Open Channelization of a Roadway Culvert System
Assessing wildlife benefits of a project in Connecticut
Anthropogenic land-use changes to riparian wetlands in urban settings result in scale-specific animal responses (Mensing et al. 1998; Nerbonne and Vondracek 2001). This includes disturbances to natural streambeds and engineered stormwater structures (e.g., culverts and open channels) intended to divert or change water flow velocity with the construction of roadways and crossings. The term open channel is used in this article to distinguish it from naturally occurring stream channels.
Both culverts and open channels involve different design considerations. Culvert engineering primarily involves hydrological calculations (e.g., peak discharge, minimum and maximum velocity, runoff and drainage amounts, and headwater depth) and design factors (e.g., type, roughness, size, length, and slope). Culverts can be placed within natural stream channels to control water flow or alternatively open channels are used. Channel engineering also involves similar hydrological calculations; however, it must also consider the channel roughness related to whether it will be an unlined, concrete lined, or grass-drainage channel. In open channel designs, the size, slope, and shape of the bottom and sides must be factored to avoid stream sediment erosion, erosion not being a factor for culverts except at the outlet/discharge, where it can be reduced by placing a splash pad. Whereas culverts are not intended to function as a best management practice (BMP) or improve water quality of collected stormwater or runoff, channels can have such intent. Additionally, the design of an open channel might also include the restoration of a streambed to mimic natural conditions after its location to provide suitable habitat for aquatic animals.
The disturbance history of a stream channel can change the benthic macroinvertebrate fauna inhabiting a streambed, particularly if flooding results in sediment scour or fill (Matthaei and Townsend 2000, Effenberger et al. 2006). Additionally, BMPs can change instream physical habitat variables (e.g., sedimentation, bank stability, and embeddedness of substrate), and therefore invertebrate and fish assemblages (Nerbonne and Vondracek 2001). Whereas information regarding the construction of BMPs and their stormwater management benefits (e.g., pollution reduction, runoff volume reduction, stream channel protection, and peak flow control) are discussed in US stormwater quality manuals (e.g., Connecticut Department of Environmental Protection 2004), they lack specific information on the post-construction effects of structures on aquatic biota. Biomonitoring that is required from US federal or state wetland permitting tends to focus on mitigation areas rather than impact sites.
The culverting and open channelization of rivers and streams is most often the focus of fisheries, civil engineering, stream restoration and bank stabilization, and flood control projects in urban environments (Riley 1998). Where the water quality of culverted or channeled streams is measured, Rapid Bioassessment Protocols (RBPs) (Plafkin et al. 1989) provide one tool to assess the potential threat of pollution to aquatic indicator animals, such as freshwater benthic macroinvertebrates. Comparing macroinvertebrate occurrences is useful to assess local land-use management and BMPs and their potential effects on stream biota among sites (Whiles et al. 2000, Nerbonne and Vondracek 2001). Because they are sensitive to disturbance and water pollution, macroinvertebrates can indicate habitat quality and a stream’s suitability to support the different life stages of aquatic animals.
During the inland wetlands and watercourses permitting process, the impact that a proposed placement of culvert or channel has on wetland functions and resources is considered. Although the specific effects of a structure on a protected animal species can be a consideration, the potential effect on a species guild or the common aquatic animals that are part of localized stream ecology may not warrant review or be prudent to assess. More often the effects of channelization on aquatic animals is the focus of restoration ecology and fisheries (Brooker 1985, Henegar and Harmon 1971, Whitaker et al. 1979, Nakano and Nakamura 2006) rather than BMPs guiding land development and land-use decision making (Burton et al. 2005).
A historic open channel located in the urban environment of Middlebury, CT, provided the opportunity to assess the long-term stream effects from open channelization on aquatic biota. The goals of this study were to: 1) evaluate the impact of disturbance history of the Wooster Brook Open Channel (WBOC) in west central Connecticut by surveying its macroinvertebrate community, and 2) use this information to guide the design of open channels in urbanized streams. The effects of open channelization on habitat quality and biota were evaluated in the fall of 2000, 2001, 2003, 2004, and 2007. This habitat evaluation involved three parts: 1) water-quality testing, 2) RBP for wadeable streams, and 3) vegetation and wildlife surveys. The habitat evaluation made it possible to assess whether the open channelization of Wooster Brook mimics local natural conditions and has comparable water quality supporting wildlife.
