-
https://cpw.cvlcollections.org/files/original/61cc04fda6c4f3ea496a87e4a81cddfc.pdf
68bc3e014da0996cc570899c52b52500
PDF Text
Text
Evaluation of an Electric Fish Barrier on the South Canal, an Irrigation Ditch on the Lower Gunnison
River, Colorado
Dan Kowalski
Colorado Parks and Wildlife Aquatic Research Section
Montrose, CO 81401
Abstract
An electric fish barrier was installed on the east portal of South Canal to reduce fish entrainment
associated with the construction of two hydropower plants in 2012. The objective of this study
was to monitor fish entrainment and evaluate the effectiveness of the barrier. Three groups of
fish were tagged and released upstream of the barrier; fish from the canal, wild Gunnison River
fish, and hatchery reared fingerlings. Mark recapture boat electrofishing was completed and
population estimates were made with the Huggins Closed Capture model using fish length to
model capture probabilities. The study reach contained 2,994 ± 1,043 fish (>150 mm) in October
2011, 1,764 ± 279 in October 2013, 1,224 ± 239 in July 2014 and 1,900 ± 379 in October 2014. Fish
population estimates have declined after the electric barrier, significantly at the 95% level for
brown trout but not for rainbows. A total of 288 tagged fish less than 300 mm and four fish greater
than 300 mm were recovered below the barrier, representing 1.3% of all tagged fish. The electric
barrier appears to meet its objective and successfully exclude larger fish from the study reach, but
not smaller age 0, age 1, or age 2 trout. The entrainment, growth and survival of smaller fish
maintains a stable population of fish in the canal, but fewer entrained mature fish is likely a benefit
to the fish population of the Gunnison River. Further study is needed to evaluate if smaller adult
trout can be successfully excluded by the electric barrier with operational modifications.
There are over 105,000 irrigation structures on rivers and streams across Colorado, most in fish bearing
waters. Fish entrainment in irrigation canals is known to be a large problem in the western U.S. (Carlson
and Rahel 2007) and the loss of fish in irrigation canals has been shown to be a population sink for trout
in Wyoming (Roberts and Rahel 2008). The impact of fish lost to irrigation canals on fish populations in
Colorado is unquantified. The South Canal is an irrigation ditch in southwest Colorado that diverts an
average of 360,600 acre feet of water each year, about 857 cfs average daily flow March-November, from
the Gunnison River for agriculture (Bureau of Reclamation 2012). The river contains a Gold Medal trout
fishery despite documented entrainment of fish for many years in the canal. The construction of a
hydropower plant was expected to increase mortality of entrained fish so an electric fish barrier was
installed at the diversion structure in 2012. From the diversion structure and barrier, the canal travels
through a 5.7-mile-long tunnel before egressing approximately 0.5 miles above the power house (Figure
1). There is a total of 7.7 miles of earthen canal that contains the majority of fish that are entrained from
the Gunnison River. The canal diverts water from March through November each year with the amount
of water depending on water supply and irrigation demand. During winter months the canal is generally
shut off with only a very small amount of flow as a result of accretions and seepage. About twice a month
it is partially opened to run approximately 100 cfs through the canal for 24-48 hours to fill a drinking water
supply reservoir. Because of low and intermittent flows in the canal, fish survival over winter was
generally thought to be low but variable year to year depending on frequency of freezing temperatures.
However, in the winter of 2012-2013, a constant flow of 20-25 cfs was run all winter long to keep water
supply reservoirs full during construction of the hydropower plant. This resulted in what appeared to be
a much larger number of fish in the canal in spring of 2013 due to increased survival of entrained fish.
The study reach for this project was downstream of the concrete drop below the West Portal (just below
1
�the first powerhouse) and was 0.72 miles, ending at the 2nd concrete drop structure (Figure 2, UTM NAD83
258703, 4262335). The canal averaged 46.1 feet wide with 20-25 cfs in March 2013 and 70.2 feet wide at
540 cfs in October 2013. The study reach represents 9.4% of the total earthen portion of the South Canal
but is suspected of containing the highest density of entrained fish due to its proximity to the West Portal.
While fish routinely pass through the high velocity concrete portions of the canal, the majority of fish
reside in the lower gradient earthen portion of the canal.
Figure 1. Area map of the Gunnison Tunnel and South Canal (Bureau of Reclamation 2012). The study
site is between the West Portal and Drop 1.
The fish barrier was constructed in 2012 and was operational before the 2013 irrigation season. It consists
of a series of vertically suspended electrodes across the east portal of the Gunnison Tunnel (Figure 3).
