effective in increasing instream cover (p = 0.02) and substrate coarseness (p = 0.01). Density and biomass of Atlantic salmon increased rapidly after structures were installed and remained significantly higher 20 years after the pre-treatment period (1993) than the sub-basin control. By contrast brook trout did not show significant increase in density or biomass in either V-dams or half-log covers but did showed significant increases in both boulder cluster (density p = 0.002; biomass p = 0.01) 20 years later. This study suggests that instream structures placed in small boreal streams can function for more than two decades when properly installed and can maintain higher levels of salmonid abundance when habitat is limiting. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The manipulation of instream habitat has been used as an enhancement strategy to increase the abundance of stream salmonids for three centuries (Van Cleef, 1885; Tarzwell, 1935; Hunter, 1991; White et al., 2011). Restoration techniques include a range of instream structures (Cowx and Welcomme 1998; Roni et al., 2005a,b; Roni and Beechie 2013) to, for example, increase pool riffle habitat, restore natural meander sequences and increase habitat complexity. The common assumption is that installation of restoration structures will result in an increase in favorable physical habitat, which will result in an increase in population size of the targeted species (Hunt, 1994; Fausch et al., 1995; CCowx and van Zyll de Jong, 2004; Whiteway et al., 2010). Despite

∗ Corresponding author. Environmental Policy Institute, Grenfell Campus, Memorial University of Newfoundland, 20 University Drive, Corner Brook, NL A2H 6P9, Canada. Fax: +1 709632702. E-mail address: [email protected] (M. van Zyll de Jong). http://dx.doi.org/10.1016/j.ecoleng.2016.02.029 0925-8574/© 2016 Elsevier B.V. All rights reserved.

the popularity and ubiquitous use of these techniques, there are few post-implementation monitoring efforts or evaluation studies designed to test this assumption (Downs and Kondolf, 2002; Bernhardt et al., 2005; Bernhardt et al., 2007; Cowx et al., 2013), especially long term sustainability of such measures. Where evaluated (see reviews of Roni et al., 2014; Cowx et al., 2013), the levels of success are highly variable. Most of these studies have focused on the effects of placement of instream structures or logs(e.g., Blakely et al., 2006; Crispin et al., 1993; Floyd et al., 2009; Schmetterling and Pierce, 1999) on salmonid fishes (particularly Coho salmon, steelhead/rainbow trout, or other trout species) and found a positive response (increased abundance) for juvenile salmonids. However, other projects have shown unexpected results such as no response of targeted species but increase of other species (Cowx and van Zyll de Jong, 2004) or poor performance (Thompson, 2006; Stewart et al., 2009). Few studies have demonstrated longterm effects of different restoration approaches on fish populations and their habitats (Thompson and Stull, 2002; Champoux et al., 2003; Whiteway et al., 2010).

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

If the science of stream restoration is to evolve, longer term assessments of projects are required. These assessments should be focused on examining structural stability and effectiveness of instream structures and their ability to meet pre-defined objectives and end points, such as increasing fish abundance and availability of suitable habitat. In this paper we revisit and report on the long term impact of the three types of instream structures installed on Joe Farrell’s Brook, Newfoundland on the physical habitat and population abundance of Atlantic salmon (Salmo salar L.) and brook trout (Salvelinus fontinalis Mitchell) two decades after installation. The original objective of the study (van Zyll de Jong et al., 1997) was to assess the short-term effects of three types of habitat improvement structures (boulder clusters, V-dams and half-log covers) installed in a river that had been adversely affected by forest harvesting activities. Fish populations and key habitat attributes were monitored prior to and, in two subsequent years after the structures were placed at selected sites in channelized reaches. Boulder clusters proved to be the most effective structure, increasing densities of juvenile Atlantic salmon (Salmo salar L.) after placement of instream devices. V-dams proved to be effective in increasing both the density of brook trout (Salvelinus fontinalis Mitchel) and Atlantic salmon through the creation of more diverse pool habitat. Half-log covers increased the number of young of the year salmon through an increase in instream cover. The objectives the current study were; (1) to assess the structural durability of the measures implemented; (2) determine if the structures produce the desired physical habitat alterations in the long term; and (3) determine if the changes in salmonid biomass and density measured in the first two post-treatment periods are persistent over two decades.