Case Study: Wooster Brook
Wooster Brook exists in New Haven County, CT, and the Housatonic Drainage Basin. It is an approximately 0.414-km-long stream that originates from Tracy’s Pond and flows through suburban, commercial, and federal lands. This stream discharges into Hop Brook and the Hop Brook Lake Dam Management Area of the US Army Corps of Engineers (USACE). As part of an effort to control future flood damage, after the 1955 Hurricanes Connie and Diane flooded the Naugatuck River, a dam was constructed in 1968 above the point where the Hop Brook flows into the Naugatuck River, Wooster Brook, a perennial stream, is part of this flood control system. Long-term impacts to the stream include its diversion, channelization, and culverting from the construction of an interstate highway (I-84) and local roadways (Connecticut Route 63 and Country Club Road). Additionally, this stream receives runoff from several roadways and effluent from unknown sources resulting in regular positive tests for fecal coliform levels (unpublished data based on past field surveys conducted along with this study). Management of the WBOC is the responsibility of the New England District of the USACE.
Wooster Brook empties into the open channel through a set of dual, concrete pipe culverts (Figure 1). The inlet flow into the channel drops approximately 10.2 cm depending on water flow. Concrete remains suggest that a splash pad once existed at the outlet. The channel dimensions are approximately 34.1 by 5.8 m. Bottom substrate consists of a gravely mix and various size rocks, and there are two sediment deposits promoting vegetation growth. The northwestern deposit is approximately 7.3 by 2.1 m, and the southwest approximately 8.5 by 4.3 m. The sidewalls are 2.1 m high and constructed of vertical placed stone. Records from the Connecticut Department of Transportation for Route 63 include a site plan dated 1935, suggesting the open channel was constructed after the culvert installation, which was commissioned in 1932. Based on this information, the WBOC is best described as a channel or canal for floodwater retention, although this most probably was not the original function of the design, because it predates regional flooding caused by the September 1938 hurricane.
Water-quality testing of the open channel followed standard RBPs (Barbour et al. 1999, Beauchene et al. 2002) for the rapid bioassessment of streams. Water samples were collected and tested onsite for water-quality measures of dissolved oxygen (DO), pH, nitrate-nitrogen (NO2), and orthophosphate (PO4) using LaMotte Company (Chestertown, MD) water monitoring test kits. Average measures from Wooster Brook were compared to its closest second order stream, Hop Brook (Sullivan 2001–2006), the Shepaug River (USGS 2005), Housatonic regional water basin, and statewide values available from Beuachene et al. (1998).
Macroinvertebrate Community Survey
Freshwater benthic macroinvertebrates were grab-sampled using a kick net of 800- by 900-µm mesh and D-framed dip-netting (650-µm mesh) following the RBP II Method (Plafkin et al. 1989; USEPA 1997a, b; Barbour et al. 1999). Onsite sorting and laboratory identification of specimens took place, and then they were returned to the point of collection to avoid death, following USACE park management regulations. Identification was done using a stereomicroscope and keys and descriptions developed by Beauchene and Hoffman (2000) for the Connecticut Ambient Water Quality Monitoring Program, Connecticut Department of Environmental Protection.
Following the RBP II Method, macroinvertebrates were identified to the least family level, assigned pollution and disturbance tolerance values, and classified to feeding groups to calculate RBP indices (USEPA 1997a, b; Barbour et al. 1999; Beauchene and Hoffman 2000). RBP indices were based on individual and cumulative taxa counts over the period 2000–2007 to reflect an overall response to long-term effects. RBI indices calculated included taxa richness, EPT (Ephemeroptera-Plecoptera-Trichoptera) Index, S:C:F (Scraper: Collector/Filter ratio), EPT:C (Chironomidae) ratio, contribution of the dominant family, modified HBI (Hilsenhoff Biotic Index), and the CLI (Community Loss Index, Plafkin et al. 1989). The formula for calculating the HBI is:
HBI = S(tv)i ni N-1
where (tv)i is the tolerance value of the i-th taxon, ni is the abundance of the i-th taxon, and N is the total number of individuals in the replicate. The modified HBI is the HBI value divided by the total number of specimens in the sample.
