2.5 Instream habitat structures

2.5.1 Common instream habitat structures

The placement of physical structures into lotic environments to create pools, to alter channel morphology, and to provide cover and habitat for fish and other aquatic organisms has a long history (White, 1996, 2002). It was one of the first methods used to mitigate habitat degradation and to increase fish production in streams and rivers (Tarzwell, 1934, 1937, 1938) and also is one of the most common and widespread methods in use. Many different configurations of instream structures and methods have been used over the years to improve habitat (White and Brynildson, 1967; Vetrano, 1988; Hunt, 1993; Riley and Fausch, 1995; Cowx and Welcomme, 1998). These techniques involve placement of materials such as large woody debris (LWD), boulders, and other materials into the active stream channel or actual manipulation of the active channel itself in an effort to improve fish habitat. They can be categorized by purpose (e.g. create pools, trap gravel) and material and can include such structures as boulder or log weirs, dams and deflectors, cover structures, rootwads and brush bundles, gabions, logjams and placement of spawning gravel (Table 8; Figure 9). Because these activities seek to enhance habitat rather than restore a deficient process (e.g. riparian, hydrology) or return a stream to some predisturbance state, they are called instream habitat enhancement rather than restoration.

TABLE 8
Common types of structures placed in streams to improve fish habitat. Adding channel meanders is discussed in section on floodplain rehabilitation. Modified from Roni (2005).

Most instream enhancement techniques used today were developed in the 1930s for use in the relatively low-gradient streams of the midwestern United States (Hunt, 1993; White, 1996). They were designed to enhance trout habitat that had been simplified and degraded through agriculture, forestry, and other land use. The primary purpose of early instream structures was to narrow channels, provide cover for trout and other game fish, and reduce mortality due to predation (Tabor and Wurtsbaugh, 1991; Gregory and Levings 1996). Instream enhancement techniques developed in the Midwest were later applied to streams throughout North America and Europe with varying degrees of success (Ehlers, 1956; Armantrout, 1991; Iversen et al., 1993; O’Grady, 1995; Cowx and Welcomme, 1998; Gortz, 1998; Laasonen et al., 1998). In higher energy streams, such as those in mountainous areas of western North America, the major goal of placement of instream structures has been to create pools and trap gravel. In more recent efforts in low energy, low gradient (< 1 percent gradient) channelized streams such as those found in floodplains and agricultural land, the goals are generally to increase habitat heterogeneity (create a suite of both shallow and deep water habitats), and either narrow or widen the channel depending on the channel and its level of degradation.

2.5.2 Effectiveness of instream structures

Though instream habitat enhancement is popular, widespread, has a lengthy history, and more than 60 published evaluations exist, most techniques have not been completely evaluated.

FIGURE 9. Three common instream habitat enhancement techniques or structures: rock weir (top), log weir (middle), and log jam or multiple log structure (bottom).

The need for rigorous monitoring and evaluation has been noted for many decades (Tarzwell, 1937; Reeves et al., 1991; Roni, 2005). There is considerable debate about the appropriateness of instream enhancement, because these techniques historically occurred without an assessment of what factors caused the lack of habitat complexity, what processes in the watershed might be disrupted and need to be corrected, and what factors might be limiting the physical and biological production of a system (Frissell and Nawa, 1992; Roni et al., 2002; White, 2002). For example, the failure of instream wood and boulder structures to increase fish abundance in a stream reach may be due to a lack of riparian vegetation, caused in turn by land use activities such as agriculture or even urbanization. In such cases, if the underlying cause of the problem is not addressed, instream enhancement may represent only a short-term improvement in habitat and in heavily modified areas, it may be the only feasible option. Thus habitat enhancement projects are not without merit. Short-term habitat improvements may be needed to protect rare fishes and biota or to provide short-term benefits where some watershed process cannot be restored (e.g. woody debris in urban streams). They are particularly relevant when coupled with restoration of natural processes, such as riparian rehabilitation or sediment reduction (i.e. road treatments and slope stabilization), otherwise these habitat enhancement techniques run the risk of treating the symptoms rather than addressing the underlying problem causing the degradation or simplification of instream habitat. It is important to reiterate that instream habitat enhancement techniques seek to improve habitat rather than restore a disrupted process. Below we review their physical and biological effectiveness.

2.5.2.1 Physical habitat

Durability-The majority of monitoring and evaluation efforts of instream projects have focused on the effects of the project on channel morphology and instream habitat (Hunt, 1988; Reeves et al., 1991; Binns, 1999; Roni et al., 2002). Early evaluations reported on the durability of instream structures, whether they created pools, and their longevity (Ehlers, 1956; Gard, 1972; Armantrout, 1991; Frissell and Nawa, 1992). Reported failure rates for various types of wood and boulder structures in North American streams are highly variable, ranging from 0 to 100 percent (Roni et al., 2002; Table 9). A few studies such as Ehlers (1956) and Thompson (2002) have examined long-term durability (>18 years) of habitat enhancement structures. For example, Thompson (2002) found that weirs, deflectors, and cover structures constructed in the 1930s and 1950s in two Connecticut rivers were no longer functioning. He also reported that failed cover structures have actually lead to reduced streamside vegetation and overhead cover and that some bank protection and grade control structures have prevented the recovery of the channel and increased erosion. This and other studies on functionality of instream structures demonstrate the need to thoroughly understand channel dynamics and long-term impacts of stationary structures.