2. Site description The Salmon River flows in a northeast direction from the upper section of the Great Northern Peninsula into Ariege Bay, an inlet of Hare Bay (51◦ 107 1077 N; 56◦ 017 2177 W). The river is 47.1 km long, drains an area of 730 km2 and has an average gradient of 5.6 m km−1 . The watershed contains over 100 pools and consists of two major tributaries: Salmon River and Southwest Brook. Southwest Brook enters the Salmon River 254 m above the high tide. The main branch is approximately 18.4 km in length, with a maximum elevation of 275 m and average gradient of 5.2 m km−1 . Treatment areas were confined to a small section of Southwest Brook locally referred to as Joe Farrell’s Brook. Joe Farrell’s Brook has an axial length of 4.2 km, meander length of 6.5 km and an average gradient of 19 m km−1 . The watershed lies in predominantly Ordovician sedimentary rock with some acidic intrusive rocks and gneisses. The landscape is gently rolling with Rubus-Balsam fir and TaxusBalsam fir the most common forest types (Meades and Moores, 1989). The Salmon River watershed was extensively logged from 1946 to 1971 and during this time approximately 450 km2 of forest was harvested (62% of the watershed). Forest harvesting consisted of clear cutting and no buffer zones were left along stream banks. Riverine transport of pulpwood was facilitated by channelization, removal of large instream structures, and the construction of large dams on lakes to hold water to support downstream floatation of logs. These activities reduced the amount and diversity of stream habitat, specifically the loss of vital stream bank communities and the normal pool-riffle sequence. The main objectives of the original project were to improve the salmonid stocks, especially brook trout, through stream habitat improvement measures. A pre-improvement study provided baseline information for an analysis of the factors most likely to be limiting the development of the salmonid fish communities. Sites were identified from this analysis for remedial treatments, includ-

149

ing placement of boulder clusters (n = 4), V-dams (n = 4) and half log covers (n = 2). Boulder clusters (according to size classification outlined in Wentworth scale reference) were intended to increase habitat diversity, thereby increasing refuge and rearing habitat for all age classes of juvenile salmonids. V-dams were designed to create a deep pool environment with roughened surface water, which was expected to provide high quality cover area for larger brook trout, particularly during low flow periods. Half log covers were intended to provide hiding–resting–security cover for yearling and older brook trout. A sub-basin control site was chosen to compare results obtained from the improved sites with an untreated, historically logged reach. Biological and physical habitat variables were sampled at 10 sampling sites on Joe Farrell’s Brook and one subbasin control site on Eastern Pond Brook from 1993 (pre-treatment year) to 1994–1995 (post-treatment years). 3. Methods 3.1. Durability of instream structures In July and August 2014, the 10 treatment sites on Joe Farrell’s Brook were resampled to determine whether the effects of boulder clusters (n = 4), V-dams (n = 4) and half-log covers (n = 2) on habitat and salmonid populations had persisted 21 years after installation. The classification system of Roper et al. (1998) used by White et al. (2011) was adopted to assess the longevity and stability of current structures. Structures were described as intact, moved or altered by natural processes or absent from the site. Secondly the integrity and functionality of the structures was described. For V-dams it was determined whether the structures continued to form pools, or whether water flowed underneath or around V-dams and hence failed to form either a dammed pool above or a plunge pool below. Boulder clusters were assessed with respect to the degree that the clusters were intact and the criteria used to assess half-logs were if they were still in place and had structural integrity providing shade. 3.2. Physical habitat measurements Stream habitats measurements were collected at 10 crosssectional transects, spaced every 4 m apart. Wetted width (measured to 0.1 m) was measured for each transect and stream depth (measured to 0.1 m) was measured at 30-cm intervals along each transect. Bottom substrate was visually examined according to a modified Wentworth scale (Gibson, 1993) every 30 cm. Substrate types were subsequently combined into a single coarseness rating by determining the relative proportion of each substrate type in the site, and multiplying by each component index, then summing to give an overall rating for each site (Gibson, 1993). Water velocity (measured to 0.01 m s−1 ) was measured at 30-cm intervals at 60% of maximum depth using an impeller velocity meter. Cover was visually assessed along each transect as instream, overhanging and canopy cover types and recorded as a percentage for each station (Gibson, 1993). 3.3. Electric fishing Fish were sampled from each 40-m section of stream enclosed with a 6-mm mesh size barrier net, with a Smith-Root LR-24 backpack electrofisher. Fish population abundance and biomass and associated variance were estimated using the maximum likelihood model in the MicroFish software package (Van Deventer and Platts 1985). Each run consisted of one pass upstream moving zigzagging laterally in a bank-to-bank fashion. The maximum number of passes was based on the rate of decline in successive fishing with a minimum of three passes. All fish were analyzed between each run. Fish captured were anaesthetized with CO2 by dissolving

150

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

Table 1 Two-way ANOVA examining the main effect of restoration procedures (i.e., boulder clusters, V-dams and half-log covers) on stream habitat parameters between control and treatments in post-treatment years (1994–1995, 2014) 20 years after installation (˛ = p ≤ 0.05). Parameter

