We examined the impact of road crossings on fish assemblages in

Report as inappropriate Your report has been received

Shared by: 6ncVz9M

Tags

Stats

views:: 2
posted:: 11/23/2011
language:: English
pages:: 25

Public Domain

Document Sample

ABSTRACT

We examined the impact of road crossings on fish assemblages in the Upper

Greenbrier watershed in the Monongahela National Forest and found significant

differences between the fish above and below culverts. Fish assemblages above culverts

were significantly poorer in both number of species present and in relative abundance of

fishes compared to downstream populations. A priori best professional judgment of good

and bad culverts was not a consistent and reliable predictor of impact. There was a

strong correlation between the product of pipe length and culvert slope and depauperate

upstream fish assemblages, suggesting that culvert design can restrict movement of

aquatic organisms. We suspect that features of culvert design that increase both water

velocity in culverts and the distance between resting pools prevents fishes from being

able to pass through culverts, leading to isolated upstream populations. This study shows

that road crossings can negatively impact non-migratory aquatic communities.

INTRODUCTION

Roads have a tremendous impact on surrounding terrestrial and aquatic

ecosystems (Forman, 2003). In the eastern U.S., road crossings on private and

government land number in the hundreds of thousands and average more than X road

crossing per stream mile (Table X). This may potentially have a dramatic impact on the

fragmentation of both migratory and non-migratory fishes. Road crossings that act as

barriers to fish movement may prevent both migration of anadromous species to

spawning areas as well as preventing recolonization of species to areas that have suffered
localized extinctions due to the effects of flood, drought, fire or other damaging events

(Morita 2002, Katopodis 1999).

Culvert design plays a large role in fish passage. Road crossings that simulate

stream morphology pass fish more effectively, especially small, non-game species

(Warren, 1998). Culverts may act as velocity, exhaustion, jump or behavioral barriers to

fish movement because of steep slopes, long pipes, outlet drops and substrate (Katopodis,

1999; Warren 1998).

Road crossings also affect stream habitat by increasing sedimentation and altering

both large-scale and small-scale stream channel morphology (Forman, 2003, Wellman

2000, Harper 2000). Small-scale changes include deep pools below culverts formed by

scouring, loss of riparian vegetation and channelization of the stream (Lim; Katopodis;

Kosicki). Sediment accumulation increases below culverts but is not necessarily linked

with changes in fish communities (Wellman, 2000). At a larger scale, culverts can

restrict the movement of streams, preventing streams from meandering (Forman, 2003).

Road crossings that act as a barrier to fish movement can also prevent the recolonization

of stream areas impacted by natural disturbances.

Culvert slope, outlet drop, and the ability of the culvert to replicate natural stream

habitat are consistently identified as important factors in determining whether or not a

culvert can pass fish effectively. In this study we evaluated changes in fish assemblage

and relative abundance upstream of culverts. We classified culverts as “good” or “bad”

for fish passage based on best professional judgment in addition to actual culvert

measurements of slope, pipe length and outlet perch.
STUDY SITE

This study was conducted in the Upper Greenbrier, Glady Fork and Shavers Run

watersheds, three fifth level watersheds in the Monongahela National Forest, West

Virginia (Figure 1). The majority of culverts used in the study were found in the Upper

Greenbrier watershed; only four were located in Glady Fork and Shavers Run

watersheds.

METHODS

Culvert Selection and Measurement

We surveyed all culverts on streams greater than 1m wetted width within the

Upper Greenbrier watershed and classified them as a priori “good” or “bad” for fish

passage based on best professional judgment. After a priori classification, a total of 33

culverts were then selected based on approximately equal distribution of “good” (n=16)

and “bad” (n=17) groups.

A level, leveling rod and tape measure were used to determine the physical traits

of each culvert (Figure 2). Elevation readings at the culvert inlet and outlet were taken to

yield culvert slope. Elevation readings at the pool bottom, water surface and pool outlet

were taken to determine the outlet drop and outlet perch. Stream elevations 50m above

and below the culvert were also taken to ensure upstream slope was similar to that below.

Measurements of pipe length, pipe size, dimensions of the pipe and pipe corrugations

were taken with measuring tape. Stream channel and culvert substrate composition were

determined by using a modified Wentworth scale to estimate the three most dominant
substrates in each situation. We selected culverts on streams that were not significantly

different in terms of their habitat upstream and downstream (Table 1).

Fish Sampling

At each site, a buffer equal to 20 times the channel width or a minimum of 80m of

stream separated the culvert from the sections sampled for fish assemblage (Figure 3).

