Abstract
In June 2008, operation of gates in the
Golden Lotus Dam resulted in the release of large amounts of organic sediment
into the Pigeon River downstream of the dam in Otsego and Cheboygan counties,
MI. Records from the U. S. Geological
Survey (USGS) gage located about 1.2 miles downstream from the dam at Sturgeon
Valley Road show that on the afternoon of June 22, 2008 stream discharge rose
rapidly from around 40 cubic feet per second (cfs) and peaked at about 170 cfs
just after midnight. Then on the
morning of June 23, 2008 estimated stream discharge plummeted rapidly to around
7 cfs at 7:30 am. Discharge was
restored to near the median level of 60 cfs after 6:00 pm on June 23rd.
Electrofishing surveys and dead fish collections made
within the first week after the initial event showed that a large fish kill
composed of at least 16 species of fish had occurred. During the remainder of June and the month of July trout
population estimates were made in 12 sections of the river between the dam and
Pigeon River Road Bridge. Michigan
Department of Natural Resources (MDNR) determined that the kill zone extended
at least to M-68 Bridge, which is located nearly 25 miles downstream of the
dam.
MDNR estimated that the event in June 2008 resulted
in the deaths of 474,486 trout in the Pigeon River between the dam and highway
M-68 Bridge. Numerical losses were
highest for young-of-year trout, particularly rainbow trout. Nearly 25,000 yearling-and-older trout were
killed by the event and over 2,000 of these fish were 12 inches or longer. MDNR could not quantitatively estimate the
numbers of fish other than trout that were killed but it is certain that many
thousands were lost.
Low populations of trout in the Pigeon River in 2008
after the event were not due to some natural phenomenon. Overall density of brook trout and brown
trout in the Pigeon River in the 1970s and 1980 were similar to expected
densities for contemporary trout streams in the Northern Lower Peninsula but
their density was far lower than normal after the 2008 fish kill.
Our economic valuation of this fish kill makes use
of contemporary and historical data collected from the Pigeon River, data from
published economic literature, and the American Fisheries Society (AFS)
recommended guidelines for determining costs associated with a fish kill like
that being reported on here. Typically,
monetary damages from fish kills consist of replacement costs, lost user value,
and lost nonuse value. However
valuation of nonuse costs in this case could not be reliably quantified and are
not included in this report.
Undoubtedly, MDNR’s economic valuation of this fish kill is an
underestimate of the true cost to the public by virtue of no reliable
information on nonuse costs and having to rely on average values published for
angler use and expenditures in Michigan.
Replacement costs for the trout killed downstream of the Golden Lotus
Dam as a result of the event in June 2008, costs associated with investigating
and managing the kill, and compensation for the recreational opportunity lost
by anglers total $668,236. It should be
noted however that investigation and management costs included here are only
through March 2009 and additional costs will continue to accrue.
Continued operation of the Golden Lotus Dam as it
has been operated for the past several decades poses a growing risk of more
frequent fish kills because the impoundment has nearly filled with
sediment. Water storage capacity of the
impoundment continues to decrease as sediment volume increases. Thus over time, the continued existence and
operation of dam poses an ever higher risk that more sediment will be
mobilizing and released into the river downstream. The 2008 fish kill was the result of human error and mechanical
failure that occurred over only a 6-hour period. The only way to eliminate the risk of future kills and restore
the Pigeon River to its full productive potential is to completely remove the
dam to restore a natural flow regime, reduce summer water temperatures, and
restore free fish passage at the dam site.
Introduction
The Pigeon River, located in Otsego and Cheboygan
counties Michigan, has only one dam located on the main stem of the river. The Golden Lotus Dam, located at T32N R1W
Section 19 is listed in the Michigan Department of Environmental Quality (MDEQ)
database as having a head of 13 feet, a pond area of 45 acres, and a storage
capacity of 565 acre feet of water (MDEQ Land and Water Management
Division). The present dam was rebuilt
after the original dam washed out on May 15, 1957. The catastrophic failure in 1957 occurred after a floating log
jammed the spillway gate, which ultimately led to erosion and sudden breaching
of the earth fill dam and spillway (Waters 1960). This dam failure and flood resulted in substantial damage to
aquatic habitat including destruction of natural and man-made fish cover and
deposition of organic and inorganic sediments from the impoundment (Waters
1960). In July 1984 while the gates of
the dam were being repaired an intentional release of an abnormal amount of
organic material and sediment from the dam resulted in a large fish kill in the
river downstream. Alexander and Ryckman
(1986) estimated that the release killed 67% of legal-sized trout (≥8.0
inches) and 52% of sublegal trout in the first 6 miles of river downstream of
the dam. They estimated that overall
stocks of trout further downstream were 30% lower than normal. The present report evaluates the effects of
a third event that occurred in June 2008.