The waterway at the barrier is 74 ft wide, 16 ft deep, and has water velocities between 0.2-0.7 m/s (0.662.3 fps) and conductivity of 180 µs/cm. The system is powered by three 1.5 KVA Smith Root pulsators
with a max power output of 4.5 kW and is designed to operate with a frequency of 2Hz, pulse width of
0.005 s and a field strength of 1v/inch (0.4v/cm). The barrier was designed to exclude “brood stock”
rainbow and brown trout but target size was not specified (Smith Root 2011). The barrier is believed to
have operated continuously as planned throughout the entire 2013-2014 irrigation seasons.
Communication has been lost for brief time periods (i.e. 6 out of over 6,000 hours of operation in 2013)
but operation of the barrier was thought to be unaffected and it is assumed that is has functioned
continuously during irrigation season the last two years (J. Heneghan, personal communication).
The purpose of this study was to estimate fish populations in the South Canal before and after the barrier
2
�and investigate the entrainment of fish from the Gunnison River. To accomplish this, fish population
estimates were compared before and after the barrier was built over different seasons and across years
while tagged fish were used to document any movement across the barrier.
Study
Reach
Figure 2. Fish sampling site on the South Canal. The sampling reach was 0.72 miles long (3,802 feet) and
was between the first and second concrete drop structures below the West Portal.
METHODS
South Canal was sampled with mark-recapture electrofishing (October 2011, October 2013, July 2014 and
October 2014) and multiple pass removal (March 2013) to estimate fish populations of adult and juvenile
trout. The study reach for all three occasions was the same but differing methods were used in the spring
sampling because of the different habitat and flows when water is not being diverted (20-25 cfs vs. 500900 cfs).
On March 29, 2013, the canal consisted of two distinct habitat types, consisting of the concrete stilling
basin just below the first drop and the earthen portion of the canal below. The density of fish was much
higher in the stilling basin and the physical habitat dictated that different sampling methods be used in
the two locations. The reach was stratified by habitat types and two sampling reaches were chosen. The
entire stilling basin was sampled with 50 ft. bag seine that was 6 ft. deep with 1/8 in. mesh. Multiple seine
hauls were made through the stilling basin so a depletion population estimate could be made (Zippin
1956, White et. al 1982). Fish were held in a live pen and then measured for total length to the nearest
millimeter. Capture probability was high (estimated to be 0.74 for rainbows and 0.79 for browns) and
model assumptions of closure appeared to have been met due to the isolated and simple structure of the
3
�stilling basin. The high capture probability and lack of evidence of size selectivity of the seine is expected
to help meet assumptions of the removal model and there was no evidence in the data to indicate an
unacceptable amount of bias. The portion of the canal below the stilling basin consisted of shallow, slow
moving channel that was 46.1 ft. wide 3,528 ft. long. A sampling reach was randomly chosen in this
portion of the study reach that was 1,000 ft long and block nets were used to ensure closure. Five Smith
Root LR24 backpack electrofishers were used to complete a two pass removal population estimate. Fish
were held in a live pen and then measured to nearest millimeter and weighed to the nearest gram, and
then returned to the canal. After the March estimate, 876 fish were removed from the canal in an effort
to depopulate the study reach before the barrier’s first season of the operation. One hundred and twentyfive fish from the stilling basin were tagged with coded wire tags (CWT) and adipose fin clips and
transported by aerated fish truck to the Gunnison River in East Portal. They were stocked at the boat
ramp approximated 0.7 miles above the East Portal and the barrier.
Because electrofishing removal estimates are known to be biased low due to size selectivity and individual
capture heterogeneity, we took several approaches to reduce this bias recommended by Riley and Fausch
(1992) and Saunders et al. (2011). First efforts were made to use sufficient effort for high capture
probabilities. Second, capture probabilities were modeled by fish species and length to account for
heterogeneity. The data was analyzed in Program Mark with the Huggins Closed Capture Model (White
and Burnham 1999, Huggins 1989). To reduce the bias associated with the size selectivity of electrofishing,
capture probabilities were modeled with length as a covariate similar to the approach described in
Saunders et al. 2011. Four models were built by estimating capture probabilities by length, species,
species + length, as well as a constant capture probability for all fish. Model selection was done with AICc
and population and parameter estimates were made by model averaging across all four models with AICc
weights (Burnham and Anderson 2002). To estimate the total trout in the study reach in March 2013, the
two pass removal estimate was expanded for the length of canal that contained similar habitat and added
to the estimate for the stilling basin. The confidence intervals were calculated by summing the variances
of each estimate (Delta Method) and multiplying by 1.96.