This was done in order to minimize the impact of the road itself, such as the alteration of

the stream channel and loss of riparian vegetation. The upstream and downstream

sampling areas were twenty times the channel width long and broken down into four

sections of equal length. The sections were sampled by one pass backpack electrofishing

using two netters. Data for the four sections were kept separate. Larger sections of

stream often used multiple netters and two shockers. Fish were identified and measured.

Several specimens were preserved in 95% ethanol for later identification and as vouchers;

all others were released.

In addition, we made a relative abundance classification based on the number of

specimens from each species found in each of the four sections. Classifications were

defined as: abundant:3 or more individuals found in all 4 study sections; common: at least

1 individual found in all 4 study sections; patchy: individuals found in only 3 study

sections; rare: individuals found in 1 or 2 study sections; absent: no individuals found in

any study section (Figure 4).

Statistical Analysis

We used a Chi-square test to compare relative abundance and species richness

above all culverts and also to compare our a priori “good” and “bad” classifications for
species richness and relative abundance. A Student’s t-test was used to compare culvert

measurements of factors associated with species loss.

RESULTS

5,924 specimens were collected from 33 sites. Significantly fewer fish species

were found above culverts (Table 1). There was no significant difference between the a

priori good and a priori bad culverts in terms of species loss (Table 2). Relative

abundance significantly decreased above culverts (Table 3), and decreased to an even

higher degree of significance above a priori bad culverts (Table 4).

There was no significant difference in the means of culvert slope, outlet drop, and

pipe length in the post-sampling analyses we performed (Table 5). The product of pipe

length and culvert slope in culverts that lost species upstream of culverts was

significantly higher than culverts that had no species loss upstream of culverts.

Seven culverts had fish species downstream but none upstream. We separated

these culverts from the other culverts that had lost species upstream but still had at least

one species found upstream. The product of pipe length and culvert slope was

significantly higher than in other culverts for the culverts that had no upstream species

(Fig. 5). These seven culverts were also significantly different than all other culverts in

terms of upstream channel width, downstream channel width, and downstream channel

slope (Table 6).

DISCUSSION
This study demonstrates a relationship between road crossings and depauperate

fish assemblages. We suspect that road crossings can act as barriers to fish movement

when factors such as culvert slope, outlet drop or pipe length exceed a threshold, which

varies by species. However, our study was unable to demonstrate a consistent

relationship between these individual factors and decreased fish populations (Table 5).

We believe that many culverts we a priori classified as bad for fish passage are indeed

acting as barriers to fish movement, but upstream populations have been self-sustaining.

These populations would be more vulnerable to localized extinctions that accompany

events such as fire, drought or floods. The long-term viability of isolated populations

such as those above a barrier is unlikely (Morita, 2002).

After sampling, we compared culverts that either lost no fish species in the

upstream sections or maintained an equal number of species as the downstream section to

those culverts that had fewer species upstream than they did downstream. No single

factor explained species loss, but there were significant differences between the two

groups when a combination of factors was examined. This suggests that a number of

factors are critical for determining whether or not a culvert is suitable for fish passage.

Comparing the products of pipe length and culvert slope may be a better reflection of

culvert suitability than any other factor alone because it combines two features that are

consistently identified as vital to determining fish passage.

Channel widths among the culverts that had no species upstream were much

narrower than other culverts, and we believe this reflects that vulnerability of smaller

streams to localized extinctions. The process of extinction may also be accelerated in

smaller streams due to the smaller populations found in these streams. Although culverts
that had no fish upstream had steeper slopes than other culverts, brook trout, one of the

most common species among the small mountain streams we sampled, have

demonstrated that they can migrate through streams that are steeper than any we sampled

(Adams, 2000). We believe that channel slope did not play a large role in limiting the

upstream movement of species in our study.

Water velocity data are not available for the culverts we sampled, making it

difficult to compare our study sites to experiments done in laboratory settings that test

swimming performance of fishes in relation to water velocity. However, we believe that

a number of culverts used in our study may have water velocities approaching or

exceeding the limits of the burst speed of many fish species, especially during high flow

events. Figure 3 shows a culvert that was included in our study that illustrates some of

the problems we believe are affecting fish movement. Water velocity within the culvert

appears to be high as well at normal flows and likely much greater during high flow

events. This poses a significant problem to benthic species that are poor swimmers, like

sculpin and darters, because a high flow event may be the only time that the stream is

raised to the point that the culvert outlet is submerged. A velocity barrier in high flows

may replace a jump barrier during normal flows.