Recreational trout angling across the Great Lakes
basin is known to generate significant economic activity. For example, the average recreational trout
angler living and fishing within a geographic region covering parts of
southwest Wisconsin, southeast Minnesota, northeast Iowa, and northwest
Illinois spends about $200 on various expenses on each outing (NorthStar Economics,
Inc. 2008). The average angler spends
nearly twice as much per trip if he travels from outside the region, and it’s
estimated that trout anglers contribute more than $1 billion per year to the
regional economy (NorthStar Economics, Inc. 2008). There is little doubt that the Pigeon River in Otsego and
Cheboygan counties, MI. had been a popular trout fishing destination for
anglers up until the event that occurred in June 2008. It is difficult to estimate how long it will
be until the Pigeon River becomes such a place again, but it is fair to expect
it will be years before the fishery returns to its former status. An earlier study of the effect on trout
populations and angler use from a similar event at this same dam (Alexander and
Ryckman 1986) reported it would take a minimum of 3 to 4 years before the
abundance of large trout would provide the high quality trout fishing
opportunities enjoyed by anglers prior to the sediment release; even after
extensive stocking of replacement fish had occurred. It is also reasonable to assume that both then and now, anglers
placed value on the fishery resources of the Pigeon River that goes well beyond
replacing the trout killed by this recent event, that trout anglers traveling
to the area contribute significant dollars to the local and regional economies
each year, and that trout anglers visiting the Pigeon River Country enjoy more
than just fishing for trout when they visit there. Unfortunately, estimating the full economic losses of
recreational values compromised by operations of Golden Lotus Dam can not be
reliably accomplished and we are limited here to accounting for the replacement
costs of trout killed by the event in June 2008, the cost of our investigation
of this fish kill, and costs associated with lost angling opportunities after
the event killed approximately 80% of the trout in nearly 25 miles of one of
the state's trout streams.
Objectives of this analysis and report were to: 1)
Determine the distance the 2008 fish kill extended down from the dam, 2)
Estimate the numbers trout killed, 3) Estimate the species composition and size
ranges of non-trout species killed, 4) Evaluate whether trout escaped to
refugia in tributaries or downstream, 5) Compare trout density in the Pigeon
River in the past and in summer 2008 to contemporary densities in other trout
streams in the Northern Lower Peninsula, 6) Make an economic valuation of the
value of trout killed along with costs of investigation and losses of
recreational opportunity resulting from the fish kill in June 2008.
Methods
Determination of the downstream extent
of the fish kill
MDNR
used one-pass electrofishing surveys to make initial estimates of how far the
fish kill extended downstream from the dam.
Electrofishing, simply defined as the use of electricity to capture
fish, was used to collect samples of live fish from the Pigeon River at seven
sites on the mainstem river between the dam and the Agnes Andreae Nature
Preserve, located 27.7 miles down from the dam, within the first 10 days after the
event (Figure 1). Additional data on survey dates, lengths of
stations surveyed and their locations is given in Table 1.
MDNR
used direct current electrofishing gear because direct current produces a field
in which fish typically exhibit forced swimming toward anodes held by wading
electrofishers, who then net the fish with a dip net so that they can be
identified and measured. Thus, trout
can be effectively captured even when waters are turbid or if they are hidden
in cover such as log jams because they can be drawn towards the surface where
they are visible.
Trout
catches during one-pass electrofishing surveys in 2008 were analyzed by
comparing them to catches in previous surveys at or near the same sites. If catches were substantially lower in 2008
than during previous surveys MDNR interpreted this as evidence that the fish
kill had extended at least as far downstream as the site sampled.
MDNR
further refined their estimate of the downstream extent of the kill by
comparing mark-and-recapture populations estimates made in July 2008 to
estimates made in the same stream sections in prior years. Locations where population estimates were
conducted in 2008 are shown in Figure 2, and more detailed information on the
surveyed reaches is given in Table 2.
Estimation of numbers of brook and brown trout killed
by the event
Estimates of numbers of brook and brown trout
killed by the 2008 event were computed by subtracting population estimates made
in 2008 from population estimates made in previous years in the same sections
of the river. These previous years when
populations were estimated can be referred to as reference periods or reference
years.