TABLE 9
Summary of key studies evaluating instream structure durability. Note that most studies did not break down success rate by structure type. NA = not available and a blank space indicates that that type of structure was not examined. Modified from Roni et al. (2002).

a Functioning defined as in-place and functioning as intended
b Functioning defined as improving habitat
c Functioning defined as functioning as intended
d Functioning defined as in-place or largely in-place but shifted
e Functioning defined as no movement or moved less than one bankful width

The physical success of many structures varies widely and is influenced by many factors including structure type, materials and design, stream power, and often by the investigators definition of "failure" or "success". As previously mentioned, early failures of structures often resulted from the use of inappropriate structures or lack of understanding of larger watershed processes (e.g. Kondolf et al., 1996). More recently, highly artificial techniques such as log weirs are being replaced with techniques that mimic natural wood accumulations or habitat. These newer methods have been demonstrated to be more durable in some instances and more effective at producing changes in habitat than traditional structures (Cederholm et al., 1997; Thom, 1997; Roni and Quinn, 2001a).

Changes in habitat-Despite highly variable results in durability, many studies have reported large (>50 percent) and significant increases in pool frequency, pool depth, woody debris, habitat heterogeneity, complexity, spawning gravel, and sediment retention following placement of instream structures (e.g. Crispin et al., 1993; Cederholm et al., 1997; Reeves et al., 1997; Binns, 1999; Roni and Quinn, 2001a; Bates et al., 1997; Gerhard and Reich, 2000; Brooks et al., 2004). Similarly projects in the midwestern United States have also demonstrated physical habitat changes including increased depth, cover, and narrower channels as a result of instream enhancement projects (Hunt, 1988; Kern, 1992; Avery, 2004). Other studies designed to aggrade highly incised stream channels have produced increases in depth, width, pool area, and bed elevation (reduced incision) following boulder weir placement (Newbury and Gaboury, 1988; Shields et al., 1993; Shields et al., 1995a).

Several European studies in low gradient channelized streams have demonstrated changes in physical habitat following placement of structures. Gerhard and Reich (2000) demonstrated increased habitat complexity, depth, organic matter retention, and macroinvertebrate species richness and abundance. Näslund (1989) reported increased depth and pool area in reaches of a Swedish stream treated with boulder weirs, while log and boulder deflectors decreased wetted area and boulder clusters had little effect on these parameters. Pretty et al. (2003) found increased depth and flow heterogeneity in rehabilitated stream reaches than control reaches in 13 English streams. Zika and Peter (2002) found increased pool volume and number following placement of LWD in a Swiss stream. Jungwirth et al. (1995) reported changes in physical habitat complexity (e.g. depth, velocity, substrate) following placement of instream structures in Austrian streams. These studies indicate that while a variety of factors can affect the level of response, many instream structures lead to substantial changes and improvements in physical habitats.

2.5.2.2 Biological effectiveness of instream structures

Biological evaluations of different enhancement techniques have been less frequently than physical evaluations and have produced inconsistent results depending on the technique, region, species and life stage examined, as well as the duration of monitoring. The majority of these evaluations have focused on trout or juvenile anadromous salmonids, which historically were the focus of most instream enhancement efforts in North America and Europe. Monitoring of instream enhancement has been more intensive in North America than most regions of the world. For some regions of the United States, there are compendiums summarizing the effectiveness of most instream rehabilitation projects within a particular region (e.g. Hunt, 1988; Binns, 1999; Avery, 2004). The results of these and other studies, however, have been variable and there is still much debate in the scientific community about the biological effectiveness of different instream enhancement methods. In an effort to summarize the extensive biological evaluations, the following subsections discuss juvenile and salmonid fishes (anadromous and resident), nonsalmonids, macroinvertebrates, and other biota.

Juvenile anadromous salmonids responses-In a synthesis of the effectiveness of such efforts for Pacific salmon (Oncorhynchus spp.), Roni et al. (2002) reviewed published evaluations of anadromous salmonid response to instream enhancement structures. They found that while many reported positive responses of juvenile coho salmon, steelhead, and cutthroat trout to instream enhancement, few of these responses were of long enough duration to detect statistically significant increases. Their synthesis and previous reviews (Reeves et al., 1991; Beschta et al., 1994; Chapman, 1996) emphasize the need for rigorous long-term evaluations of the responses of juvenile fishes to instream enhancement. While the results from these projects and those reviewed by Roni et al. (2002) are variable, they do indicate that when implemented correctly instream rehabilitation techniques benefit species and life stages of Pacific salmonids that prefer pool habitats.

Several studies have also demonstrated the positive effects of instream structure placement on Atlantic salmon. O’Grady (1995) reported higher levels of age 1+ Atlantic salmon and brown trout parr in sections of Irish streams treated with boulder structures including weirs, stone mats, and boulder clusters. Kelly and Bracken (1998) reported increased juvenile Atlantic salmon densities following placement of boulders, riprap, and deflectors in the Rye Water in Ireland, but no response for brown trout. In a similar study examining Atlantic salmon and brown trout in 13 Irish streams, Gargan et al., (2002) found significantly higher levels of both Atlantic salmon and brown trout parr in treated stream reaches, but no difference for older trout or for trout or salmon fry. Clarke and Scruton (2002) reported use of gravel additions by Atlantic salmon and high numbers of parr at gravel addition sites, though their monitoring was discontinued after 2 years. These studies suggest that placement of instream structures benefit juvenile Atlantic salmon, a species with habitat preferences similar to steelhead (rainbow trout).