Instream cover Pool % Coarseness

Boulder clusters

V-dams

Half log cover

F

p

F

p

F

p

3.630 3.650 0.229

0.045 0.050 0.874

0.340 0.493 1.199

0.797 0.694 0.352

13.12 0.737 16.98

0.015 0.582 0.010

an Alka–Seltzer tablet in a few liters of water. After measurement they were placed into an instream flow-through holding box before being release back into the station after the completion of the sampling. Each salmonid was identified to species, fork length measured to the nearest mm, and weighed to the nearest 0.1 g. For young of the year (YOY), fish were grouped by species, counted, measured for length (mm), and pooled weights were obtained for all YOY. Statistical Analyses: For each restoration treatment a two-way ANOVA was used to test if stream habitat features (i.e., pool %, coarseness, and instream cover) and fish population parameters (i.e., density and biomass) changed significantly between the pretreatment periods (1993) and post-treatment periods (1994–1995 and 2014). Bonferroni’s method was used to further test if there were significant differences in fish population parameters between treatment and control for all pre- and post-treatment periods. Before analysis was conducted, all explanatory variables were logtransformed to correct for heterogeneous variance. 4. Results 4.1. Durability of structures Boulder clusters were installed in the study stream in 1994, and 95% of cluster structures were in place and functioning in 2014. A small percentage (5%) of clusters had moved slightly during heavy flooding and ice events, and thus their integrity compromised by individual boulders being moved and distributed more than one meter apart from each another. V-dams were also installed in the study stream in 1994 and in 2014 all structures were in place (100%) and functioning to create a deep pool environment with roughened surface water, which was expected to provide high quality cover area for larger brook trout, particularly during low flow periods. The V-dams showed no signs of imminent failure and were completely intact. This condition persisted despite no maintenance during the intervening 20 years. Half log covers were installed in the study stream in 1994 and by 2014 only 60% of the original covers were in place. The log structures were exposed to low flows and high flood periods during the last two decades. Many of the cover structures were transported downstream by high flow and water velocity events, most likely uprooted by ice and high flow events. The furthest displacement was 4.3 km downstream. Half logs were intended to provide hiding-resting-refuge cover for yearling and older brook trout (Fig. 1). 4.2. Effects of structures on habitat A significant increase in percentage of pool habitat was found associated with the boulder cluster treatment in 2014 compared with the pre-treatment period. In addition, boulder clusters had no effect on coarseness rating but increased instream cover compared with the pretreatment period (Table 1; Fig. 2). No change in percentage of pool habitat was found with V-dam treatments in 2014 compared with the pre-treatment period. In addition, V-dams had

Table 2 Two-way ANOVA examining the main effect of restoration procedures (i.e., boulder clusters, V-dams and half-log covers) on Atlantic salmon and brook trout density and biomass between pre (1993) and post-treatment years (1994–1995) 20 years after installation including simple main effects of observed means (˛ = p ≤ 0.05). Boulder Clusters

Salmon density Salmon biomass Trout density Trout biomass – V-dams Salmon density Salmon biomass Trout density Trout biomass – Half Log Cover Salmon density Salmon biomass Trout density Trout biomass

Main effects

Simple main effects

F

p

1993

1994

1995

6.78 6.86 8.29 4.49

0.006 0.006 0.003 0.025

0.064 0.034 0.368 0.050

0.003 0.001 0.002 0.010

0.002 0.001 0.002 0.010

2.53 4.26 2.17 1.43

0.106 0.029 0.144 0.281

0.050 0.005 0.026 0.388

0.33 0.42 0.22 0.12

0.776 0.280 0.306 0.400

7.67 6.50 2.17 1.03

0.039 0.050 0.144 0.468

0.698 0.570 0.945 0.945

0.060 0.175 0.380 0.941

0.024 0.016 0.477 0.238

no effect on coarseness rating or instream cover compared with the pre-treatment period (Table 1; Fig. 2). There was no increase in percentage of pool habitat found with half-log treatments in 2014 compared with pre-treatment period. Half-log covers caused a significant increase in coarseness rating and instream cover compared with the pre-treatment period (Table 1; Fig. 2). 4.3. Effects of structures on fish density and biomass 4.3.1. Atlantic salmon Both density and biomass of Atlantic salmon increased at boulder clusters sites compared with the sub-basin control (Table 2). Density was significantly higher in 2014 than in the pre-treatment period 1993 and the post-treatment period 1995 with no significant difference found when compare to 1994 post treatment (Fig. 3). Biomass was significantly higher in 2014 than the pre-treatment year 1993 and both post-treatment years 1994 and 1995 (Fig. 4). V-dam treatments resulted in increases in biomass but not density when compared with the sub-basin control (Table 2). Densities were higher in 2014 than the pre-treatment year but did not differ from both post-treatment periods 1994 and 1995 (Fig. 3). A similar responses was found for biomass 2014 when compared with pre-treatment year 1993 and post-treatment years 1994 and 1995 (Fig. 4). Both density and biomass increased in the half-log treatment areas when compared with the sub-basin control (Table 2). Salmon densities were higher in 2014 in comparison to the pre-treatment year but were found to be similar to both post-treatment periods 1994 and 1995 (Fig. 3). Biomass was greater in 2014 when compared to the pre-treatment year 1993 but not significant in both post-treatment years 1994 and 1995 (Fig. 4). 4.4. Brook trout Density and biomass of brook trout increased significantly in the boulder cluster treatments when compared with the sub-basin control (Table 2). Trout densities were greater in 2014 when compared to the pre-treatment period 1993 and post-treatment year 1995 but not significantly different from the first post-treatment period 1994 (Fig. 3). Biomass was higher in 2014 than the pretreatment year 1993 and 2 post-treatment years 1994 and 1995 (Fig. 4). Density and biomass did not increase significantly in the V-dam treatments when compared with the sub-basin control (Table 2).