Weak-swimming fish have difficulty maintaining burst speeds in currents as low

as 25-cm/s, and we believe that velocities equal to or exceeding this may occur in such a

culvert (Toepfer 1999). More important to weak swimming fish than the current may be

the distance between the pipe outlet and the stream surface, which in this case exceeds

two feet. The combination of potentially swift currents and a substantial outlet perch lead
us to a priori designate this culvert as bad for fish passage because we believe it acts a

barrier to fish movement.

Brook trout, blacknose dace and sculpin spp. were found in both upstream and

downstream sampling sections of the culvert shown in Fig. 3, and only fantail darters

were found downstream but not upstream. We suspect that brook trout may be able to

successfully pass through this culvert, but blacknose dace and sculpin are not robust

swimmers and thus may not be able to leap over 2 feet into the culvert outlet. Blacknose

dace and sculpin may be surviving in the upstream areas even though they are isolated

from the downstream populations. With an average bankfull channel width of over 19

feet and a great deal (see if we can get a real number to put here) of viable habitat

upstream, populations of fish upstream may be able to successfully maintain steady

numbers for years after they have been isolated.

We witnessed several occasions in which crayfish and salamanders were unable

to move through culverts. In these instances, the crayfish and salamanders were typically

found near the entrance of a culvert with a submerged outlet. The current was not strong

enough to prohibit them from entering the pipe, but they were washed out before they

could approach the culvert inlet. This may be due to the water velocity increasing toward

the middle of the culvert or because the crayfish and salamanders had become exhausted

and were not able to continue. We also observed one instance of a brook trout beaching

itself on the ground near a culvert in an attempt to leap into the culvert. The culvert in

this case consisted of multiple slick pipes that were all raised above stream grade. All of

the pipes were relatively small, approximately 2 feet in diameter. This culvert was not
used in study because it was too close to the main stem and thus did not have sufficient

area for either a buffer or the minimum sampling distance of 80m.

Consecutive culverts may damage stream communities further. Within the Upper

Greenbrier watershed, we found several streams that had multiple culverts installed

within several hundred meters of each other. We believe that this may lead to

fragmentation of the stream environment to a greater extent than streams with only one

crossing. Our observations of the consecutive culverts installed on streams leads us to

suspect that any fish populations between these culverts may be especially vulnerable to

natural disturbances.

In order to confirm that a culvert is capable of passing fish, we must be able to

demonstrate that a fish has moved from an area downstream of a culvert to an area

upstream of a culvert. We hope to be able to mark individual fish using passive

integrated transponder (PIT) tags and monitor their movement over time. Information

gained from this could provide useful information about which culverts are acting as

barriers to fish movement. Additionally, we believe that examining the genetic diversity

in streams with culverts may reveal if populations are indeed isolated from downstream

reaches.

Acknowledgements

Funding for this project was provided by the USDA Forest Service’s National Aquatic

Ecology Unit. The USDA Forest Service Center For Aquatic Technology Transfer

provided support in our sampling. We would also like to thank the Department of
Biology at James Madison University and the Greenbrier Ranger Station in the

Monongahela National Forest for their assistance.
Table 1. Comparison of the mean channel slope and channel width of upstream and
downstream sections of all culverts used in the study.

Upstream Downstream P-value

Channel Slope -3.757 -4.275 0.257

Channel Width 16.061 15.383 0.360
FIGURES, TABLES, STATS

Fish good vs. fish bad culvert slope: p=0.2503

Ditto, outlet drop, p=0.4003

Ditto, pipe length, p=0.6459

Ditto, pipe length x culvert slope, p=0.0866

The worst vs. fish bad, culvert slope, p=0.0029

Table 5

Culverts Culverts not p-value
losing losing species
species upstream
upstream
Mean Culvert slope -3.244 -1.988 0.2503

Mean Pipe length 40.21 43.96 0.6458

Mean Outlet drop 0.7236 0.4285 0.4003

Mean Pipe length x -83.16 -167.61 0.0866
culvert slope
Mean Outlet perch 0.58 0.15 0.2781

Table 6

Culverts with no Culverts with at p-value
species found least one species
upstream found upstream
Downstream -6.80 -3.54 0.002
channel slope
Upstream channel -5.56 -3.23 0.185
slope
Downstream 9.71 17.04 0.004
channel width
Upstream channel 9.43 17.99 0.002
width
100

0
1
Pipe Length x Culvert Slope

-50

-100
No Upstream Species
-150 Fewer Upstream Species
No Loss of Species Upstream
-200

-250

-300

-350

-400

Figure 5. Comparison of the product of pipe length and culvert slope based
upon changes in the upstream fish population.
Table 1 REDO CHI-SQUARE!!!!