MDNR
used averages of trout population estimates conducted in 1975, 1976, 1979, and
1980 as the reference data for computing trout mortality and losses in the
first 6 miles of river immediately downstream of the dam. Electrofishing sampling was conducted
throughout the entire 6-mile reach below the dam in 1979 and 1980 (Sections A
through E) (Figure 2). In 1975 and 1976
electrofishing sampling was conducted in about 1/3 of Sections A, C, D, and E
(Figure 2). These estimates were
standardized to numbers per mile so that all 4 years of data could be used to
compute a reference abundance level for these four sections. Populations were not estimated in Section B
in 1975-76 so MDNR used average abundance in 1979-80 as the reference period
for Section B. More recent population
data collected from the entire 6-mile reach in 1984 and 1985 were not usable
for the analyses because trout populations were depressed from a previous fish
kill caused by operation of the Golden Lotus Dam in July 1984.
MDNR
estimated trout killed in the 9.8-mile section of the Pigeon River between the
downstream end of Section A and McIntosh Landing based on population data
collected at four locations within this reach: Tin Bridge, Trail 34-Elk
Meadows, Pine Grove Campground, and Webb Road (Figure 2). This approach allowed MDNR to account for
spatial variability in trout abundance among sections within the 9.38-mile
reach. Total trout abundance for this
section in 2008 was then computed by multiplying average abundance per mile at
the four survey stations by 9.38.
Collectively, the four sampling stations were 1.52 miles long so the
expansion of the average density to the entire reach is representative of
abundance in the entire section. MDNR
estimated the percentage of trout killed in this reach by comparing population
data collected at Webb Road in 1976 and 2008.
There were no other pre-2008 population data available for this reach
except for data collected several weeks after the fish kill documented in 1984.
Five
brown trout captured at Trail 34-Elk Meadows in late July 2008 that were
obviously recent immigrants (lake-run fish from Mullett Lake) were excluded
from the analysis. These fish could be
easily distinguished from resident fish because immigrants have a silvery
coloration and robust morphology.
MDNR
used a very similar approach to estimating trout killed in the 9.5-mile reach
of the river between the upstream end of McIntosh Landing and M-68 Bridge. In this instance, comparable population data
from 1976 and 2008 were available at two sites; McIntosh Landing and Pigeon
River Road Bridge (Figure 2).
Percentage killed in 2008 was determined by comparing average abundance
in 2008 to average abundance in 1976 for these two stations. Average abundance of trout per mile in 2008
for the section was computed as the average of density at three stations in the
reach, McIntosh Landing, the Cudlip Property, and Pigeon River Road Bridge
(Figure 2). Total numbers in the reach
were obtained by multiplying by 9.5.
Estimates of trout killed were made by subtracting observed abundance in
2008 from expected abundance, which is based on the reference year of 1976.
Estimation of numbers of rainbow trout killed by the
event
Estimates
of mortality of rainbow trout were made using a different method because
historically rainbow trout were very rare in the Pigeon River. The methods used for brook trout and brown
trout would have resulted in severe underestimates of numbers of rainbow trout
killed because more of them survived the 2008 event than were present
historically. Therefore, to estimate
rainbow trout mortality MDNR first determined what percentage of brook trout
and brown trout were killed in 2008 in each river section and made the
assumption that the same percentage of rainbow trout would also have been
killed. This assumption is based on the
fact that all three species have similar intolerance to sediment or low
dissolved oxygen levels. MDNR next
computed the number of rainbow trout expected to be present in 2008 based on
the population estimates made in 2008 and expansion factors derived from the
percentage of brook trout and brown trout killed in each river section. For example, if 90% of brook trout and brown
trout in a section were killed (i.e. 9 out of 10) and if our 2008 rainbow trout
estimate was 10 fish then the expected number of rainbow trout in 2008 would be
100 fish (expansion factor of 10) and the number killed would be 90 fish (i.e.
100 – 10).
Estimates of non-trout species killed by the release
Three
300-foot sections of the Pigeon River were searched for dead fish of all
species to obtain information on the species composition and size distribution
of fish killed (Figure 3). On June 24,
2008 a survey crew conducted a search for visible dead fish in a 300-foot
section of the Pigeon River centered on Sturgeon Valley Road Bridge. Two additional searches for all visible dead
fish were conducted in 300-foot stream reaches on June 25, 2008. One reach was located upstream of Sturgeon
Valley Road, about 0.6 miles down from the dam, and the other was located about
2.2 miles downstream of the dam. All
species of fish observed were collected, sorted by species, and each individual
measured to the inch class. Fish that
could not be identified to species in the field were taken back to the Gaylord
MDNR office where they were identified using fish identification keys (Smith
1989).