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

151

Fig. 1. Joe Farrell’s Brook location showing treatment sites.

Trout density was significantly greater in 2014 than in the pretreatment year 1993 but not significantly different from both posttreatment periods 1994 and 1995 (Fig. 3). Biomass in 2014 was not significantly different between pre-treatment and both posttreatment years (Fig. 4). Both density and biomass did not increase significantly in half-log treatments when compared with the sub-basin control (Table 2). No significant differences were found between densities of trout in 2014 when compared with the pre-treatment year or both post-treatment periods 1994 and 1995 (Fig. 3). Similarly biomass was not significantly different in 2014 compared with the pre-treatment year 1993 and both post-treatment years 1994 and 1995 (Fig. 4).

5. Discussion 5.1. Structural stability The long term structural condition of the three restoration measures varied across time. Boulder clusters and V-dam demonstrated long term strength and stability and continued to sustain improved habitat volume and higher salmonid abundance and biomass in treatment sites. Failure rates for various types of wood and boulder structures in North American streams are highly variable, ranging from 0% to 85% (Roni et al., 2002, 2005a,b). Longer-term studies suggest that most instream structures persist for less than 20 years (e.g., Ehlers 1956; House 1996), although little long-term

152

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

Fig. 2. Mean percentage difference (treatment-control) in targeted habitat parameters for specific restoration treatments across four time periods; one pre-treatment period, 1 = (1993), and three post-treatment periods 2 = (1994), 3 = (1995) and 4 = 2014. Vertical bars indicate ±1 SE, calculated from the raw data.

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

153

Fig. 3. Mean difference (treatment-control) in density (number/m2 ) for Atlantic salmon and brook trout for each restoration treatment across four time periods; one pre-treatment period, 1 = (1993), and three post-treatment periods 2 = (1994), 3 = (1995) and 4 = 2014. Vertical bars indicate ±1 SE, calculated from the raw data.

monitoring has occurred. Structural stability of instream wood and rock structures in the long term is influenced by a myriad of factors, including structure type, materials, and design, stream power, and the investigators’ definition of failure or success (Roni et al., 2005a,b). Structural failures that have been documented have been attributed to inappropriate structures or a lack of understanding of larger watershed processes (e.g., Kondolf et al., 1996; Thompson and Stull, 2002). In the case of both V-dams and boulder clusters used in the current study appropriate treatment, materials and design were considered of the correct magnitude to sustain stream power and appropriate for site specific conditions. By contrast, half-log covers remained structurally intact in the first post-treatment period (1994–1995) providing an increase in

instream cover and coarseness resulting a corresponding increase in Atlantic salmon abundance, but had no detectable effect on the abundance of the target species—brook trout. In the second posttreatment period (2014), 40% of the half-log covers were displaced, but despite this loss continued benefit for Atlantic salmon was sustained. Furthermore, the expected benefit for brook trout failed to occur in the long term. These results are consistent with shortand medium-term studies on the structural performance of half log covers, e.g. Hrodey & Sutton (2008) in nine Indiana streams and Roper et al. (1998) in streams in the Pacific Northwest, where the majority of structures remained function up to 5 years under variable flood magnitudes. No long-term evaluations of half-log covers were found. These studies suggest that annual or bi-annual

154

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

Fig. 4. Mean difference (treatment-control) in biomass (g/m2 ) for Atlantic salmon and brook trout for each restoration treatment across four time periods; one pre-treatment period, 1 = (1993), and three post-treatment periods 2 = (1994), 3 = (1995) and 4 = 2014. Vertical bars indicate ±1 SE, calculated from the raw data.

maintenance is required to keep half logs functioning properly as fish habitat (Kondolf et al., 1996; Roni et al., 2002). Wood addition should be limited to physically stable streams because physical failure easily occurs in sand-slugged or erodible streams. Wood devices that create complex instream cover are recommended (Nagayama & Nakamura 2010). In a review of instream structures deployed in the northeast United States Thompson and Stull (2002) found widespread failure in long-term durability of habitat enhancement structures and highlighted the need to have a better integrated understanding of dynamic ecosystem processes at the watershed scale over a longer temporal scale to determine the potential long-term impacts and uncertainty around the success of instream structures (Frissell and Nawa, 1992; Kondolf, 1995). In the initial planning phases of this

project, emphasis was placed on direct effects of specific measures on key habitat variables of concern. Little consideration was given to the longer term effects of ecological processes (e.g. flooding events and winter ice conditions) on durability. This is probably the primary reason for the loss of 40% of half-log structures and suggests that selection could have been had been improved if the long-term effects of water level fluctuations were considered (Cowx and van Zyll de Jong, 2004; Cowx et al., 2013). 5.2. Habitat effects The ability of the measures to produce desired physical habitat features in the long term had variable results. The percentage of pool habitat and instream cover associated with boulder clus-