Resident salmonids response-Early work examining the effectiveness of instream structures on resident trout suggested moderate increases of abundance, growth, condition, and survival of trout species associated with placement of structures (Tarzwell, 1938; Gard, 1961). Three compendiums on the response of trout to enhancement of Rocky Mountain and midwestern United States streams summarize much of this work (Hunt, 1988; Binns, 1999; Avery, 2004). In a review of 71 different instream enhancement projects installed between 1953 and 1998 in Wyoming, Binns (1999) detected more than a two fold increase in wild trout abundance either after treatment or between treatment and reference reaches for the 46 projects where fish data were available. However, few evaluations included more than a few years of data collection.

Numerous techniques for improving habitat for trout including brush bundles, current deflectors, bank cover, cut or cover logs, and LWD placement have been evaluated in Midwest streams. Hunt (1988) and Avery (2004) examined evaluations of instream habitat enhancement efforts in a total of 103 Wisconsin streams. Streambank debrushing and installation of brush bundles produced rather disappointing trout response in 11 different streams in Wisconsin (Hunt, 1988; Avery, 2004). Conversely, bank cover and current deflector structures demonstrated excellent results in increasing trout mean size and biomass, with 75 percent of the projects examined showing an increase in local abundance of 25 percent or more (Hunt, 1988; Avery, 2004).

Other North American studies not covered in these reviews have produced various results. An early 2-year study found increases in number and biomass of brook (Salvelinus fontinalis), brown, and rainbow trout following addition of brush bundles, and a decrease following removal of overhanging brush in other stream reaches (Broussu, 1954). Quinn and Kwak (2000) found more rainbow trout in a rehabilitated reach of an Arkansas River following placement of instream structures and bank protection. Saunders and Smith (1962) found increased levels of brook trout 1 year after placement of combinations of weirs, deflectors, and cover structures in a small stream on Prince Edward Island, Canada, though no increase in fish growth was detected. Whole or longitudinally cut half logs, anchored mid-channel, are used extensively in the Midwestern and the Eastern United States to provide cover for trout in both streams and lakes (Hunt, 1993). However, in a Pennsylvania study, Hartzler (1983) found no effect of half log cover structures on harvest, abundance, or biomass of brown trout. The results of these North American studies on instream enhancement demonstrate that overall instream enhancement efforts may increase local trout abundance and condition. However, multiple types of enhancement practices were implemented in many stream reaches and few projects were monitored long term (>5 years), making determination of which method led to increased fish abundance unclear in most instances.

Rehabilitation on channelized European trout streams have produced more consistently positive than results from American streams. Näslund (1989) examined four types of habitat improvement structures in relatively steep (3.3 percent) gradient Swedish stream and found brown trout densities increased up to three fold compared with nearby reference reaches. The most successful structures were boulder weirs (dams) and log deflectors, while boulder clusters and boulder deflectors appeared to be ineffective. Linlokken (1997) monitored the response of brown trout before and after construction of boulder weirs (10 years total) and found a 3 fold increase in brown trout abundance. Hvidsten and Johnsen (1992) found higher levels of juvenile brown trout following placement of boulder weirs in a recently channelized section of the River Soya, Norway. Zika and Peter (2002) reported increased rainbow trout and brown trout following placement of LWD in a Swiss stream. O’Grady et al. (2002) and Gargan et al., (2002) examined the effects of a variety of boulder and wood structures on brown trout in treatment and control reaches of 20 Irish streams and found higher levels of trout parr in treated stream reaches. Kelly and Bracken (1998) reported increased salmon densities but not brown trout in preliminary studies of boulder placement in the Rye Water, Ireland. The relatively consistent results of the European studies may be in part because the studies were more rigorous monitored than many North American studies, but also because the streams were highly simplified prior to enhancement and addition of instream structure may have resulted in larger changes in habitat quality (Roni and Quinn, 2001a).

Adult salmonids response-Where spawning gravels are in low abundance or of low quality, habitat structures such as channel-spanning LWD, boulder clusters, or gabions may recruit and store gravel (House et al., 1989; House, 1996). The evaluations of adult salmonid response to enhancement structures or gravel placement have been limited to a handful of short-term studies that demonstrated adult salmonid use of gravel accumulated at structures (Avery, 1996; House, 1996; Gortz, 1998), or observations of redds or adult spawning near enhancement sites (Moreau, 1984; Crispin et al., 1993; House, 1996; Iversen et al., 1993) (Table 10). For example, Gortz (1998) found three times as many brown trout redds in restore reaches of the River Esrom, Denmark, but found low numbers of fry, suggesting the spawning was not successful. Vehanen et al. (2003) found that adult grayling preferred a section of a Finish stream enhanced with boulder and cobble structures (reefs and islands). Iversen et al. (1993) reported increased fry and trout numbers in rivers Gudenå and River Hjortvad in Denmark, though they suggested projects on other Danish streams were not as successful because gravel was placed in slow water areas or because of siltation or poor water quality. In some cases the creation of spawning riffles is coupled with sediment traps (e.g. Avery, 1996), which has shown some level of success, but it is unclear how long the sediment traps will function before needing to be reexcavated.

TABLE 10
Summary of studies reviewed that examined adult response to instream habitat improvement.