Four groups of fish were tagged and released in East Portal upstream from the Gunnison Tunnel to
challenge the barrier. One hundred and twenty-five fish (59 brown trout and 66 rainbow trout) from the
March 2013 sampling of the stilling basin were moved from below the barrier to above and received both
coded wire tags and adipose fin clips. Mean length of the tagged fish was 241 mm for brown trout (range
165-310 mm) and 232 mm for rainbows (180-392 mm). Wild fish were captured by boat electrofishing on
June 17 and 19, 2013, in the Gunnison River above the barrier and tagged with both coded wire tags and
adipose clips. A total of 1,265 fish (653 rainbow trout and 612 brown trout) were tagged, the mean length
of brown trout was 281 mm (103-737 mm) and 336 mm (82-547 mm) for rainbows. Fingerling rainbow
trout from the Rifle Falls Fish Hatchery were also tagged and released into the Gunnison River in East
Portal above the barrier. A total of 19,800 fish with a mean length of 68 mm were tagged with coded wire
tags on June 24-26, 2013 and stocked into the Gunnison River 0.7 m above the barrier on July 26. Due to
the results of the first study season, the focus in 2014 was on tagging larger fish and 1,841 wild fish from
the Gunnison River above the barrier were tagged with 32 mm half duplex PIT tags. The mean length was
396 mm (200-545 mm) and an estimated 21.7% of the fish larger than 200 mm in the Gunnison River
above the barrier were tagged. A total of 23,031 trout from 68mm to 737mm were tagged in the 20132014 and released in the Gunnison River above the barrier.
Mark recapture population estimates in the study reach were conducted in October 2011, October 2013,
July 2014 and October 2014 with a 14 ft aluminum jet boat with Smith Root 2.5 GPP electrofisher. The
study reach, equipment and methods for all occasions were the same. Fish were measured to the nearest
4
�millimeter and all fish on the recapture pass were weighed to the nearest gram. All captured fish were
examined for fin clips and checked for coded wire tags with a Norwest Marine Technology T-Wand
Detector and for PIT tags with an Oregon RFID handheld reader. On the marking pass all fish greater than
150 mm were marked with a caudal fin punch and held in a live pen to ensure recovery. Fish were
returned by boat throughout the study reach to ensure redistribution in the population. The recapture
pass was completed 72 hours after the marking pass and generally accepted methods were followed for
mark recapture studies (Curry et al. 2009). The interval between capture events was chosen to maximize
redistribution of marked fish throughout the population but to attempt to meet demographic and
geographic closure assumptions of the model. The first power plant served as an upstream migration
barrier further ensured geographic closure; block nets downstream were not feasible to the high volume
of water in the canal (600-900 cfs). Model assumptions appear to have met well as marked fish were not
observed to be encountered in any temporal or spatial pattern in the canal. Capture probabilities were
good and the catch per unit effort of fish was similar between the passes.
A stationary PIT tag antenna was constructed above the penstock of the power plant but below the barrier
in the spring of 2014. The objective was to differentiate fish deterred by the barrier and turbine mortality
as well as increase detection of tagged fish. The antenna was operational for less than two months as the
extreme velocities of the water (900 cfs in a 10.5 ft. wide concrete channel) made it impossible to keep in
place. No tags other than test tags were detected by the antenna.
Fish population estimates were made with the Huggins Closed Capture Model in Program Mark (Huggins
1989, White and Burnham 1999). Four models were built by estimating capture probabilities by length,
species, species + length, as well as a constant capture probability for all fish, identical to a Lincoln
Petersen model (Seber 1982). Model selection was done with AICc and population and parameter
estimates were made by model averaging across all four models with AICc weights (Burnham and
Anderson 2002).
Figure 3. The electric fish barrier on the east portal of the South Canal.
5
�RESULTS
The results of the population estimates are summarized in Table 1 and Figure 4. Length frequency
histograms from the fall 2013 sampling are presented in Figures 5-7. Model selection results from are
summarized in Appendix A, Tables 2-6.
The population modeling exercise in Program Mark provided good results and estimates appeared
accurate (all years) and relatively precise (except for October 2011). The expected bias of population
estimates should be low due to model assumptions being met, and the ability to model the size
selectivity of electrofishing with fish length covariates. The top population model for the October 2011
data contained terms that varied capture probability by length and time while the second ranked model
that contained terms for species, length and time was 2.40 ΔAICc units behind. Models with a term for
fish length contained 0.98% of the model weights. Capture probabilities were lower during this survey
(0.10) compared to subsequent surveys due to the higher flows and lower total number of fish captured.
In March 2013, the top population model for the canal and the stilling basin had a single capture
probability for all fish regardless of species or length while the second ranked model contained a term
for species. These two models are essentially identical to the simple Zippin two pass removal model and
had 73.2% of the model weight (Zippin 1956). Although it has been shown that electrofishing surveys
generally have a size related bias, this effect was not seen in these data because of how few fish were in
the canal outside of the stilling basin and there was little variation in fish size compared to the fall
surveys. Because of the low density of fish, moderate capture probabilities and similar sized fish, the
data from the canal were too sparse to support more detailed models. There was no evidence of size
selectivity in the stilling basin with the small mesh seine.