A bioassessment of the benthic community and water quality followed Platkin et al. (1989) based on the calculation of percent similarity among WBOC, Hop Brook, and the closest state reference (Beauchene et al. 2002), which is the Shepaug River. Using the metric scores, RBP II thresholds for condition, water quality, and attributes were then determined. These metrics are useful for distinguishing the degree of impact and disturbance between sites (Plafkin et al. 1989, Whiles et al. 2000). The percent similarities among these sites were compared using the Jaccard Coefficient of community similarity (Plafkin et al. 1989). The Jaccard Coefficient is defined as:
J = c/(a+b+c)
where J is the similarity among two samples, “a” is the number of taxa in sample A, “b” is the number of taxa in sample B, and “c” is the number of taxa shared by both samples.
Vegetation and Wildlife Survey
Additionally, observations of vegetation and wildlife were made during each field season. The vegetation survey considered the entire open channel as a single stand for which the most dominant plants were identified (modified from Mueller-Dombois and Ellenberg 1974 and Barbour et al. 1987). Native and non-native status of plants followed Dowhan (1979), and plants identified by the Connecticut Invasive Plants Council were noted. Additionally, the wetland dependency of plants found in the northeast region followed the US Fish and Wildlife Service (1988). The wildlife survey included aquatic netting, minnow trapping, and visual encounter searches for living individuals, signs of activity, and animal remains.
Results and Discussion
The inorganic substrate components of the WBOC include gravels (65%), sands (20%), and clay (15%). Organics of the stream bottom include vegetated detritus (60%) and muck/mud (40%). These characteristics reflect an artificial substrate bottom developing for over 70 years and now approximating natural upstream and downstream conditions.
The results of water-quality testing are presented and compared to local, regional, and statewide values (Table 1). Regional differences in temperature and dissolved oxygen are perhaps the most significant factors affecting aquatic animal diversity among streams within a regional watershed (Tarplee et al. 1971). The water quality of the WBOC tends to be most similar to its nearest brook, Hop Brook. Regionally, the WBOC is more acidic and has a higher NO3/NO2 than the Shepaug River reference site (i.e., one of the rivers having the highest water quality in Connecticut) and Housatonic Drainage Basin. WBOC pH levels are more acidic than the range of means for Connecticut streams and rivers, but approximate rainwater tests for Middlebury (Long 2005). The WBOC has been affected by the upstream presence of fecal coliform bacteria and filamentous spirogyra algae, both of which occur also in the channel.
|a USGS 2005
|a TV (Tolerance Values); b Abbreviations: C-F (Collector-
Filterer); C-G (Collector-Gatherer); P (Predator);
Sc (Scraper); Sh (Shredder)
A total of 15 macroinvertebrate taxa have colonized the WBOC from an aggregate sample of 162 specimens collected from five years between 2000 to 2007 (Table 2). Annual abundance ranged from eight to 68 specimens with the greatest number collected in the first season (2000). Taxonomic richness ranged from two (2003) to seven (2000 and 2004) with Hydropsychidae found in four collecting seasons (Table 3). Dominant taxon groups include Hydropsychidae, Isopoda, and Chironomidae. Non-insects represent the most common functional feeding group, collector-gatherers, followed by Chironomidae and Baetidae. The pollution tolerance values (TV) ranged from 8 (least) to 3 (most sensitive) with an even taxa frequency among values 3, 6, and 8. The greatest number of macroinvertebrates represented moderately pollution sensitive taxa (TV=4, n=72) and included Baetidae, Hydropsychidae, Elimidae, and Arachnida.