This lack of rigorous evaluation of adult salmon and trout response to instream enhancement stems in part from the multiple generations and thus long time frame (>10 years) needed to detect an adult response to habitat alterations (Bisson et al., 1997; Korman and Higgins 1997). Nonetheless, evaluating adult response is critical for projects focusing on enhancement of spawning habitat and review of the limited published studies suggests increases in adult use of treated areas and increased fry (Table 10).

Nonsalmonid fish response-Less information exists on the response of nonsalmonid fishes to instream habitat enhancement (Bilby et al., 1996; Lonzarich and Quinn, 1995; Roni 2003) (Table 11). A review of habitat enhancement projects in warm-water streams throughout the Midwest found that a wide variety of techniques were successful in altering stream morphology and increasing stream cover, although rigorous physical and biological evaluation of individual projects was generally lacking (Lyons and Courtney, 1990). Lonzarich and Quinn (1995) found no effect of woody debris cover and depth on threespine stickleback (Gasterosteus aculeatus) or coast range sculpin (C. aleuticus) habitat use, growth, or survival in an artificial stream channel. Roni (2003) examined response of reticulate sculpin (Cottus perplexus), torrent sculpin (C. rhotheus), larval lamprey (Lampetra spp.), and Pacific giant salamanders (Dicamptodon ensatus) to LWD placement in the coastal Pacific Northwest of the US and found higher densities of lamprey in treated stream reaches, but little difference for other nonsalmonid species.

Other studies on nonsalmonid fishes have examined changes in species diversity. Shields et al. (1993, 1995b) found large increases in fish species diversity following boulder weir placement in two streams in Mississippi (USA). Angermeier and Karr (1984) examined responses of 10 warm-water fishes to wood removal and placement in a small stream in Illinois (USA) and found more and larger fish in stream sections with woody debris. Jungwirth et al. (1995) examined instream structures on fishes in seven Austrian rivers and found increases in both fish biomass and species diversity when treated sections were compared with natural sections. Pretty et al. (2003) found no improvement in fish species abundance or diversity in 13 English streams treated with flow deflectors or artificial riffles. However, bullhead (Cottus gobio) and stone loach (Barbatula barbatula) tended to be more abundant in rehabilitated riffles than control areas.

Monitoring the response of the entire fish community is important in determining both project effectiveness and whether instream techniques are restoring fish communities rather than just manipulating habitat for individual species. The studies discussed above suggest three major findings. First, in very diverse fish communities the increase and habitat complexity provided by instream habitat structures can lead to an increase in diversity. Second, projects designed for salmonids may have little effect on other species. Finally, little work has been done on nonsalmonid fishes and additional monitoring and evaluations considering all members of the fish community are needed.

TABLE 11
Studies examining response of non-salmonids or entire fish community to placement of instream structure.

Response to different techniques - Most instream rehabilitation projects involve the use of a variety of techniques to improve habitat, making comparison of different techniques difficult, and few studies have attempted to compare effectiveness among techniques. While we review them here, the results should be viewed with caution, as the effectiveness of a particular technique at a particular project can result from a variety of factors including project planning, construction, and recovery or integrity of watershed conditions and processes. Binns (1999) reported larger increases of trout numbers in constructed plunge pools or log weirs (129 percent increase) than at logjams (69 percent increase) and rock weirs (66 percent gain), but cautioned that many techniques were employed at various sites. Avery (2004) attempted to compare effectiveness among structure types using several fish success metrics and suggested that cover, current deflectors, and beaver dam removal achieved the best success ratings for resident trout. Slaney et al., (1994) and Van Zyll De Jong et al., (1997) provided more detailed comparisons of techniques within a given stream. Slaney et al., (1994) examined rootwad, floating cribs, pseudo beaver lodges and debris catchers in the Keogh River, British Columbia, and found higher levels of juvenile Chinook and steelhead in treated than untreated areas, with debris catchers appearing to be the most cost-effective. Van Zyll De Jong et al., (1997) examined the effects of three types of structures in a Newfoundland, Canada, stream and found increased Atlantic salmon (age 0+, age 1+, and age 3+) at boulder cluster sites, increased density of both brook trout and juvenile Atlantic salmon with v-weir placement, and increased age 0+ Atlantic salmon at half log (cover log) sites.

Other studies have found a strong correlation between the numbers of structures placed or amount of physical habitat change and fish response (i.e. Kennedy and Johnston, 1986; Roni and Quinn, 2001a). Kennedy and Johnston (1986) examined placement of stones in shallow areas (riffles) to improve salmon habitat in drainage maintenance (channelizing and periodic dredging of stream) and found a strong positive correlation between the number of stones placed in riffles and the juvenile Atlantic salmon density, while Roni and Quinn (2001a) found a positive correlation between amount of woody debris placed in the channel and juvenile coho salmon response to instream improvement. Roni and Quinn (2001a) found that the type of structure was not as important as whether a significant change in physical habitat occurred. This may also be the reason why projects in highly simplified streams seem to show the largest biological response. Moreover, as previously indicated, the success or failure of particular techniques at creating physical change often has to do with the appropriateness of the technique for that geomorphic situation and whether other larger factors such as water quality and riparian condition have been addressed.