The top population model for the October 2013 data contained terms that varied capture probability by
length, species and time while the second ranked model that contained terms for length and time.
These two models accounted for 100% of the model weights and had much higher support than a simple
Lincoln-Petersen (19-27 ΔAICc units behind). Capture probabilities were high (0.33) due to the lower
flow conditions than 2011. Model selection uncertainty was taken into account in all surveys by model
averaging across all four models with model weights to get parameter estimates and population
estimates.
The significant increase (95% level) of total fish in the canal from April 2013 to October 2013 is evidence
for fish successfully running the barrier and surviving the turbines. After the March estimate, when 876
fish were removed from the canal, the population estimate increased by 1,057 fish by October. The
total number of estimated fish was significantly greater (at the 95% level) in October than in April and
was not significantly different than in the October 2011, before the barrier. The size structure and
species composition of the fish in the 2013 also provide evidence of fish entrainment, specifically for
brown trout (Figures 5, 6 and 7).
6
�Table 1. Fish Population Estimates and 95% Confidence Intervals from the South Canal 2011-2014.
These estimates are for age 1 fish and older, the stocked CWT tagged rainbows are excluded from the
rainbow trout estimates.
Date
October 2011
March 2013
October 2013
# Caught
Brown Trout
415
2,359±981
Rainbow Trout
108
634±354
Brown Trout
683
924±52
Rainbow Trout
495
659±46
Brown Trout
573
1,035±150
Rainbow Trout
277
728±235
Stocked CWT Rainbow
246
1,486±768
2
NA
0
NA
Brown Trout
225
586±52
Rainbow Trout
132
638±469
Stocked CWT Rainbow
25
NA
0
NA
0
NA
Brown Trout
305
964±258
Rainbow Trout
277
936±278
Stocked CWT Rainbow
15
NA
0
NA
4
NA
CWT, Adipose Clipped
Brown
CWT, Adipose Clipped
Rainbow
July 2014
CWT, Adipose Clipped
Brown
CWT, Adipose Clipped
Rainbow
Oct 2014
Population Estimate in
Species
CWT, Adipose Clipped
Brown
CWT, Adipose Clipped
Rainbow
7
Study Reach
�4,500
4,000
Number of Fish
3,500
3,000
2,500
2,000
1,500
1,000
500
0
Oct 2011
Mar 2013
Apr 2013
Oct 2013
Jul 2014
Oct 2014
Figure 4. Estimated total number of trout age 1 and older and 95% confidence intervals in the South
Canal study reach. After the March 2013 estimate, 876 fish were removed from the canal study reach
and the barrier was operational at the start of the irrigation season in April 2013. There are about 1,094
fewer fish in the study reach since the barrier was installed but the decline is not significant at the 95%
level, mostly because of the low capture probability and corresponding high uncertainty around the
October 2011 estimate.
Number Sampled
70
60
CWT Rainbow
50
Brown Trout
40
Rainbow Trout
30
20
10
0
5
45
85
125
165 205 245
Length (mm)
285
325
365
405
445
485
525
Figure 5. Length frequency histogram of trout captured in the South Canal in October 2013. A total of
246 coded wire tagged rainbows were captured that had been stocked upstream of the guidance system
(plus 10 recaptures). They had a mean length of 163 mm (123-204). Two other coded wire tagged fish
were captured, a 310 mm brown and 337 mm brown (the 310 mm fish was also recaptured). No tag loss
was observed, all of the larger fish were double marked and no fish were observed with an adipose clip
but without a CWT.
8
�Figure 6. Length frequency histogram of brown trout captured in March and October 2013.
Figure 7. Length frequency histogram of rainbow trout captured in March and October 2013.
9
�The top population models for the July and October 2014 data contained terms that varied capture
probability by length, species and time. The top two models that included length accounted for 100% of
the model weights. Capture probabilities were good in July (0.11-0.32) and (0.17-0.20) October. Model
selection uncertainty was taken into account in all surveys by model averaging across all four models
with model weights to get parameter estimates and population estimates.
The population modeling exercise for all the mark recapture data indicated that modeling capture
probabilities by length was important under these conditions which agrees with previous work on the
topic (Saunders et. al 2011). Using a simple Lincoln-Petersen model under these conditions could
underestimate population size by overestimating the capture probability for small fish, even when using
a length cutoff designed to exclude age 0 fish. Figure 8 shows an example of the estimated capture
probability by length and Figure 9 shows a comparison of population estimates with and without the
length covariate.