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

ter treatments increased over the last two decades. Similar effects of boulder groupings have been documented in other regions (e.g. Näslund, 1989; House, 1996). By contrast, V-dams did not significantly change the representation of any targeted habitat features. In other studies, V-dams and instream log structures have resulted in increases in pool frequency, pool depth, and woody debris and sediment retention (e.g., Ehlers 1956; Armantrout 1991; Crispin et al., 1993; Fissell and Nawa 1992; Cederholm et al., 1997; Whiteway et al., 2010). In the current study V-dams initially created a small plunge pool and over time it deepens and forms a suitable deep pool habitat directly in front of the structure. However, the downstream sections of Vdam sites narrowed over time eliminating some shallower side pool habitat, which was replaced by longer, narrow, more extensive riffle areas, thus favoring salmon over trout. The long term relative effect to increase pool habitat was likely due to longer term changes in channel morphology and river functioning. Half-log covers had the desired effect of increasing the overall coarseness and instream cover in the treated sections compared with the pre-treatment conditions over the two decades of the study. Direct addition of instream cover enhanced the reaches structural complexity. Before half-log additions, no difference existed between the total amount of cover in treatment sites and the sub-basin control site. The addition of half logs to each treatment site increased the amount of instream cover by an average 33%. Half-logs provided critical resting-hiding-shelter areas for juvenile salmonids. 5.3. Effect on fish abundance A true measure of the performance of habitat rehabilitation measures can be related to the ability to demonstrate improvement in habitat quality, which can then be linked to observed changes in fish populations (van Zyll de Jong et al., 1997), i.e. setting quantifiable targets for restoration outcomes (Cowx et al., 2013). van Zyll de Jong et al. (1997) suggested boulder clusters were successful in creating habitat diversity as evident from an increase in the density of Atlantic salmon in both post-treatment years. The initial increase in salmon abundance was attributed an increase in the number and diversity of foraging positions and shelter. Prior to the placement of the boulder clusters, the pre-treatment condition and sub-basin control were extremely uniform and channelized with little habitat heterogeneity. The greater habitat complexity created by the addition of boulders increased the habitat complexity and this has remained intact over time with increases in instream cover and small pool habitat. These alterations have increased variability in depth, substrate, cover, and current velocity, which is probably regulating Atlantic salmon abundance (Gibson, 1993). In most Newfoundland streams, brook trout and salmon coexist, but with juvenile salmon generally dominating riffles and brook trout pools (Gibson, 1993). Salmon juveniles typically displace brook trout from riffles where salmon densities are high and there is competition for food (Gibson, 1973). The larger pectoral fins of salmon juveniles enable them to hold to the substrate in riffle areas. By contrast, brook trout have smaller pectoral fins and are more buoyant, which gives them a distinct advantage in pool habitats, displacing salmon juveniles through antagonistic interactions (Gibson, 1993). Several studies have found a similar increase in salmonid densities following the placement of boulder clusters and attributed the change to an increase in habitat complexity and stabilization of smaller bed materials where juveniles overwinter (e.g. Ward and Slaney, 1979; House and Boehne, 1985; House et al., 1991; O’Grady and King, 1991; Kelly and Bracken, 1998; Gargan et al., 2002; Clarke and Scruton, 2002). Brook trout density and biomass also increased in the boulder cluster treatment areas and were greater two decades after implementation. This gradual increase in brook

155

trout abundance can be attributed to increases in the availability in habitat diversity. V-dams were installed to create pool habitat for brook trout. The structures were expected to increase instream habitat complexity and provide variability in flow, substrate, and depth. There was no significant change in brook trout biomass or densities immediately post-treatment years, but increases in brook trout density were realized two decades later. The lack of any short term positive effect on brook trout was attributable to a lack instream or overhanging cover in the vicinity of the pool area created at the V-dam sites. The positive response 20 years later may be related to more overhanging and instream stream cover in addition to an increase in the proportion of pool habitat. An increase in cover was not observed at the sub-basin control site therefore it was not possible to test the two factors independently. A similar increase in the density of Atlantic salmon was observed. These changes were attributed to increases in gravel, narrowing of the stream channel, and increase in relative amount of riffle habitat (House et al., 1991; Binns, 1994). Half-log covers placed in Joe Farrell’s Brook were intended to provide hiding-resting-security cover for yearling and older brook trout. Brook trout density and biomass at the half-log cover sites two decades later were similar to the pre-treatment year and subbasin control. Hartzler (1983) also found no effect of half log cover structures on harvest, abundance, or biomass of brown trout in Pennsylvania. By contrast, half-log treatment produced significant increases in density and biomass of Atlantic salmon two decades later compared with the pre-treatment period and sub-basin control. The increase in Atlantic salmon is attributable to the half-logs providing better over-winter habitat and critical areas for restinghiding-shade in open exposed riffle areas. The result suggests that holistic watershed planning with the explicit incorporation of longterm ecosystem dynamics as an important factor that needs to be considered when selecting sites and deploying this enhancement measure. We conclude that installing instream structures has resulted in substantial and long-term increases in salmonid abundance and biomass, although variation was found between the measures installed and the efficacy of the measures. Boulder structures provided the most sustainable long-term improvements, whereas half-log covers were less durable and probably need maintenance after a number of years or following extreme flow events, which disrupt the structures. Sustained increases in salmonid abundance in this study can be attributed to the stream channels that were chosen. The sites were relatively small and stable allowing the structures to continue functioning to create desired habitat components, and because habitat was initially limiting. Salmonid densities of both brook trout and salmon increased during the posttreatment periods. These increases in salmonid abundance cannot be solely attributed to an improvement in physical habitat. Other factors play a role in determining the productivity of the stream reaches treated. Density and biomass are controlled by physical and chemical habitat variables and are modified through inter and intra-specific competition (Gibson, 1973; Gibson, 1993). In this study, the conservation or restoration of natural habitat diversity appears to have minimized inter- and intra-specific competition and thus allowing a sustained increase in production. While this study suggests a variety of factors can affect the level of response, instream measures can lead to substantial changes and improvements in physical habitats and associated biological components. It also highlights the need for long-term assessment of the efficacy of restoration measures and the need for adaptive management to respond to failure or deterioration of the restoration action. Given the scale of this study and the many factors that were not accounted for, it may be appropriate to further such research in different catchments and using modeling approaches, possibly in combination with information theoretic analysis (Paulsen and Fisher, 2005) to