Macroinvertebrate indices, based on aggregate samples, characterizing the WBOC include EPT=2, which is low compared to the reference site (10) and Connecticut range (9 to 16). The S:C:F ratio (0.098) calculated also was low compared to the reference site (0.97) and Connecticut range (0.22 to 5.04). EPT:C ratio (64) was higher than the reference site (8.0) and within the Connecticut range (8 to 90.5), and the Percent Dominant Family (Hydropsychidae=59%) was higher than the reference site (23%) and out of range (21% to 38%). HBI (4.89) also was higher than the reference site (3.32) and out of range (2.64 to 3.33). The overrepresentation of Trichoptera and non-insects (e.g., Isopoda) resulted in high HBI values for 2003 and 2004. Based on these metrics, the percent comparison of the WBOC to the nearest stream (Hop Brook) was 60% and reference site (Shepaug River) was 53%. The taxonomic similarity (Jaccard Coefficient) was 0.15 to Hop Brook and 0.17 to the Shepaug River.
The vegetation survey (Table 4) resulted in 17 dominant plant species, of which 83% (n=5) of the plants with a wetland indicator status are hydrophytes. Eighty-eight percent of the plants (n=8) were native. Three species, purple loosestrife (Lythrum salicaria), Japanese honeysuckle (Lonicera japonica), and multiflora rose (Rosa multiflora) are invasive species in Connecticut.
Fisheries and wildlife occurred infrequently during the five collecting seasons. Fish found included individuals of native yellow perch (Perca flavescens), rainwater killifish (Lucania parva), and common shiner (Notropis cornutus). A single salamander larva of northern dusky salamander (Desmognathus fuscus) inhabited shallow water with a stony bottom.
Channel type, water source, and water quality are known factors affecting macroinvertebrate assemblages in urban environments (Burton et al. 2005). WBOC non-insects (40%) and tricopterans (54%) dominated the macroinveretebrate assemblage. In urban environments, those aquatic families that are most tolerant to disturbance tend to be dominant (Barbour et al. 1999; Effenberger et al. 2006). The stone-walled WBOC most likely mimics a concrete-lined channel, and therefore was expected to have low measures of taxa abundance and higher measures for non-insects (Burton et al. 2005). The WBOC reflects this characterization; however, the dominance of the moderately pollution tolerant (Tolerance Value=4, where 10 is the least tolerant) Hydropyschidae suggests suitable water quality. Hydropsychidae also occur in Hop Brook and Shepaug River.
|a Abbreviations: EPT (Ephemeroptera-Plecoptera-Trichoptera Index); Sc (Scraper); CF (Collector-Filterer); HBI (Modifi ed Hilsenhoff Biotic Index); bHyd (Hydropsychidae); NI (Non-Insect)
|a I (Introduced); I* (Introduced Invasive); N (Native) after Dowhan 1979
b FAC (Facultative); FACU (Facultative Upland); FACW (Facultative Wetland); NR (Not Reported); OBL (Obligate);
UPL (Obligate Upland); WI (Wetland Indicator Status); - (Regionally found less frequently in wetlands); + (Regionally
found more frequently in wetlands) after USFWS 1998
However, the overall presence and abundance of Hydropsychidae and Chironimidae, considered pollution tolerant, in the WBOC sample indicates a water quality that is under stress even though the habitat condition may be suitable. This is suggested by the bioassessment of the benthic community and water quality (Plafkin et al. 1989) between WBOC and Hop Brook, which suggested that channel habitat is slightly impaired and water quality very good to good to the local natural conditions and is moderately impaired and fair to poor compared to the reference site. Additional support is provided by the resulting Jaccard Coefficients among the WBOC, Hop Brook, and the Shepaug River. WBOC only shared three aquatic insect families with the reference site and only two with Hop Brook. Not surprisingly, WBOC is most similar to Hop Brook (Jaccard Coefficient=0.27) than to the reference site.
Management Implications and BMPs
The results suggest that aquatic macroinvertebrates, fisheries, and wildlife will reclaim even small open channels, and the design of culvert systems should consider this option. This finding complements other investigations of channel effects on riparian wetlands and aquatic animals (Whitaker et al. 1979, Edwards et al. 1984, Burton et al. 2005). Information regarding the characteristics, environmental condition, and biological condition of the WBOC is useful in developing BMPs that can result in increased water quality and habitat suitability. Additionally, visual observations (Dirrigl, unpublished data) from a larger channel system (Trout Brook) in West Hartford, CT, also assist this review. The recommendations and BMPs that are presented might help guide the design of open channels and consider their use in combination with roadway culvert systems with a focus on design benefits to wildlife. The following summaries adapted from available literature and their applicability to the WBOC and Trout Brook Channel are explored.