In October 2011, there were an estimated 2,994±1,043 fish in the South Canal study reach. In the spring
of 2013 there were an estimated 1,583±70 in the study reach, 89% in the stilling basin. Eight hundred
and seventy-six of these fish were removed from the study reach leaving an estimated 707 fish when the
irrigation flows first began in the spring of 2013. In October 2013 the estimated population had
increased to 1,764±279 trout. The population estimate of total fish in the study reach decreased from
October 2011 to 2013 but that difference was not significant at the 95% level, likely due to the
uncertainty around the 2011 estimate caused by lower capture probability likely due to higher flows.
Subsequent sampling occasions had much higher capture probabilities generally in the 20% range (0.110.33). In 2014 the study reach contained 1,224±239 fish in July and 1,900±379 in October.
In 2013, a total of 248 coded wire tagged fish from 123 mm to 337 mm were documented passing
through the barrier, mostly smaller stocked rainbow trout (n=246 mean length 163 mm in October).
Only two larger wild brown trout were confirmed passing the barrier (310 and 337 mm). Of the tagged
fish that were documented to have run the barrier in 2013, the stocked rainbows represent 1.24% of the
fish marked in East Portal and the wild brown trout were 0.3%. Overall, 1.17% of all the tagged fish in
East Portal were captured in the study reach in 2013. In the 0.72-mile study reach there was an
estimated 1,486±768 coded wire tagged rainbows or 7.5% of the tagged fish in East Portal.
These results do not represent a direct estimate of entrainment rates as only 9.4% of the total length of
the canal was sampled at a single time interval. Rather, the results represent the number of entrained
fish in the study area that were detected. It should be interpreted as a minimum number of fish that
navigated the barrier because fish would have to pass the through the guidance system, travel the 5.7mile-long tunnel, avoid entrainment in two small lateral canals, survive passage through the hydropower
turbines and remain in the first 0.72 miles of the 7.7-mile canal to be detected. If the density of fish in
the study reach is representative of the rest of the canal, then an estimated 15,809 coded wire tagged
rainbow fingerlings would have been entrained or 79.8% of those fish and 3.2% of the larger marked
wild brown trout in 2013. This is most likely an over estimate of entrained fish because the study reach
could have a higher density of fish than the other reaches of the canal, but it demonstrates the same
trend of high entrainment rates for small fish and relatively low rates for larger fish and can be
interpreted as a potential maximum number. While estimating robust entrainment rates with the
barrier is not possible in this study, the true rates are likely between these minimum and maximum
values; 8-80% for small rainbow trout and 0.3-3.2% for larger brown trout.
In 2014, a total of 44 tagged fish were encountered, 40 of the hatchery rainbows (mean length 326 mm
10
�at the time of capture). Four CWT and fin clipped wild rainbow trout (296-398 mm) were found. It is
unknown exactly when or what size all the tagged fish in 2014 passed the barrier because fish lived and
grew in the canal throughout the study. The 2013 data give the best idea of size of fish that ran the
barrier because they were in the canal for a maximum of seven months. The large number of CWT
tagged rainbows could have passed the barrier as small as 68 mm and then survived to be captured at a
larger size.
By the end of the study 288 small or medium sized fish had been documented passing the barrier. Only
four fish >300 mm, and no fish >400 mm were documented passing the guidance system. Only 1.3% of
all tagged fish were recovered in the canal study reach in two years. While turbine mortality and fish
excluded from the study reach by the trash racks on the penstock cannot be differentiated from fish
excluded by the barrier, very few large fish have been observed passing these barriers. The number of
large fish (>350mm) in the study is not significantly different after the barrier (Figure 10) even though
there is no evidence that fish of that size are passing through in great numbers. Large numbers of
smaller fish have been shown to run the barrier as evidenced by both the number of marked fish and
the stable trout population in the canal even after the barrier was in use. The lack of a decline in fish
populations in the canal after the barrier is likely related to higher than expected survival and growth of
small fish entrained in the canal.
In July 2014, 17% of the fish captured during the population estimate (37% greater than 350 mm) had
been handled the previous October by the presence of a healed caudal punch scar. This indicates that
there is fair to good over winter survival in the canal. Growth of fish that live in the study reach is also
relatively high; coded wire tagged rainbows grew an average of 6.4 inches from age 1 to age 2. With the
good annual survival and growth rates, the large numbers of smaller fish that pass the barrier maintain a
relatively stable population of fish in the study reach, even though large fish do appear to be excluded
from the canal by the electric barrier.
Figure 8. Estimated capture probability by length and 95% confidence interval (dashed lines) for trout in
the South Canal in October 2014.
11
�Figure 9. Brown trout population estimates from the Huggins Closed Capture model in Program Mark
comparing models with a fish length covariate to a standard Lincoln-Petersen. The estimates that used
length to model capture probabilities were on aver 23% higher (6-41%) than the Lincoln-Petersen.
Models containing length as a covariate had between 98-100% of the model weight across all mark
recapture sampling occasions.