Fish movement and watershed-scale evaluations-Examination of changes in fish abundance at instream enhancement projects can be complicated by immigration and emigration of fishes from nearby habitats or watersheds or by the effects of instream structures on fish movement. This is particularly evident for highly migratory fishes such as salmonids. For example, Riley and Fausch (1995) demonstrated an increase in abundance and biomass of three species of trout associated with structure installation in Colorado (USA) streams; however, a subsequent analysis of the same projects demonstrated that increases in biomass and abundance were strongly influenced by movement of trout into treated stream reaches (Gowan and Fausch, 1996). In contrast, Roni and Quinn (2001b) found little movement of juvenile salmonids between treatment and nearby reference reaches in a small Western Washington stream. Rinne (1982) reported reduced movement of gila trout O. gilae following the placement of high (>0.5 m) log dam structures. However, these studies did not examine population level responses but only local shifts in abundance.

The question of whether instream enhancement increases fish production or simply shifts fish distribution should be an important component of project evaluation (Gowan et al., 1994; Frissell and Ralph, 1998; Roni and Quinn, 2001b). Underlying this is the need to examine both physical and biological responses at a population or watershed scale. Few studies have examined whole watershed response to habitat modifications. Solazzi et al., (2000) found significantly higher abundance of juvenile coho following treatment. Conversely, Reeves et al., (1997) examined fish response to a suite of rehabilitation efforts in the Fish Creek Watershed in Oregon, and found little evidence for increased fish production at a watershed scale. The difference in these two watershed-scale studies is likely a result of both differences in study design and changes in physical habitat due to rehabilitation actions.

While most instream projects occur at a site or reach scale, these projects may produce responses or affect responses of physical habitat and fish production in downstream reaches, in adjacent habitats, or throughout a watershed. This has been a particular concern for highly migratory fishes such as salmonids (Northcote, 1992), which typically have seasonal habitat preferences (Nickelson et al., 1992; Roni, 2002). Thus, changes in one stream reach may affect salmonid abundance in adjacent stream reaches (Gowan and Fausch, 1996; Kahler et al., 2001; Roni and Quinn, 2001b). Assessing the biotic responses and the physical responses at a watershed scale is arguably more important (and more difficult) than examining reach-scale responses such as changes in local fish abundance. Both the costs and the difficulty in implementation have made examining the effect of individual or multiple projects on physical habitat, fish populations, and other biota at a watershed-scale relatively rare. Yet watershed-scale monitoring and evaluation is a particularly important perspective to consider and is needed to thoroughly adequately understand fish responses to habitat rehabilitation.

Macroinvertebrate response-Equally important to monitoring responses of fishes to habitat enhancement is the need to examine the responses of aquatic macroinvertebrates, which are an important food source for fishes and highly sensitive to habitat alteration and disturbance (Merritt and Cummins, 1996; Karr and Chu, 1999). Several authors have considered the response of macroinvertebrates as a component of their monitoring strategy, if not focusing exclusively on them (Table 12). Early investigations by Tarzwell (1938) and Gard (1961) demonstrated an increase in macroinvertebrates following instream enhancement. More recently, Gortz (1998) detected an increase in some species of macroinvertebrates following placement of instream structures and Wallace et al., (1995) found changes in functional feeding groups only within habitats altered by wood placement. Harper et al., (1998) and Ebrahimnezhad and Harper (1997) examined artificially constructed riffles in a lowland English streams and found that constructed shallow riffles (< 30 cm) had more diverse macroinvertebrate communities than deeper constructed riffles and communities more similar to natural control riffles. Conversely, Hilderbrand et al. (1997), Laasonen et al. (1998), Larson et al. (2001), Brooks et al. (2002), Black and Crowl (1995) and Roni (unpublished data) detected no difference of macroinvertebrate density or diversity in enhanced and untreated stream reaches. These studies suggest that macroinvertebrates were likely limited by primary productivity rather than habitat complexity. Several studies have examined the recovery time of macroinvertebrates following instream rehabilitation efforts (Table 12). Tikkanen et al. (1994) found only short-term impacts of habitat rehabilitation efforts (boulder placement) on macroinvertebrate abundance. These disparate results may be the result of differences in scale of measurement, metrics examined, differences in project objectives and physical habitat change, or perhaps that macroinvertebrates are not a good indicator of success of instream habitat enhancement for fish.

TABLE 12
Summary of studies examining macroinvertebrate response to instream habitat enhancement. All studies examined response at a reach or individual habitat unit scale.

A few studies have examined the recovery of macroinvertebrates following construction of instream enhancement efforts. Muotka et al. (2002) found that macroinvertebrate communities recovered within 4 to 8 years following disturbance with heavy equipment during construction of instream rehabilitation actions. Biggs et al. (1998) indicated that macroinvertebrate communities and aquatic macrophytes recovered within a year or more following intensive rehabilitation activities on the Brede, Cole, and Skerne rivers in Denmark and the United Kingdom. Limited evidence from these studies suggests that construction impacts are relatively short lived and that organic matter retention increases following increases in habitat complexity and instream structure.

Effects on organic matter and aquatic vegetation - The biological response to instream rehabilitation has focused on fishes and macroinvertebrates, though these activities can influence other species as well as other stream processes. Wood and other instream structure help increase retention of organic matter (Bilby and Likens, 1980) and placed wood has been shown to function in a similar manner (Trotter, 1990; Gerhard and Reich, 2000). In addition, habitat complexity and the retention of organic mater and nutrient spiralling are important. Muotka and Laasonen (2002) found increased organic matter retention following boulder and other structure placement in eight Finish headwater streams, but moss cover decreased dramatically following rehabilitation efforts. Wallace et al. (1995) and Negishi and Richardson (2003) found increased organic matter retention following wood placement and boulder clusters, respectively. These studies suggest that instream structures may have an influence on watershed processes such as nutrient and organic matter retention and utilization.