156

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157

test the relative support of multiple working hypotheses provided by observations or experiments to account for the range of habitat variables that may affect salmonid densities. Acknowledgements We would like to thank the Atlantic Salmon Conservation Foundation for funding this study. We would also like to thank the Canadian Department of Fisheries and Oceans and Canadian Forest Service for contributing field equipment and finally we would like to thank our community partner the White Bay Central Development Association for their support. IGC was supported under the EU REFORM project (REstoring rivers FOR effective catchment Management)- Grant Agreement 282656. References Armantrout, N.B., 1991. Restructuring streams for anadromous salmonids. Fisheries Bioengineering Symposium: American Fisheries Society Symposium 10, p. 136. Bernhardt, E.S., Palmer, M., Allan, J.D., Alexander, G., Barnas, K., Brooks, S., Carr, J., et al., 2005. Synthesizing U. S. river restoration efforts. Science 308 (5722), 636–637, http://dx.doi.org/10.1126/science.1109769 (Washington). Bernhardt, E.S., Sudduth, E.B., Palmer, M.A., Allan, J.D., Meyer, J.L., Alexander, G., Pagano, L., 2007. Restoring rivers one reach at a time: results from a survey of US river restoration practitioners. Restor. Ecol. 15 (3), 482–493, http://dx.doi. org/10.1111/j.1526-100X.2007.00244.x. Blakely, T.J., Harding, J.S., Mcintosh, A.R., Winterbourn, M.J., 2006. Barriers to the recovery of aquatic insect communities in urban streams. Freshwater Biol. 51 (9), 1634–1645. Binns, N.A., 1994. Long-term responses of trout and macrohabitats to habitat management in a Wyoming headwater stream. North Am. J. Fish. Manage. 14 (1), 87–98, http://dx.doi.org/10.1577/1548-8675(1994) 014<0087:LTROTA>2. 3.CO;2. Cederholm, C.J., Bilby, R.E., Bisson, P.A., Bumstead, T.W., Fransen, B.R., Scarlett, W.J., Ward, J.W., 1997. Response of juvenile coho salmon and steelhead to placement of large woody debris in a coastal Washington stream. North Am. J. Fish. Manage. 17 (4), 947–963. Champoux, O., Biron, P.M., Roy, A.G., 2003. The long-term effectiveness of fish habitat restoration practices: Lawrence Creek Wisconsin. Ann. Assoc. Am. Geogr. 93 (1), 42–54, http://dx.doi.org/10.1111/1467-8306.93104. Clarke, K.D., Scruton, D.A., 2002. Evaluating efforts to increase salmonid productive capacity through habitat enhancement in the low diversity/production systems of Newfoundland, Canada. In: Proceedings of the 13th International Salmonid Habitat Enhancement Workshop, Westport, Ireland, 16–19 September 2002 Irish Freshwater Ecology and Management Series, Central Fisheries Board, Dublin, Ireland, pp. 160–182. Cowx, I.G., Welcomme, R.L. (Eds.), 1998. 204 pp. Fishing News Books Blackwell Science, Oxford. Cowx, I.G., van Zyll de Jong, M., 2004. Rehabilitation of freshwater fisheries: tales of the unexpected? Fish. Manage. Ecol. 11 (3–4), 243–249, http://dx.doi.org/10. 1111/j.1365-2400.2004.00410.x. Cowx, I.G., Angelopoulos, N., Noble, R.A., Slawson, D., Buijse, T., Wolter, C., 2013. Measuring success of river restoration actions using end-points and benchmarking Deliverable D5.1 EU FP7 project REstoring rivers FOR effective catchment Management (last accessed 10.10.15.). http://www.reformrivers. eu/system/files/5.1%20Measuring%20river%20restoration%20success.pdf. Crispin, V., House, R., Roberts, D., 1993. Changes in instream habitat, large woody debris, and salmon habitat after the restructuring of a coastal Oregon stream. North Am. J. Fish. Manage. 13 (1), 96–102, http://dx.doi.org/10.1577/15488675(1993) 013<0096:CIIHLW>2.3.CO;2. Downs, P.W., Kondolf, G.M., 2002. Post-project appraisals in adaptive management of river channel restoration. Environ. Manage. 29 (4), 477–496, http://dx.doi. org/10.1007/s00267-001-0035-X. Ehlers, R., 1956. An evaluation of stream improvement devices constructed eighteen years ago. Calif. Fish Game 42 (3), 203–217. Fausch, K.D., Gowan, C., Richmond, A.D., Riley, S.C., 1995. The role of dispersal in trout population response to habitat formed by large woody debris in Colorado mountain streams. Bull. Francais de la Peche et de la Pisciculture 337/338 (339), 179–190 http://dx.doi.org/10.1051/kmae:1995020. Frissell, C.A., Nawa, R.K., 1992. Incidence and causes of physical failure of artificial habitat structures in streams of western Oregon and Washington. North Am. J. Fish. Manage. 12, 182–197, http://dx.doi.org/10.1577/15488675(1992)012<0182:IACOPF>2.3.CO;2. Floyd, T.A., MacInnis, C., Taylor, B.R., 2009. Effects of artificial woody structures on Atlantic salmon habitat and populations in a Nova Scotia stream. River Res. Appl. 25 (3), 272–282. Gargan, P., O’Grady, M., Delanty, K., Igoe, F., Byrne, C., 2002. The effectiveness of habitat enhancement on salmon and trout stocks in streams in the Corrib Catchment. In: O’Grady, M. (Ed.), Proceedings of the 13th international