Figure 10. Estimated total number of trout greater than 350 mm in the South Canal study reach in the
October sampling periods. While very few (4) fish greater than 300 mm have been documented passing
the barrier and turbines, growth and survival of smaller entrained fish supports a stable number of
larger fish in the study reach.
12
�DISCUSSION AND MANAGEMENT RECOMMENDATIONS
The South Canal contained approximately 1,094 fewer fish in October 2014 after the barrier, than in
October 2011. While the total fish estimates in the canal have declined since the barrier was installed,
there is not a significant difference at the 95% level mostly due to the low capture probability (0.8-0.12)
and corresponding high uncertainty around the October 2011 estimate. The number of brown trout
only is significantly lower at the 95% level in 2014 (two years after the barrier was installed) while the
number of rainbow trout has remained relatively stable (Figure 11).
4,000
Brown Trout
Population Estimate
3,500
Rainbow Trout
3,000
2,500
2,000
1,500
1,000
500
0
Oct 2011
Mar 2013
Oct 2013
Jul 2014
Oct 2014
Figure 11. Population estimates of rainbow and brown trout and 95% confidence intervals for the South
Canal Study reach 2011-2014. There were signifcantly fewer brown trout in 2014, two years after the
installation of barrier. Rainbow trout numbers have raimained relatively stable.
Of the 23,031 tagged fish, only 1.3% were recovered in the canal study reach in two years. At the end of
the study, 288 small or medium sized fish had been documented passing the barrier. Four fish greater
than 300 mm and no fish greater than 400 mm were documented passing the barrier. This size
selectivity is expected with electrically based barriers and electrofishing is known to be highly size
selective as well (Saunders et. al 2011). It is also likely that turbine mortality is higher on larger fish,
further selecting for smaller fish to make it into the study reach. The growth and survival of fish in the
canal is higher than expected as evidenced by the high proportion of recaptured fish from October 2013
to July 2014. The practice of running 100 cfs into the canal twice a month in the winter and relatively
mild recent winters apparently allowed for good fish survival during the winter of 2013-2014.
Fish in the Gunnison River are successfully passing the electric barrier and surviving the turbines, but
mostly smaller fish. Their growth and survival in the canal maintains a stable fish population that is
lower than before the barrier, significantly (at the 95% level) for brown trout only. The difference
between species is likely due to two factors; larger size of age 0 brown trout and potential spawning of
rainbow trout in the study reach. Because brown trout emerge about 8-10 weeks earlier than rainbow
13
�trout they are larger during their first summer. Because the barrier is size selective, brown trout fry are
expected to be entrained at a lower rate than rainbows. The canal is first filled with water around April
1st of each year, just before rainbow trout spawn. Large numbers of age 0 rainbow trout were observed
in the canal in July 2014 (they were smaller than the 150 mm size cut off used in the fish population
estimates). It is unknown if they were entrained fish from the Gunnison River or were spawned in the
canal, both are likely. Brown trout spawn in October in the Gunnison River and flows are generally shut
off in the canal around October 31. Water flow is then stagnant or 100 cfs (twice a month for 24 hours)
in the canal in winter. There is very little spawning habitat for brown trout in the canal and it is variable
and poor quality compared to rainbow trout, which spawn at higher flows that are stable or increasing.
A combination of higher entrainment rates and better potential spawning success in the canal likely
leads to higher number of small rainbow trout in the canal.
If the barrier is successfully excluding many of the fish greater than 300 mm and most of the fish greater
than 400 mm, then it is excluding approximately 15% of the trout greater than 150 mm and 26-71% of
sexually mature fish based on 2013 data on the Gunnison River. Low numbers of age 2 fish in the
Gunnison are sexually mature (mostly males) while most age 3 fish are mature (E. Gardunio, Colorado
Parks and Wildlife, unpublished data). So while the barrier is generally meeting its stated objective, it’s
not protecting all of the sexually mature fish. Excluding higher proportion of small trout from
downstream passage is likely to be difficult and will be dependent on several factors including the
voltage gradient of the barrier and the approach velocities of the water at the barrier. Excluding a larger
proportion of adult fish than is currently occurring is a more reasonable expectation for the East Portal
barrier with some operational changes.