A few other studies have examined the effects of instream structures on aquatic macrophytes. For example, O’Grady (1995) found differences in aquatic macrophyte diversity between reaches of two Irish streams following placement of boulders. These studies suggest some improvement in macrophyte composition and growth from instream rehabilitation likely from changes in velocity and depth, while other studies (e.g. Iversen et al., 1993; Muotka and Laasonen, 2002) suggest disturbance due to construction of rehabilitation projects can lead to short-term impacts to aquatic macrophytes.

2.5.3 Conclusions - instream habitat structures

Our review suggests instream enhancement projects produce improvements in physical habitat and biota when implemented correctly and provides support for their continued use. However, given the variability in results for various species and structure types, the limited number of statistically rigorous studies, differential responses by different species or life stages, and the cost of instream enhancement projects, it is apparent that these projects should be undertaken with careful consideration of scale, watershed conditions and processes, and coupled with a rigorous monitoring programme. Several books are dedicated to the appropriate application and design of instream habitat rehabilitation and should be consulted (e.g. Slaney and Zoldakis, 1997; Cowx and Welcomme, 1998).

The following are key findings and factors to consider when implementing instream habitat projects.

• Placement of logs and log and boulder structures leads to improvements in physical habitat and local fish abundance when implemented correctly.

• Success of efforts is often tied to larger watershed-scale factors such as water quality, hydrology, sediment delivery, and riparian and upslope conditions.

• Gradient and stream power will influence the effectiveness of structures and many structural techniques are not appropriate in higher energy channels.

• The potential benefits of most instream structures may be short lived (< 10 years) unless coupled with riparian planting or other process based restoration activities that will lead to long term recovery of deficient processes.

• While placement of instream structures appears to be successful at increasing local fish abundance, particularly for salmonid fishes, results are highly variable by species and life stage as well as structure type.

• The most successful projects are those that create large changes in physical habitat and mimic natural processes.

• Watershed and population level responses to placement of instream structures for both physical habitat and biota are still largely unknown.

2.6 Lake habitat enhancement

2.6.1 Common lake habitat enhancement techniques

As with the placement of instream structures, the placement of artificial structures in lakes has been conducted since at least the early 1930s (Brown, 1986; Bolding et al., 2004). Habitat enhancement in reservoirs is particularly common because of the lack of substrate complexity and cover (Vogele and Rainwater, 1975; Prince and Maughan, 1978; Wills et al., 2004). The purposes of artificial structures in ponds, lakes, and reservoirs are similar to those for artificial reefs in the marine environment: to concentrate fish, increase harvest levels, and increase localized or overall biological production in the water body (Rountree, 1989; Bassett, 1994; Bohnsack et al., 1997; Lindberg, 1997; Bolding et al., 2004). In addition, structures are often placed in aquaculture ponds to increase fish production or in natural floodplain lakes and other habitats to attract fish to improve harvest (Welcomme, 2002). As no studies have examined physical effects, we summarize the biological effectiveness of structures in lakes.

Lake habitat enhancement structures can be categorized based on the types of material used and whether they are designed for cover or for spawning habitat. These include cover structures made of trees or artificial materials, addition of gravels or boulders, and planting of vegetation (Table 13; Figure 10). Below we briefly discuss each type and what is known about their effectiveness.

2.6.2 Effectiveness of lake habitat enhancement

The goals and effectiveness of structures placed in lakes and ponds to improve fish habitat have been periodically reviewed and summarized (Prince and Maughan, 1978; Brown, 1986; Tugend et al., 2002; Wills et al., 2004). For example, in a recent survey in the United States, Tugend et al. (2002) found that 75 percent of the lake habitat enhancement included placement of structures, while 25 percent were spawning structures or substrate manipulations.

Habitat enhancement structures can produce positive results through increased spawning recruitment and survival, refuge or cover from predation, shade and sites for orientation and schooling, can increase prey as well as feeding efficiency of adults, and provide increased harvest and fishing opportunities (Bolding et al., 2004; Wills et al., 2004). Prince and Maughan (1978) suggested that habitat structures in ponds and lakes increase both fish prey and production. However, a number of factors affect the success of structures including the structure type, construction material, interstitial spaces, depth and location within the lake, lake morphometric characteristics, substrate complexity, presence of aquatic macrophytes, and the fish species and life stage examined (Walters et al., 1991; Lynch and Johnson, 1989; Moring et al., 1989; Rogers and Bergersen, 1999; Tugend et al., 2002; Wills et al., 2004; Bolding et al., 2004). Moreover, fish use of structures varies daily (deil variation), seasonally, and with water temperature (Prince and Maughan, 1978; Moring and Nicholson, 1994). Similar to studies of artificial structures in the marine environment, the effect of these structures on fish population size or fish production has not been examined (see Hoff, 1991; Tugend et al., 2002).

TABLE 13
Common pond, lake, and reservoir physical habitat enhancement techniques and their goals and effectiveness. See section on lake habitat enhancement for references.

SHASTA-TRINITY NATIONAL FOREST

TENNESSEE WILDLIFE RESOURCES AGENCY

KENT GILGE, MONTANA DEPARTMENT OF FISH WILDLIFE AND PARKS

FIGURE 10. Photographs of common lake habitat enhancement techniques used in North America including brush bundles (top), stake bed (middle), and rock shoal for walleye spawning habitat (bottom).