salmonid habitat enhancement workshop. Central Fisheries Board: Dublin, Ireland, pp. 220–223. Gibson, R.J., 1973. Interactions of juvenile Atlantic salmon (Salmo salar L.) and brook trout (Salvelinus fontinalis Mitchill). Int Atl. Salmon Found. Spec. Publ. Ser. 4 (1), 181–202. Gibson, R.J., 1993. The Atlantic salmon in fresh water: spawning, rearing and production. Rev. Fish Biol. Fish. 3 (1), 39–73, http://dx.doi.org/10.1007/ BF00043297. Hartzler, J.R., 1983. The effects of half-log covers on angler harvest and standing crop of brown trout in McMichaels Creek, Pennsylvania. North Am. J. Fish. Manage. 3 (3), 228–238, http://dx.doi.org/10.1577/15488659(1983)3<228:TEOHCO>2.0. CO;2. House, R., 1996. An evaluation of stream restoration structures in a coastal Oregon stream, 1981-1993. North Am. J. Fish. Manage. 16 (2), 272–281, http://dx.doi. org/10.1577/1548-8675(1996) 016<0272:AEOSRS>2.3.CO;2. House, R.A., Boehne, P.L., 1985. Evaluation of instream enhancement structures for salmonid spawning and rearing in a coastal Oregon stream. North Am. J. Fish. Manage. 5(2B, 283–295, http://dx.doi.org/10.1577/15488659(1985) 5<283:EOIESF>2.0.CO;2. House, R., Crispin, V., Suther, J.M., 1991. Habitat and channel changes after rehabilitation of two coastal streams in Oregon. Fisheries Bioengineering Symposium: American Fisheries Society Symposium 10, p. 150. Hrodey, P.J., Sutton, T.M., 2008. Fish community responses to half-log additions in warmwater streams. North Am. J. Fish. Manage. 28 (1), 70–80. Hunt, R.L., 1994. Trout Stream Therapy. University of Wisconsin Press, Madison, Wisconsin, USA. Hunter, C.J., 1991. Better Trout Habitat. Island Press, Washington, D.C, pp. 320. Kelly, F.L., Bracken, J.J., 1998. Fisheries enhancement of the Rye Water, a lowland river in Ireland. Aquat. Conserv.: Marine Freshwater Ecosyst. 8 (1), 131–143, http://dx.doi.org/10.1002/(SICI) 1099-0755(199801/02) 8:1<131:AIDAQC258>3.0.CO;2-H. Kondolf, G.M., 1995. Five elements for effective evaluation of stream restoration. Restor. Ecol. 3, 133–136, http://dx.doi.org/10.1111/j.1526-100X.1995.tb00086. x. Kondolf, G.M., Vick, J.C., Ramirez, T.M., 1996. Salmon spawning habitat rehabilitation on the Merced River, California: an evaluation of project planning and performance. Trans. Am. Fish. Soc. 125 (6), 899–912. Meades, W., Moores, L., 1989. Forest Site Classification Manual: a Field Guide to the Damman Forest Types of Newfoundland. Canadian Forest Service, Newfoundland and Labrador Region, St John’s Newfoundland, 285pp. Nagayama, S., Nakamura, F., 2010. Fish habitat rehabilitation using wood in the world. Landsc. Ecol. Eng. 6 (2), 289–305. Näslund, I., 1989. Effects of habitat improvement on the brown trout Salmo trutta L., population of a northern Swedish stream. Aquacult. Res. 20 (4), 463–474, http://dx.doi.org/10.1111/j.1365-2109.1989.tb00373.x. O’Grady, M.F., King, J.J., 1991. The effectiveness of two physical in-stream works programmes in enhancing salmonid stocks in a drained Irish lowland river system. In Strategies for the rehabilitation of salmon rivers. In: Mills, D. (Ed.), Proceeding of a Joint Conference held at the Linnean Society. The Atlantic salmon Trust, The Institute of Fisheries Management and The Linnean Society of London: London,U.K, pp. 154–178. Paulsen, C.M., Fisher, T.R., 2005. Do habitat actions affect juvenile survival? An information-theoretic approach applied to endangered Snake River Chinook salmon. Trans. Am. Fish. Soc. 134 (1), 68–85. Roni, P., Beechie, T.J., Bilby, R.E., Leonetti, F.E., Pollock, M.M., Pess, G.R., 2002. A review of stream restoration techniques and a hierarchical strategy for prioritizing restoration in Pacific Northwest watersheds. North Am. J. Fish. Manage. 22 (1), 1–20. Roni, P., Hanson, K., Pess, G., Beechie, T., Pollock, M., Bartley, D., 2005a. Habitat rehabilitation for inland fisheries: global review of effectiveness and guidance for restoration of freshwater ecosystems. In: Fisheries Technical Paper 484. Food and Agriculture Organization of the United Nations, Rome, Italy, pp. pp. 116. Roni, P., Liermann, M.C., Jordan, C., Steel, E.A., 2005b. Steps for designing a monitoring and evaluation program for aquatic restoration. In: Monitoring stream and watershed restoration. American Fisheries Society, Bethesda, Maryland, pp. 13–34. Roni, P., Beechie, T., 2013. Stream and watershed restoration—A guide to restoring riverine processes and habitats. John Wiley & Sons Ltd. Roni, P., Beechie, T., Pess, G., Hanson, K., 2014. Wood placement in river restoration: fact, fiction, and future direction. Can. J. Fish. Aquatic Sci. 72 (3), 466–478. Roper, B.B., Konnoff, D., Heller, D., Wieman, K., 1998. Durability of Pacific Northwest instream structures following floods. North Am. J. Fish. Manage. 18 (3), 686–693, http://dx.doi.org/10.1577/15488675(1998) 018<0686:DOPNIS>2.0.CO;2. Schmetterling, D.A., Pierce, R.W., 1999. Success of instream habitat structures after a 50-Year flood in Gold Creek Montana. Restor. Ecol. 7 (4), 369–375.24, http:// dx.doi.org/10.1046/j.1526-100X.1999.72031.x. Stewart, G.B., Bayliss, H.R., Showler, D.A., Sutherland, W.J., Pullin, A.S., 2009. Effectiveness of engineered in-stream structure mitigation measures to increase salmonid abundance: a systematic review. Ecol. Appl. 19 (4), 931–941 http://dx.doi.org/10.1890/07-1311.1. Tarzwell, C.M., 1935. Progress in lake and stream improvement. Transactions of the Twenty-first American Games Conference 21, 119–134.