The
numbers of dead fish picked up within the three sections were expanded to give
a qualitative view of the relative numbers and sizes of different fish species
killed in the first 3 miles of river downstream of the dam. Thus, the total numbers of dead fish
collected in the three sections were multiplied by 17.6 (i.e. (15,840 feet in 3
miles) / (900 feet searched for dead fish).
A fourth dead fish survey was conducted
concurrently with an electrofishing survey on June 25, 2008. A fisheries survey crew completing a
one-pass electrofishing survey for living fish in the entire stream reach
between Sturgeon Valley Road and Golden Lotus Dam picked up and later measured
90 dead trout they observed (Figure 3).
Dead non-trout species were not collected during this survey.
Evaluation of whether trout escaped to refugia
During
the first 10 days after the event, MDNR survey crews conducted one-pass
electrofishing surveys at eight sites and a visual survey at one site located
downstream of the dam (Figure 1, Table 1).
Electrofishing catch rates of trout collected during these surveys were
standardized by converting them to numbers caught per 1,000 feet of
stream. The 2008 catch rates were then
compared to catch rates per 1,000 feet from the same stream sections during
surveys conducted in previous years.
The existence of substantially lower catch rates in 2008 as compared to
catch rates in years when there was not a documented fish kill was evidence
that trout had not found refugia in tributaries or springs within or near a
sampled section.
The
2008 electrofishing survey conducted at the Agnes Andreae Preserve, located
about 28 miles downstream of the dam, was conducted because MDNR knew from
previous survey data that few trout reside in the river this far downstream. If trout abundance had been high at this
site in 2008, it would have been evidence that some trout had escaped the
noxious conditions caused by the event by fleeing downstream. Similarly, if trout abundance in the Little
Pigeon River had been higher than expected based on previous surveys it would
have been evidence that trout had found refuge in the only significant
tributary between the dam and highway M-68.
The visual survey of Nelson Creek, a very small tributary to the Pigeon
River, was conducted to determine if large trout from the Pigeon River were
visible. Because the stream is so
small, any large individuals that fled there would have been readily visible.
Comparisons
of population estimates conducted during July 2008 to estimates conducted in
prior years in the same river sections were used as the most quantitative test
of whether trout found refugia. Because
the population estimates were conducted well after the 2008 event, if fish had
escaped to refugia they would have returned to the mainstem of the Pigeon River
by the time the estimates were made.
Thus, if population abundance of trout at sites downstream of the dam
remained low in July 2008 compared to prior years, this showed that they did
not find escape refuge from the event.
Comparisons of trout abundance in the Pigeon River to
abundance in other trout streams.
The
purpose of these comparisons was to show that the Pigeon River trout
populations ranked lower relative to other streams after the 2008 fish kill
than during prior times. Since 2002, a
large number of trout streams across Michigan have been sampled using
standardized protocols under the Stream Status and Trends Program (Wills et al.
2006). Population estimates collected
as part of this program provide reference data that allowed MDNR to compare
trout density from other streams in the Northern Lower Peninsula to abundance
in the Pigeon River in 2008 as well as during previous periods.
MDNR
graphically compared combined brook trout and brown trout density in the Pigeon
River to density estimates derived from contemporary population estimates made
on 21 other Michigan trout streams located in either the Central Lake Michigan
or Northern Lake Huron Management Units (Figure 4). Trout streams in this region are similar in terms of hydrology,
water chemistry, substrate, and channel morphology because they are associated
with similar geology and are subject to a similar climate. MDNR used data only from streams that were
30 feet wide or wider so they would be more comparable to the Pigeon River,
which averages about 41 feet wide downstream of the dam. The average width of our 21 reference
streams was 48 feet. For the majority
of these streams, three or more mark-and-recapture population estimates were
conducted in the summer of 2002 or later so our reference density levels were
based on 71 population estimates for brook trout and brown trout. No trout were stocked in stream sections
used to develop the database. Trout
abundance data for each stream was first averaged for all surveys at the same
fixed-site section of stream, most of which were 1,000 feet long. Trout density was computed as numbers and
pounds per surface acre of water to make the data more comparable between streams
of different widths. From these density
data MDNR computed the 75th percentile (75% of sites have fewer
fish), the median (half of sites have fewer fish), and 25th
percentile (25% of sites have fewer fish).