Above culverts Fewer Species Equal More Species Total
22 9 2 33

C
21
REDO CHI-SQUARE!!!!
Table 2
33
Above culvert More Species Equal Species Fewer Species Total

Good 2 4 10 16

Bad 0 5 12 17

Total 2 9 22 33

Table 3

Chi-square=18.31; p=0.0001; significant.

Relative Greater Equal Less Total
abundance Abundance Abundance Abundance
above culvert
29 63 71 163

Table 4 – REDO CHI-SQUARE!!!! (have not made changes to this since switching ef09
from good to bad)

Relative Greater Equal Less Total
abundance Abundance Abundance Abundance
good vs bad
Good 25 52 36 113
Bad 4 11 35 50
Total 29 63 71 163
LITERATURE CITED

Angermeier, P.L. and M.R. Winston. 1999. Characterizing fish community diversity

across Virginia landscapes: prerequisite for conservation. Ecological

Applications 9(1):335-349

Chipps, S.R., W.B. Perry and S.A. Perry. 1993. Status and distribution of Phenacobius

teretulus, Etheostoma osburni, and “Rhinichthys bowersi” in the Monongahela

National Forest, West Virginia. Virginia Journal of Science 44(1):47-58.

Chipps, S.R., W.B. Perry and S.A. Perry. 1994. Fish assemblages of the central

Appalachian Mountains: an examination of trophic group abundance in nine West

Virginia streams. Environmental Biology of Fishes 40:91-98.

Chipps, S.R., W.B. Perry and S.A. Perry. 1994. Patterns of Microhabitat Use Among

Four Species of Darters in Three Appalachian Streams. American Midland

Naturalist 131:175-180.

Dennis, T.E., S.E. MacAvoy, M.B. Steg and A.J. Bulger. 1995. The association of water

chemistry variables and fish condition in streams of Shenandoah National Park.

Water, Air and Soil Pollution 85:365-370.
Ead, S.A., N. Rajaratnam, C. Katopodis, and F. Ade. 2000. Turbulent open-channel

flow in circular corrugated culverts. Journal of Hydraulic Engineering(NEED

VOLUME NUMBER!)750-757.

Fitzhugh, R.D., T. Furman, J.R. Webb, B.J. Cosby and C.T. Driscoll. 1999.

Longitudinal and seasonal patterns of stream acidity in a headwater catchement

on the Appalachian Plateau, West Virginia, U.S.A. Biogeochemistry 47:39-62.

Forman, R.T.T., editor. 2003. Road Ecology: Science and Solutions. Island Press,

Washington, D.C.

Kelly, G.A., J.S. Griffith and R.D. Jones. 1980. Changes in the distribution of trout in

Great Smoky National Park, 1900-1977. U.S. Fish and Wildlife Service,

Technical Paper of the Fish and Wildlife Service 102.

Katopodis, C. 1999. DON’T KNOW FORMAT- article title Sustaining fish migrations:

The challenge of road culverts. Appears in “Proc., Nordic Conf. On Fish Passage” also

“Direktoratet Naturforvaltning,” Olso, 138-141. ILL notice cites

Kamula, R. and Laine, A. (eds) of Proceedings of the Nordic Conference on Fish

Passage, pub. 1999, ISBN 82-7072-316-9; BIOL 639.2 Nordisk symposium om

fiskepassajer; article is a chapter out of that book.
Kosicki, A.J. and Davis, S.R. CHECK FORMAT- I THINK IT’S A GOVT PUB.

Consideration of stream morphology in culvert and bridge design. Transportation

research record 1743, paper no. 01-2466. 57-59.

Lim, S.Y. 1995. Scour below unsubmerged full-flowing culvert outlets. Proceedings of

the Institution of Civil Engineers- Water, Maritime and Energy. 112:136-149.

Little, J.D. and J.J. Mayer, 1993. Pages 221-226. In C.F. Bryan and D.A. Rutherford,

editors. Impacts on Warmwater Streams: Guidelines for Evaluation. Southern

Division, American Fisheries Society, Little Rock, Arkansas.

Lonzarich, D.G., M.L. Warren, Jr., and M.R.E. Lonzarich. 1998. Effects of habitat

isolation on the recovery of fish assemblages in experimentally defaunated stream

pools in Arkansas. Canadian Journal of Fisheries and Aquatic Sciences.

55:2141-2149.

Isaak, D.J. and W.A. Hubert. 2000. Are trout populations affected by reach-scale stream

slope? Canadian Journal of Fisheries and Aquatic Sciences. 57:468-477.