M. van Zyll de Jong, I.G. Cowx / Ecological Engineering 91 (2016) 148–157 Thompson, D.M., 2006. Did the pre-1980 use of in-stream structures improve streams? A reanalysis of historical data. Ecol. Appl. 16 (2), 784–796 http://dx. doi.org/10.1890/1051-0761(2006)016[0784:DTPUOI]2.0.CO;2. Thompson, D.M., Stull, G.N., 2002. The development and historic use of habitat structures in channel restoration in the United States : the grand experiment in fisheries management. Géographie physique et Quaternaire 56 (1), 45–60, http://dx.doi.org/10.7202/008604ar. Van Deventer, Platts, 1985. ‘Microcomputer software system for generafing population estimates from electrofishing data: user guide for Microfish 3.0 , USIDA Forest Service Intermountain Research Station General Technical Report INT-254. 29 pp. van Zyll de Jong, M.C., Cowx, I.G., Scruton, D.A., 1997. An evaluation of instream habitat restoration techniques on salmonid populations in a Newfoundland

157

stream. Regul. Rivers-Res. Manage. 13 (6), 603–614, 34.10.1002/(SICI) 1099-1646(199711/12) 13:6<603:AID-RRR487>3.0.CO;2-Y. Van Cleef, J.S., 1885. How to restore our trout streams. Trans. Am. Fish. Soc. 14, 50–55, http://dx.doi.org/10.1577/1548-8659(1885) 15[50:HTROTS]2.0.CO;2. White, S.L., Gowan, C., Fausch, K.D., Harris, J.G., Saunders, W.C., 2011. Response of trout populations in five Colorado streams two decades after habitat manipulation. Can. J. Fish. Aquatic Sci. 68 (12), 2057–2063 http://dx.doi.org/10. 1139/f2011-125. Whiteway, S.L., Biron, P.M., Zimmerman, A., Venter, O., Grant, J.W.A., 2010. Do in-stream restoration structures enhance salmonid abundance? A meta-analysis. Can. J. Fish. Aquatic Sci. 67 (5), 831–841, http://dx.doi.org/10. 1139/F10-021.