Brush bundles and log cover structures-Brush bundles, logs, and half-log structures are made of natural material and are typically placed on or suspended near the bottom (Brown, 1986). Brown (1986) and Tugend et al., (2002) reviewed published evaluations of brush bundles and indicated that several studies have reported increased abundance and catches of fishes at treatment (brush structures) versus control sites (no brush added) particularly for centrarchids and cyprinids (Brown, 1986; Graham, 1992; Moring and Nicholson, 1994; Tugend et al., 2002), but less so for other species such as yellow perch (Perca flavescens) (Moring et al., 1989).

Cover logs or placement of whole logs or trees into lacustrine environments typically have the same goals as brush bundles. Similarly they have been shown to have a positive response for centrarchids and other warmwater fishes (Moring et al., 1989; Wills et al., 2004). For example, smallmouth bass (Micropterus dolomieu) numbers, nest density, and nest success improved following placement of half logs in both lakes and reservoirs (Hoff, 1991; Wills et al., 2004). Submerged logs left over from logging operations supported five of 10 species found in reservoir in Maine, US (Moring et al., 1989). There is also some evidence that a combination of brush, vegetation, and gravel placed in the littoral zone can benefit young of year pikeperch (Stizostedion lucioperca) even in the presence of a predator (Northern pike, Esox lucius) (Lapínska et al., 2001).

While a number of studies have examined the response in North America lakes, placement of brush and logs into lakes and floodplain habitats to enhance fish production is also utilized in the developing world (Welcomme, 2002). For example, dropping of logs to rehabilitate designated fish sanctuaries has occurred in the Tonle Sap Great Lake in Cambodia (Thuok, 1998). Similarly, Vinci et al. (2003) described three Indian variations of placing bamboo and brush into floodplain ponds in India. Brush bundles or brush parks are widely used in floodplain lakes and ponds to enhance both native and cultured fish production in developing countries (Hoggarth et al., 1999; Quasim, 2002). These are primarily installed to provide cover for fishes and many different modifications exist particularly for fish culture practices (Quasim, 2002; Figure 11). However, in some areas this practice is illegal or they are designed as refuges or fish sanctuaries and fishing in these areas is illegal. Studies of their effectiveness are rare, but evaluations of fish sanctuaries and brush parks in Bangladesh have shown them to be highly effective at both attracting fish and recovering endangered species (Poulsen and Hasan, 2004).

Cover structures made of artificial materials - A variety of artificial materials have also been used for creating cover and habitat complexity in lakes. Stake beds, concrete slabs or structures, plastic crates, boxes, and pipes and tire were the most commonly reported in the literature (Brown, 1986). Stake beds are typically composed of hundreds of stakes driven into the substrate or prefabricated onshore prior to installation. These artificial structures are designed to provide cover, holding, and in some cases spawning habitat.

The results for structures made from artificial materials have been less consistent than for brush bundles or logs. A handful of studies have reported positive results from placement of highly artificial structures in lakes. Moring and Nicholson (1994) examined fish use of cinder blocks, tire bundles, and brush bundles and found higher levels of pumpkinseed sunfish (Lepomis gibbosus), chain pickerel (Esox niger), brown bullhead (Ameiurus nebulosus), common shiner (Luxilus cornutus), and golden shiner (Notemigonus crysoleucus) and found higher densities of all species near artificial structures than sites without. Sampling before and after placement of cinderblock and log structures, Graham (1992) found higher numbers of small bluegill (< 15cm) in Lake Anna, Virginia (USA) following treatment. Prince and Maughan (1979a,b) found that tire reefs significantly concentrated both bass and sunfish in another Virginia lake. Wege and Anderson (1979) reported higher growth rates of largemouth bass (Micropterus salmoides) and higher catch rates of both bluegill and largemouth bass in ponds treated with tires and stakes beds or brush bundles than control ponds.

In contrast, other studies have found no change in fish use or higher fish use of structures composed of natural materials versus artificial materials. Wills et al. (2004) examine plastic box shaped shelter devices as a habitat enhancement device and found no change in fish use before and after placement of plastic cover structures (Aqua Cribs). Pardue and Nielsen (1979) examined tilapia (Tilapia spp.) and bluegill (Lepomis macrochirus) response to pine board structures in small experimental ponds and found little evidence that structures increased fish biomass. Wilbur (1978) examined clay pipes and cement blocks with brush as fish attraction devices and found several warmwater species used the structures. However, structures that contain trees or brush were preferred by fishes over structures made of purely artificial materials (Wilbur 1978).

FIGURE 11. Brush parks or "samrah". A typical samrah brush park on the Tonle Sap River, Cambodia with bushes and branches gathered together with submerged poles or fences to prevent them being carried away (top). Highly effective at attracting fish, they are illegal in many areas. Modification on the samrah in Bangladesh using water hyacinth (bottom). Here the brush park is legally used as a fish sanctuary where fishing is not allowed.

Rogers and Bergersen (1999) examined a suite of highly artificial structures fastened to steel frames and found higher largemouth bass use of structures containing trees and no use of any of the artificial structures by northern pike. Johnson and Lynch (1992) found fish use (bluegill sunfish and white crappie, Pomoxis annularis) and angler success higher at brush bundle and evergreen sites than stake bed structures. Moreover, Hunt et al. 2002) found that artificial structures were only successful at providing cover and nesting habitat for largemouth bass if they were similar to natural habitat (logs or brush) and placed in the littoral zone where bass typically spawn and woody debris was historically most abundant. The advantages of log structures and brush bundles over structures made of artificial materials apart from their apparent higher level of success is that they mimic what may have been found in many lakes prior to lakeshore development and removal of riparian vegetation.