* The open channel allows light and light penetration into the stream bed. The WBOC provides for a break in the culvert system that provides light to the aquatic organisms in the channel water. Light penetration differentially supports, aquatic plant, macroinvertebrate feeding groups, and fish (Liboriussen et al. 2005).
* The open channel provides a portion of stream bed substrate that approximates natural conditions. The original WBOC stream bed consisted of gravel and rock riprap without any consideration to mimicking the natural condition of the surrounding stream. However, after years of water flow and flooding episodes resulting in the deposit of sediments, some degree of naturalness has developed. This is not meant to imply that stream restoration efforts are not necessary; it is only after 70 years of that the WBOC approximates a natural stream supporting aquatic life.
* Potentially, an open channel can function as a small refugium for developing aquatic life. Salamander larvae and juvenile crayfish occur in the WBOC area. However, as fish are flushed into the channel, the opportunity for predation can increase until the fish are later flushed out into the natural stream where cover would be provided. Unfortunately, flood management restrictions prevent any placement of living or dead vegetation in the WBOC that would provide cover from prey. The height of the culverts entering and exiting the channel also prevents any animal movement below the effluent from entering the channel. In this scenario, the choice of open-bottom box culverts connecting the channel rather than concrete piping would be of greater benefit to wildlife.
* Increases in aquatic and terrestrial wildlife use can result from open channelization of culvert systems. The WBOC supports invertebrate wildlife; however, except for aquatic amphibians and reptiles, only urbanized mammals (e.g., gray squirrel (Sciurus carolinensis), Norway rat (Rattus norvegicus), opossum (Didelphis virginiana), and raccoon (Procyon lotor) most likely frequent this small area. Access to the channel is provided only through the culverts or climbing down the vegetation on the steep banks. Although a bird survey was not part of the WBOC study, numerous different birds perch on the sidewall and streamside vegetation and drink from the stream. Design considerations such as bank slope and channel access, should consider the habitat requirements of local wildlife. If not for the WBOC location in a busy road and commercial area, it might be expected that animals could fall into and be harmed by the open channel because of the extreme height of the sidewalls.
* Aquatic life can benefit from the inclusion of different water depths in the design of open channels. Likewise, the flow of the water should mimic the character (e.g., riffles, pooling, flats, scour holes) of the stream waters found above and below the structure. Arranged rock placement (e.g., deflectors) and structures (e.g., gabions) can assist with creating these water characteristics (Barton and Cron 1979). The wildlife survey indicated that different animals occurred in the shallow versus pool portions of the WBOC. In the Trout Brook channel, riprap check dams provided water pools that Canada geese (Branta Canadensis) regularly frequent during the winter months.
* Management should focus on increasing plant diversity and abundance and reducing turbidity (Stewart and Downing 2008). Vegetation should be planted rather than allowed to naturalize in the open channel. The plant selection should include species that are tolerant of flooding episodes if one of the channel’s functions is floodwater retention. Also, salt-tolerant species can be included if roadway runoff is high in salt content. This would be a consideration in US states (e.g., Connecticut) where road and ice removal programs include rock salt treated with liquid calcium chloride. The naturalization of the WBOC resulted in almost equal amounts of native versus non-native species and promoted the growth of invasive species. Whenever invasive plants are found, management must include their control based on each species’ biology.
* In the US, state minimum requirements for channel construction must always be part of design and engineering. However, engineering that considers the geometry of the channel (triangular v-ditch, rectangular, or one-stage trapezoidal versus two-stage natural designs) in addition to discharge intervals can result in more effective control of water flow, reduce bank erosion, and improve water quality (Doll et al. 2001, Evans et al. 2004, Ward et al. 2004). The WBOC was not designed following the recent Connecticut culvert standards for fisheries; however, it was oversized to accommodate large flood episodes.