MDNR first plotted the combined density of brook trout and brown trout
young-of-the-year fish per acre and yearling-and-older fish per acre estimated
for different sections of the Pigeon River at different times. Plots were then inspected to determine how
trout density in the Pigeon River compared to density in other streams in the
region in 2008 and during previous times.
MDNR made similar plots of total pounds per acre of brown trout and
brook trout combined.
MDNR
also graphically compared numbers of young-of-the-year and yearling-and-older
rainbow trout in the Pigeon River in 2008 to density estimates made for 12
other Michigan trout streams wider than 30 feet and located in either the
Central Lake Michigan or Northern Lake Huron Management Units (Figure 4). Again, most of these streams had been
surveyed 3 or more times since 2002 so the reference data were based on a total
of 40 mark-and-recapture population estimates for rainbow trout.
Economic valuations of the 2008 fish kill
MDNR’s economic valuation of this fish kill makes use of
contemporary and historical fish and angler use data collected from the Pigeon
River, data from the 2006 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation in Michigan (U.S. Dept. of the Interior, Fish
and Wildlife Service, and U.S. Dept. of Commerce, and U.S. Census Bureau 2008),
and the American Fisheries Society procedures and guidelines for determining
costs associated with a fish kill like that being reported on here (Southwick
and Loftus 2003). Typically, monetary
damages from fish kills consist of replacement costs, lost user value, and lost
nonuse value. However, valuation of
nonuse costs in this case could not be reliably quantified and are not included
in this report.
MDNR’s valuation of replacement costs includes estimates
of the cost of replacement fish and the costs of investigating and managing the
kill. MDNR used replacement costs for
brown, brook, and rainbow trout in the Great Lakes-Big Rivers, FWS Region 3
reported in Southwick and Loftus (2003).
Those guidelines provide replacement costs for these trout species by
1-inch size categories beginning at 1-inch through the largest sizes they could
obtain data for. For brook, brown, and
rainbow trout the size categories ranged from 1 to 18 inches. Because MDNR wanted to minimize the
possibility to overvalue smaller fish and undervalue larger fish when using an
average replacement cost, and because we report our population estimates of
trout killed in four broad size categories (1 to 4.9 inches, 5 to 7.9 inches, 8
to 11.9 inches, and 12 inches and longer), we averaged the replacement costs
reported in Southwick and Loftus (2003) within these broader inch groupings to
obtain an average replacement cost for a fish within a given category. Also, replacement costs for fish in Southwick
and Loftus (2003) are reported in 2002 dollars, so we adjusted the total cost
to reflect the 2008 value using an inflation calculator available on the
internet based on the Consumer Price Index (CPI), from 1800 to 2007 (http://www.westegg.com/inflation/). As discussed in Southwick and Loftus (2003)
these replacement costs for the fish killed should be considered a minimum
value or underestimate of the true economic value of the lost fish.
MDNR also tracked Fisheries Division costs for investigating
and managing the fish kill and include those costs from the time the
investigation began through March 2009 in the estimate of replacement costs
here. That estimate includes the cost
of labor, materials, and travel associated with this investigation.
User
value costs were calculated by multiplying an estimate of the number of trips
taken annually to fish the impacted 25 mile section of the Pigeon River
(Alexander and Ryckman 1986) by the reported average dollars spent per trip
fishing in Michigan (U.S. Dept. of the Interior, Fish and Wildlife Service, and
U.S. Dept. of Commerce, and U.S. Census Bureau 2008). MDNR limited the calculation of the average dollars spent per
outing to trips taken to fish on inland waters and excluded trips taken to fish
on the Great Lakes proper.
Field methods for mark-and-recapture population
estimates
During summer 2008 trout populations were
estimated using standard mark and recapture methods (Ricker 1975) in 14 river
reaches encompassing 46,074 feet (8.73 miles) of stream (Figure 2). Two reaches were located upstream of the dam
and the remainder were located downstream of the dam. Trout were captured by wading electrofishing crews using 240 volt
DC electrofishing units with three capture electrodes. Crews commenced electrofishing at the
downstream end of sampling stations and swept the electrical field throughout
the stream while wading in an upstream direction. Two sampling runs were made in each river reach. On the first, or marking run, all trout
captured were measured, marked by clipping the tip of their caudal fin, and
then immediately released back to the water near where they were captured. On the second, or recapture run, captured
trout were again measured and examined for marks (caudal fin clips).