As approach velocities increase above 2.5 fps the probability of excluding small salmonids with an
electric barrier decreases (Demko et al. 1994, Pugh et al. 1970). The approach velocities at the South
Canal barrier varied between 0.7 and 2.3 fps in October 2011 when the tunnel was flowing about 730 cfs
and the river below the tunnel was about 580 cfs. Under those flows, better deterrence of small trout
should be possible with operational adjustments, but more work is needed to determine approach
velocities at various flows. The field strength of the South Canal barrier is currently about 1v/inch or
0.4v/cm (Smith Root 2011) which is relatively conservative compared to other barriers designs. Most
other downstream oriented barriers are graduate-field fish barriers (GFFB) where several rows of
electrodes produce increasing voltage gradients between 0.2-1.2 v/cm while other designs have utilized
voltage gradients as high as 3.0 v/cm (Raymond 1956, Burger et al. 2015). The GFFB technology appears
more effective in deterring downstream movement of fish but was not applied at the South Canal due to
site specific conditions. Diverting downstream moving fish is one of the more difficult applications of
electric fish barriers (Burger et al. 2012). Achieving complete deterrence of all fish is unlikely in
scenarios like East Portal. The objective there should be to reduce the amount of entrainment much as
feasible within the constraints of the system. More work is necessary to determine if increasing the
voltage gradient, or other operational changes at the East Portal barrier could improve performance on
smaller fish.
The electric fish barrier on the South Canal of the Gunnison River appears to effectively exclude large
fish from the south canal, resulting in fewer entrained fish from the river. Fish populations in the South
canal, while lower than before the barrier, appear stable due to the number of entrained smaller fish,
potential spawning of rainbow trout and better than expected growth and survival of fish in the canal.
The electric barrier on the South Canal should continue to be operated whenever feasible during the
irrigation season and future study is needed to examine if operational changes of the current barrier can
increase the probability of excluding more adult fish.
14
�REFERENCES
Bureau of Reclamation. 2012. Final Environmental Assessment South Canal Hydropower
Project. Western Colorado Area Office Upper Colorado Region, U.S. Department of the Interior.
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a
practical information-theoretic approach, 2nd edition. Springer-Verlag, New York.
Burger, C.V, J.W. Parkin, M. O’Farrell and A. Murphy. 2015. Barrier technology helps deter fish at hydro
facilities. Hydo Review 34(5).
Burger, C.V, J.W. Parkin, M. O’Farrell, A. Murphy, J. Zelgis. 2012. Non-lethal electric guidance barriers
for fish and marine mammal deterrence: a review for hydropower and other applications.
HydroVision Brazil, September 26, 2012.
Carlson, A.J. and F.J. Rahel. 2007. A basinwide perspective on entrainment of fish in irrigation canals,
Transactions of the American Fisheries Society, 136:1335-1343.
Curry, R.A., R.M. Hughes, M.E. McMaster, and D.J. Zaft. 2009. Coldwater fish in rivers.
Pages 139-158 in S.A. Bonar, W.A. Hubert, and D.W. Willis, editors. Standard methods for
sampling North American freshwater fishes. American Fisheries Society, Bethesda, Maryland.
Demko, D.B, S.P. Cramer, D. Neeley and E.S. Van Dyke. 1994. Evaluation of a sound and electrical fish
guidance system at the Wilkins Slough diversion operated by Reclamation District 108. Annual
Report S.P. Cramer & Associates Inc., Gresham OR.
Huggins, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika
76:133–140.
Pugh, J.R., G.E. Monan and J.R. Smith. 1970. Effect of water velocity on the fish-guiding efficiency of
and electrical guiding system. Fishery Bulletin 68 (2):307-324.
Roberts, J.J. and F.J. Rahel. 2008. Irrigation Canals as Sink Habitat for Trout and Other Fishes in a
Wyoming Drainage. Transactions of the American Fisheries Society 137:951-961.
Riley, S. C., and K. D. Fausch. 1992. Underestimation of trout population size by maximum
likelihood removal estimates in small streams. North American Journal of Fisheries
Management 12:768–776.
Saunders W.C., K.D. Fausch, and G.C. White. 2011. Accurate estimation of salmonid
abundance in small streams using nighttime removal electrofishing: an evaluation using marked
fish. North American Journal of Fisheries Management 31:403-415.
Seber, G. A. 1982. The estimation of animal abundance and related parameters, Second edition.
Charles Griffin and Company, Ltd, London.
Smith Root. 2011. Electric Fish Barrier Design, Hardware Supply, Commissioning and Training Gunnison
Tunnel, CO. 16 pp.
15
�Raymond, H.L. 1956. Effect of pulse frequency and duration in guiding salmon fingerlings by electricity.
U.S. Fish and Wildlife Service Research Report 43, Washington D.C.
White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capture-recapture and
removal methods for sampling closed populations. Los Alamos National Laboratory, Report LA8787-NERP, Los Alamos, New Mexico.
White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations
of marked animals. Bird Study 46(Supplement): 120–139.
Zippin, C. 1956. The removal method of population estimation. Journal of Wildlife
Management 22:82-90.
16
�Appendix A. Model Selection Results for Population Estimation Models
Table 2. Model Selection Results for the Mark Recapture Electrofishing in October 2011. Population
estimates and capture probabilities were calculated by model averaging across all four models using
model weights. The “Time” and “Time+Species” models are identical to the standard Lincoln Petersen
model.