While only a few studies have examined cost-effectiveness, it appears that more natural materials, such as brush, trees, logs, or rock and gravel, are not only more likely to be utilized by fishes but are more cost-effective. This is in part a result of the high cost of fabricating many structures composed of artificial material (Johnson and Lynch, 1992). Studies such as Paxton and Stevenson (1979) and Wege and Anderson (1979) have reported higher catch rates at artificial structures than natural unstructured areas and suggested that artificial structures may lead to increased exploitation from fish harvest rather than an increase in population.

Similar to studies on fish aggregation devices in the marine environment (Lindberg, 1997), logs and other floating material have also been demonstrated to attract fishes in lakes (Helfman, 1979). For example, Helfman (1979) reported higher numbers of bluegill, black crappie, and golden shiner were more abundant near or under floating structures, while several other cyprinids and centrarchids showed little affinity for structures. Floating structures provide cover for both schooling fishes and predators; thus it is unclear which species or trophic levels benefits most from these types of structures. As discussed in the section on instream rehabilitation techniques, rigorous studies are needed to determine whether placement of structures in lentic habitats leads to increased production or simply just concentrates fish.

Artificial shoals or spawning areas-Creating spawning areas within lakes typically involves the placement of gravel, cobble, boulders, or concrete rubble to create suitable areas. As previously discussed, many cover structures such as brush bundles and logs are designed to provide cover for both juveniles and spawning fishes (Tugend et al., 2002; Wills et al., 2004). In the Great Lakes of North America, artificially constructed reefs or spawning shoals account for much of the lake trout (Salvelinus namaycush) spawning areas in Lakes Ontario and Michigan (Fitzsimons, 1996). In a recent review of research on use of artificial reefs in the Great Lakes, Fitzsimons (1996) reported higher levels of eggs, fry, and young-of-year at artificial reefs and spawning areas. However, success of naturally produced lake trout in the Great Lakes is poor and other factors besides spawning appear to be limiting the success of artificial reefs at increasing fish production.

Artificial reefs can provide spawning habitat for a number of warm-water fishes such as sunfish (Lepomis macrochirus) and walleye (Stizostedion vitreum vitreum) and provide for increased catch rates for yellow perch and other species (Brown, 1986; Kelch et al., 1999). Geiling et al. (1996) reviewed the effectiveness of creation of artificial reefs to create walleye spawning habitat in the Great Lakes and found that it rarely increased adult walleye abundance and suggested that this was not a successful technique for this species. Similarly in a review of 20 rock placement projects designed to improve walleye habitat in Wisconsin lakes (USA), Neuswanger and Bozek (2004) reported that 85 percent showed no signs of increased fry production. However, the success or failure of artificial spawning areas in lakes or rivers at increasing fish populations is driven in part by their location relative to rearing areas for juveniles (Jones et al., 2003). The information on the effectiveness of spawning habitat enhancement in lakes is limited but appears to only be successful if they are carefully designed and placed with consideration of the biology and ecology of the species of interest.

Shoreline vegetation-The planting of lakeshore vegetation (riparian vegetation) provides a number of benefits to lakes including shade, source of organic matter, protection from wind (which can change mixing), bank protection, and food (terrestrial insects), and is a source for brush, logs and other cover structures. In addition, lakes or reservoirs with seasonally inundated areas are planted with vegetation which provides good habitat for young of year fish such as bass (Brouha and von Geldern, 1979; Brown, 1986). Little biological evaluation has occurred following planting of lake shores, but the planting of water tolerant vegetation such as Salix spp. helps stabilize banks and prevent erosion, which may also improve fish habitat near structures and in the waterbody.

Concentration versus increased harvest-In both the freshwater and marine environment, there is a concern that placement of structures is only leading to concentration of fish and increased harvest rates rather than increased production (Roni and Quinn, 2001b). Paxton and Stevenson (1979), Wege and Anderson (1979) and others have reported higher catch rates at artificial structures than natural unstructured areas and suggested that artificial structures may lead to increased exploitation from fish harvest rather than an increase in population. Additional work is needed to clarify whether this is true or not.

2.6.3 Conclusions - lake habitat enhancement

The existing literature on addition of structures to lakes and ponds to improve habitat suggests that natural structures such as logs and brush bundles lead to increases in local abundance of some species. The relative success of structures made of natural materials is also supported by ecological studies which demonstrate the importance of riparian areas and woody debris along the shoreline and littoral zone of lakes (see Christensen et al., 1996). The literature also indicates that placement of habitat structures in lakes is an effective strategy for providing refuge areas, fishing opportunities and has a long history of use in aquaculture. We make the following recommendations based on our review of the literature.

• Addition of logs, wood, and brush bundles is a suitable form of habitat enhancement in areas were cover is limiting fish production.

• Structures made of artificial materials are not as effective as those constructed of logs or brush.

• Spawning reefs constructed of gravel, rock and boulders have not been demonstrated to be and effective method for increasing production.

• A number of factors effect project success including: fish species, limiting factors, location of structures (depth, vicinity to shoreline, proximity to other important habitats), predators-prey interactions, increased harvest due to fish concentration, and whether the structures mimic natural habitats.

• Additional studies are needed to determine population level responses to habitat enhancement in lakes.