Computational
methods for population estimates
Population estimates (P) of surviving trout
were computed using the Chapman modification of the Petersen formula where P =
((M + 1) * (C + 1))/(R + 1)) – 1, where M signifies the number of fish marked
on the first run, C signifies the total number of fish caught on the recapture
run, and R is the number of marked (fin clipped) fish caught on the recapture
run (Hayes et al. 2007). Fisheries
researchers have known for over a half century that the capture efficiency of
electrofishing gear varies with fish size (Sullivan 1956). Capture efficiency for smaller fish is
typically lower than for larger fish both because they are harder to see and
because they have a weaker reaction to the electrical field. Thus, trout population estimates are often
computed for each individual inch class of trout and then summed for a total
population estimate. When numbers of
fish marked or recaptured for particular inch classes are low, the probability
of statistical bias in the computed population for that size class
increases. MDNR minimized the
probability of such bias by combining size classes of trout with similar
catchability (1-3 in, 4-5 in, 6-7 in and 8-in-and-larger trout for each trout
species present at each sampling site) before computing population
estimates. These combined inch class
estimates were then apportioned into numbers for individual inch classes based
on the proportion of new (once-caught) fish caught on both runs (Wills 2006).
Selection of reference periods or single years for
comparison to 2008 abundance
Different
reference periods were used to compute trout mortality for different reaches of
the river based on availability of data and other factors that are described
below. The most extensive pre-kill data
were available for the 6 miles of river immediately downstream of the dam
(sections A-E). MDNR estimated
mortality in this river section by subtracting abundance estimates for July
2008 from the average of population estimates made in August 1975-76 and/or
September 1979-80. MDNR did not make
comparisons of abundance in 2008 to abundance during the period from 1949-65
because there is compelling evidence that abundance of trout, particularly
large brown trout and juvenile rainbow trout was much higher immediately prior
to the 2008 fish kill than during 1949-65.
These reasons are described below:
The
number of large dead trout observed after the 2008 event was substantially
higher than the abundance of large trout observed in population estimates
conducted prior to 1975. On June 25, 2008, three days after the initial fish
kill, an electrofishing survey crew picked up 40 dead brown trout ranging from
12.0 to 23.5 inches long between Sturgeon Valley Road and the dam. This was more large brown trout than were
present in the section in any year between 1954 and 1965. Clearly, the abundance of large brown trout
in this section immediately before the 2008 fish kill was much higher than
historical levels.
Rainbow
trout (steelhead) and brown trout reproduction has increased through time due
to increases in the numbers lake-run fish from Mullet Lake that spawn in the
river. Treatments of the Pigeon River
to kill larval sea lamprey have resulted in the establishment of runs of
rainbow trout and brown trout from Mullet Lake. The Pigeon River was treated for lamprey twelve times between
August 7, 1966 and July 13, 2007 (E. Koon, U.S. Fish and Wildlife Service,
personal communication). In the past,
rainbow trout and brown trout that moved out of the Pigeon River into Mullet
Lake were heavily preyed upon by sea lamprey and rarely survived to make a
spawning run back into the Pigeon River.
Juvenile steelhead were virtually absent from the Pigeon River prior to
the late 1970s but since that time their abundance has increased
dramatically.
It
took more than two decades for the steelhead runs to become well established,
but the large size of contemporary steelhead runs was very evident on April 19,
2005 when MDNR Fisheries Division personnel counted 49 individual steelhead
redds and 2 redd complexes between Sturgeon Valley Road and the Golden Lotus
Dam, a distance of about 1.2 miles.
Given that each female steelhead may deposit from 3,000-5,000 eggs into
a redd depending on her size, the potential number of juvenile steelhead
produced in the river is clearly very large. The size of lake runs of brown
trout are less formally documented by survey data but are well known to anglers
and local fisheries biologists. In late
July 2008, fisheries survey crews captured lake-run brown trout at the Trail
34-Elk Meadows site. Brown trout that
have recently emigrated from a lake are large, silver in color, and more robust
than stream-resident trout and hence can be easily distinguished from
them. Similar runs of brown trout also
occur in the adjacent Sturgeon River that flows into Burt Lake. As in the Pigeon River, lake-run brown trout
begin to ascend the Sturgeon River in July and August prior to spawning later
in the fall. The increase in egg
deposition by large brown trout from Mullet Lake clearly has the potential to
increase recruitment of young brown trout, which in turn would be expected to
increase the abundance of older stream-resident brown trout above historical
levels because a portion of the offspring of lake-run fish remains as residents
of the stream. Zorn and Nuhfer (2007)
showed that brown trout reproduction levels in the Au Sable River system were
positively related to spawner abundance and that higher reproduction subsequently
resulted in higher abundance of older and larger trout, assuming that flows
were reasonably stable.