Morita, K. and A. Yokota. 2002. Population viability of stream-resident salmonids after

habitat fragmentation: a case study with white-spotted charr (Salvelinus

leucomaenis) by an individual based model. Ecological Modelling 155(1):85-94.
Murphy, G. and M.R. Pyles. 1989. Cost-effective selection of culverts for small forest

streams: minimizing total discounted cost. Journal of Forestry 87(10):45-50.

Newman, K. and A. Dollof. 1995. Responses of blacknose dace (Rhinichthys atratulus)

and brook char (Salvelinus fontinalis) to acidified water in a laboratory stream.

Water, Air and Soil Pollution 85(2):371-376.

Slawski, T.M. and T.J. Ehlinger. 1998. Fish habitat improvement in box culverts:

management in the dark? North American Journal of Fisheries Management

18(3):676-685.

Taylor, C.M and M.L. Warren, Jr. 2001. Dynamics in species composition of stream

fish assemblages: environmental variability and nested subsets. Ecology

82(8):2320-2330.

Toepfer, C.S., W.L. Fisher, and J.A. Haubelt. 1999. Swimming performance of the

threatened leopard darter in relation to road culverts. Transactions of the

American Fisheries Society 128(1):155-161.

Vaughan, D.M. 2002. Potential impact of road-stream crossings (culverts) on the

upstream passage of aquatic macroinvertebrates. Report of United States Forest Service,

San Dimas Technology and Development Center. Located in San Dimas? Need to have

a city.
Warren, M.L. Jr. 1998. Road Crossings as Barriers to Small-Stream Fish Movement.

Transactions of the American Fisheries Society 127:637-644.

Wellman, J.C., D.L. Combs, and S.B. Cook. 2000. Long Term Impacts of Bridge and

Culvert Construction or Replacement on Fish Communities and Sediment

Characteristics of Streams. Journal of Freshwater Ecology 15(3):317-328.

Welsh, S.A. and S.A. Perry. 1998. Habitat partitioning in a community of darters in the

Elk River, West Virginia. Environmental Biology of Fishes 51: 411-419.

THIS ONE IS ALL FUNKY. CHECK AGAIN.

Harper, D.J. and J.T. Quigley. (NEED TO FIND THE YEAR FOR THIS) No net loss

of fish habitat: an audit of forest road crossings of fish-bearing streams in British

Columbia, 1996-1999. Canadian Technical Report of Fisheries and Aquatic Sciences

2319.

D.J. Harper and J.T. Quigley. Fisheries and Oceans Canada.
Scour below an overfall: Part I. Investigation

Robinson KM, Hanson GJ, Cook KR

TRANSACTIONS OF THE ASAE

45 (4): 949-956 JUL-AUG 2002
Evolution of cutoffs across meander necks in Powder River, Montana, USA
Gay GR, Gay HH, Gay WH, Martinson HA, Meade RH, Moody JA
EARTH SURFACE PROCESSES AND LANDFORMS
23 (7): 651-662 JUL 1998
Consideration of stream morphology in culvert and bridge design
Kosicki AJ, Davis SR
HYDROLOGY, HYDRAULICS, AND WATER QUALITY; ROADSIDE SAFETY
FEATURES
TRANSPORTATION RESEARCH RECORD
(1743): 57-59 2001
Long-term impacts of bridge and culvert construction or replacement on fish
communities and sediment characteristics of streams
Wellman JC, Combs DL, Cook SB
JOURNAL OF FRESHWATER ECOLOGY
15 (3): 317-328 SEP 2000
Population viability of stream-resident salmonids after habitat fragmentation: a
case study with white-spotted charr (Salvelinus leucomaenis) by an individual based
model
Morita K, Yokota A
ECOLOGICAL MODELLING
155 (1): 85-94 SEP 15 2002
Effects of habitat fragmentation by damming on the persistence of stream-dwelling
charr populations
Morita K, Yamamoto S
CONSERVATION BIOLOGY
16 (5): 1318-1323 OCT 2002
Abundant- 3 or more specimens collected in
all 4 sections
Common- At least one specimen collected in
all 4 sections
Patchy- At least one specimen collected in 3
sections
Rare- Specimens collected from only 1 or 2
sections
Absent- No specimens collected

Figure 4. Relative abundance definitions
Figure 1. 5th level watersheds containing land that is part of the
Monongahela National Forest; insert at right depicts the roads and
streams within the Upper Greenbrier River watershed.
Figure 2. Diagram illustrating culvert measurements taken.
Figure 3. Illustration of the buffer and sampling areas within sampled
streams.