Model
AICc
Number of
Delta
AICc
Model
Parameters
AICc
Weights
Likelihood
Time+Length
893.0038
3
0
0.75
1.00
Time+Species+Length
895.4048
5
2.40
0.23
0.30
Time
900.3091
2
7.31
0.02
0.03
Time+Species
902.7106
4
9.71
0.01
0.01
Table 3. Model Selection results for the Two Pass Removal Electrofishing in March 2013. Population
estimates and capture probabilities were calculated by model averaging across all four models using
model weights.
Model
AICc
Number of
Delta
AICc
Model
Parameters
AICc
Weights
Likelihood
Constant p
117.40
1
0
0.528
1.00
Species
119.30
2
1.90
0.204
0.39
Length
119.39
2
2.00
0.195
0.37
Length+Species
121.34
3
3.94
0.073
0.14
Table 4. Model Selection Results for the Mark Recapture Electrofishing in October 2013. Population
estimates and capture probabilities were calculated by model averaging across all four models using
model weights. The “Time” and “Time+Species” models are identical to the standard Lincoln Petersen
model.
Model
AICc
Number of
Delta
AICc
Model
Parameters
AICc
Weights
Likelihood
Time+Species+Length
1760.461
5
0
0.77
1.00
Time+Length
1762.837
3
2.38
0.23
0.30
Time+Species
1779.036
4
18.58
0.00
0.00
Time
1787.185
2
26.72
0.00
0.00
17
�Table 5. Model Selection Results for the Mark Recapture Electrofishing in July 2014. Population
estimates and capture probabilities were calculated by model averaging across all four models using
model weights. The “Time” and “Time+Species” models are identical to the standard Lincoln Petersen
model.
Model
AICc
Number of
Parameters
Delta
AICc
AICc
Weights
Model
Likelihood
Time+Species+Length
663.08
5
0
0.802
1
Time+Length
665.88
3
2.80
0.198
0.2469
Time+Species
685.91
4
22.83
0.000
0
Time
693.13
2
30.05
0.000
0
Table 6. Model Selection Results for the Mark Recapture Electrofishing in October 2014. Population
estimates and capture probabilities were calculated by model averaging across all four models using
model weights. The “Time” and “Time+Species” models are identical to the standard Lincoln Petersen
model.
AICc
Number of
Parameters
Delta
AICc
AICc
Weights
Model
Likelihood
Time+Length
1106.96
3
0
0.876
1.000
Time+Species+Length
1110.87
5
3.9097
0.124
0.142
Time
1122.71
2
15.7451
0.000
0.000
Time+Species
1126.72
4
19.7574
0.000
0.000
Model
18
�
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Reports
Text
A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Evaluation of an electric fish barrier on the South Canal, an irrigation ditch on the lower Gunnison River, Colorado
Description
An account of the resource
An electric fish barrier was installed on the east portal of South Canal to reduce fish entrainment associated with the construction of two hydropower plants in 2012. The objective of this study was to monitor fish entrainment and evaluate the effectiveness of the barrier. Three groups of fish were tagged and released upstream of the barrier; fish from the canal, wild Gunnison River fish, and hatchery reared fingerlings. Mark recapture boat electrofishing was completed and population estimates were made with the Huggins Closed Capture model using fish length to model capture probabilities. The study reach contained 2,994 ± 1,043 fish (>150 mm) in October 2011, 1,764 ± 279 in October 2013, 1,224 ± 239 in July 2014 and 1,900 ± 379 in October 2014. Fish population estimates have declined after the electric barrier, significantly at the 95% level for brown trout but not for rainbows. A total of 288 tagged fish less than 300 mm and four fish greater than 300 mm were recovered below the barrier, representing 1.3% of all tagged fish. The electric barrier appears to meet its objective and successfully exclude larger fish from the study reach, but not smaller age 0, age 1, or age 2 trout. The entrainment, growth and survival of smaller fish maintains a stable population of fish in the canal, but fewer entrained mature fish is likely a benefit to the fish population of the Gunnison River. Further study is needed to evaluate if smaller adult trout can be successfully excluded by the electric barrier with operational modifications.
Creator
An entity primarily responsible for making the resource
Kowalski, Dan
Subject
The topic of the resource
Sport fish
Fishing
South canal
Research
Extent
The size or duration of the resource.
18 pages
Date Created
Date of creation of the resource.
2012
Rights
Information about rights held in and over the resource
<a href="http://rightsstatements.org/vocab/NoC-NC/1.0/" target="_blank" rel="noreferrer noopener">No Copyright - Non-Commercial Use Only</a>
Type
The nature or genre of the resource
Text
Format
The file format, physical medium, or dimensions of the resource
application/pdf
Language
A language of the resource
English