Voluntary
release of legal-sized trout by anglers has increased steadily over the past 30
years, which would be expected to result in increases in abundance of large
brown trout. Clark and Alexander (1992)
reported that voluntary release of legal-sized brown trout in the North Branch
of the Au Sable River increased from around 40% in 1976 to 80% in 1990. There are no data on voluntary release rates
in Michigan prior to 1976 because the behavior occurred at such a low frequency
that census clerks were not instructed to ask anglers about released fish.
For
the reasons outlined in the paragraphs above, MDNR estimated mortality in the
first 6 miles of river immediately downstream of the dam using reference data
from population estimates made in August 1975-76 and September 1979-80. At locations further downstream MDNR used
data collected in 1976 population surveys at Webb Road, McIntosh Landing, and Pigeon
River Bridge as reference data for computing mortality due to the 2008
event. For these three downstream
sites, data collected in 1976 were the only usable data available for
comparison to data collected in 2008.
Population data collected at Webb Road and McIntosh Landing in 1984 were
not usable for analyses because populations were depressed as a result of a
previous fish kill caused by operations of the dam in 1984.
Assessment of the comparability of trout populations
in 1976 and 2008
MDNR
compared trout population data collected in the river upstream of the Golden
Lotus impoundment, and therefore unaffected by dam operations, in July 1976 and
July 2008 to evaluate whether it was reasonable to assume that overall trout
populations were similar between these years.
Trout population data for yearling-and-older brook trout and brown trout
in July 1976 near Old Vanderbilt Road reported by Enk (1977) were compared to
population data collected by the MDNR in July 2008. Enk (1977) estimated that there were 154 yearling-and-older brook
trout and brown trout combined in a 300-m (984 ft) section of the Pigeon River
extending upstream from Old Vanderbilt Road Bridge (Enk’s stations 1 through
3). In July 2008, the MDNR estimated
that there were 137 yearling-and-older brook trout and brown trout combined in
a 1,000-foot section of the river centered on Old Vanderbilt Road Bridge. Due to a beaver dam located upstream of the
bridge, MDNR was unable to sample exactly the same section surveyed by Enk in
1976. Enk (1977) did not estimate
abundance of young-of-year trout so their abundance could not be compared
between 1976 and 2008.
Because
the 1976 and 2008 population data were very similar, MDNR concluded that
computing fish killed in 2008 by comparison to 1976 population data was a
reasonable approach. Moreover, there
were no other population data available at downstream sites on the Pigeon River
except for data collected shortly after the 1984 fish kill.
Results
Determination of the downstream extent of the fish kill
One-pass electrofishing surveys conducted downstream
of the dam within the first 10 days after the 2008 event indicate that the kill
may have extended as far as the Agnes Andreae Preserve located 27.7 miles
downstream of the dam (Figure 1). Total
catches of brook trout and brown trout combined were nearly 40 times higher
prior to the 2008 event in the river section immediately downstream from the
dam (Table 3). At the Agnes Andreae
Nature Preserve catches of brook trout and brown trout combined were 4 times
higher in 2003 than in 2008. Rainbow
trout were not considered in this analysis because their abundance level was
low in the river during the times reference data were collected in the
past. Reasons for this are given in the
methods section of this report.
Estimates of the downstream extent of the kill were
defined further after population estimates were made (Figure 2). The 2008 population estimates were compared
to reference population estimates to determine the percentages of brook and
brown trout killed at each sampling station.
Groups of years or single years used for reference data are described in
the methods section under Estimation of
numbers of brook and brown trout killed by the event. The 2008 event resulted in mortality of
almost all young-of-year and yearling-and-older brook trout and brown trout in
the first 2.3 miles or river downstream of the dam (Table 4). Mortality estimates declined somewhat at
sites located further downstream. At
Webb Road, 14 miles down from the dam, mortality of young of year was 91% and
mortality of yearling-and-older-fish was 73%.
Pigeon River Road Bridge was the site located the farthest downstream
from the dam where MDNR had usable pre-2008 data for estimating mortality
caused by the event. MDNR estimated
that 89% of young-of-year brook trout and brown trout and 66% of yearling and
older brook trout and brown trout were killed in Pigeon River Road Bridge area, which is 20.75 miles downstream of
the dam (Table 4).