* Equally important is that state guidelines for fish passage be applied to fish-bearing streams. A review of these documents shows that each culvert and channel design option has different benefits to fish. No slope and flat-bottom designed culverts (without debris racks) and roughened, sinuous channels with riffles seem to provide the best fish passage conditions and habitat to support the macroinvertebrates they feed on (Zwim 2002, Bates 2003, Saldi-Caromile et al. 2004, Halwas et al. 2005). The WBOC consists of circular, concrete culverts that empty into a flat-bottomed, straight channel that most likely was not designed for fish passage even though fish travel through the system. However, a positive response to the coarse gravel and cobble substrate of the WBOC is suggested by the macroinvertebrates, salamander larvae, and juvenile crayfish found. Wildlife using the WBOC would benefit more from a design that consisted of: 1) “habitat complexity and hydraulic diversity,” and 2) altering the “physical habitat suitability for various species, which is a function of substrate type, velocity, depth, and bank characteristics” (Saldi-Caromile et al. 2004:9).
* The banks of open channels benefit from designs that include bioengineering over structural techniques (IWLA 2006) to provide edge habitat to wildlife. The WBOC consists only of steep vertical banks of staked, quarried rock and lacks any habitat value. The Trout Brook channel has varying banks that include regular cuts, shelved banks, vegetated banks, cement banks, and riprap banks, all of which provide wildlife access to the stream. My winter observations at Trout Brook showed that wildlife traveled regularly along the cement banks of the channel to access the stream. However, it is not known if the observed wildlife activity would decrease without snow cover.
* In urban areas that are susceptible to litter, a planted tree barrier is likely necessary to prevent debris from entering an open channel. This tree barrier also could provide shading to the channel to control water temperature. With any planting plan, the selection of trees should mimic the surrounding dominant forest species, which also will increase plant survivability in the first growing season providing the proper soil organic content is established. The sides of the WBOC channel are cut into lawn, and therefore prone to litter falling into the channel. The soil around the WBOC supports lawn and not species of the surrounding natural vegetation, and therefore is prone to litter and increases in water temperature.
* Biomonitoring of open channels must include pre- and post-construction sampling of macroinvertebrates that includes defining a reference collection (King et al. 2000). Assessing the historical effects of the WBOC was not possible because this study was the first attempt to examine macroinvertebrate responses to its construction. However, even the short duration, 2000 through 2007, of this study allowed for annual changes to be observed and to access the overall ability of the system to support wildlife.
This study demonstrates that even small areas engineered as open channels within a culvert system can develop suitable aquatic habitat. Hupp (1992) estimated that disturbed stream channels can achieve a state of naturalness in 65 years, which is approximately the time period in which the WBOC was constructed. Channel designs that avoid concrete bottoms and instead include natural stream bottoms support greater macroinvertebrate diversity and abundance (Burton et al. 2005). Edwards et al. (1984) indicated that mitigated channels are beneficial to macroinvertebrate and fish diversity and relative abundance. Although the WBOC habitat cannot mimic natural conditions some, benefits can still be obtained when the open channel functions as a refugium for developing larvae (e.g., crayfish, salamanders) especially during episodes of drought (see Griswold et al. 1982).
The uniqueness of this case study is that aquatic life continues to be supported even though the WBOC 1) is a fragmented “island” surrounded by roadway and commercial development; 2) was not designed for any ecological restoration, habitat use, or fish passage (i.e., the channel contains artificial substrate and has extremely high slopes (2.1-m high)); 3) at first observation would appear to not provide any suitable habitat; and 4) must be maintained by the USACE primarily for flood control purposes (i.e., removal of vegetation and debris that could impede flow). Based on the observations of this case study, a series of management considerations and BMPs were presented. Although these BMPs and considerations are not novel, the references cited and discussions presented are meant to encourage continued collaborative canal design by biologists and engineers.
Author's Bio: Frank J. Dirrigl Jr., Ph.D., is an assistant professor of environmental science in the Department of Biology at the University of Texas-Pan American, Edinburg, TX.