Given that
very large percentages of trout were killed by the event as far downstream as
Pigeon River Road Bridge, there was no reason to expect that the kill zone
ended there. In addition, catches of
brook trout and brown trout per 1,000 feet at the Agnes Andreae Nature
Preserve, located 27.7 miles downstream of the dam were 4 times higher in 2003
than in 2008 (Table 3). Therefore, MDNR
used the highway M-68 Bridge, located 24.7 miles down from the dam as the
downstream end of the kill zone for purposes of computing trout mortality from
the 2008 event.
Estimation of numbers of trout killed by the 2008 event
Estimates of numerical abundance of trout during
different years, percentage of trout killed by the 2008 event, and estimated
numbers of trout killed are presented for individual sections of the Pigeon
River downstream of the dam in a series of tables in this report (See Tables 5
through 13). All population data
available for the river sections is presented in the tables to provide a
historical perspective as well as to allow comparisons of the fish kill due to
the 2008 event to the fish kill caused by the 1984 event.
Numbers killed in the first 6 miles
below the dam
In Section E, the first 1.17 miles of river below
the dam, MDNR conservatively estimated that over 6,000 trout were killed by the
event (Table 5). Data in Table 5 show
that the 2008 fish kill was substantially more severe than a previous kill in
1984. Over 9,500 trout were killed by
the event in Section D, whose downstream boundary is 2.4 miles downstream from
the dam (Table 6). Again, the 2008 kill
in Section D was much more severe than the 1984 kill. MDNR estimates that 21,806 trout were killed in Section C by the
event (Table 7). Less than half as many
brook trout and brown trout combined survived the 2008 kill as compared to the
1984 fish kill in Section C. In Section
B over 2,500 trout were killed in 2008 (Table 8). Estimates of trout killed in Section B were lower than for
sections upstream primarily because mortality of young-of-year trout was lower
here. However, mortality of
yearling-and-older trout in section B remained high at 69% (Table 8). MDNR estimated that 47,503 trout were killed
in Section A by the 2008 event (Table 9).
A large proportion of these fish were juvenile rainbow trout.
MDNR estimates that a total of 87,229 trout were
killed in the first 6 miles of the Pigeon River downstream of Golden Lotus dam
is Sections A through E (Tables 5 through 9).
This estimate of fish mortality is substantially higher than estimates
of losses following the 1984 kill (Alexander and Ryckman 1986), in part because
rainbow trout were a relatively minor component of the trout community at that
time compared to more recent times.
Alexander and Ryckman (1986) estimated that total abundance of trout in
Sections A-E was reduced by 53% whereas in 2008 total abundance of trout was
reduced by 75%.
Numbers of trout killed between
Section A and McIntosh Landing
MDNR estimated numbers of trout killed in the
9.38-mile section of the Pigeon River between the downstream end of Section A
and McIntosh Landing based on population data collected at four sites in the
reach: Tin Bridge, Trail 34-Elk Meadows, Pine Grove Campground, and Webb Road
(Figure 2). At the Webb Road
population index site, located 14 miles down from the dam, comparisons of
numerical changes in trout populations between 1976 and 2008 indicated that 73%
of yearling-and-older trout and 91% of young-of-year trout were killed by the
2008 event (Table 10). High mortality
here was expected given the extremely high turbidity levels at Webb Road observed
by MDNR fisheries personnel on June 24, 2008 on the day the fish kill was
officially discovered and reported but two days after the initial event.
MDNR estimated that 337,183 trout were killed in the
section of the Pigeon River between the downstream end of Section A and the
upstream end of the McIntosh Landing population index site (Table 11).
Numbers of trout killed between
McIntosh Landing and highway M-68
MDNR estimated numbers of trout killed in the
9.5-mile section of the Pigeon River between McIntosh Landing and highway M-68
based on population data collected at three sites in the reach: McIntosh
Landing, the Cudlip property, and Pigeon River Road Bridge (Figure 2). Comparisons of population data collected in
1976 and 2008 from this section of river indicated that the 2008 event resulted
in the mortality of 92% of young-of-year trout and 67% of yearling
and-older-trout (Table 12). MDNR
estimated that 50,074 trout were killed in this 9.5 mile section of the Pigeon
River (Tables 13).
Grand total numbers of trout killed
between the dam and highway M-68
MDNR estimates that the 2008 event caused by the
Golden Lotus Inc. Dam killed 474,486 trout in the Pigeon River between the dam
and highway M-68 Bridge (Table 14).
Numerical losses were highest for young-of-year trout, particularly
rainbow trout. Nearly 25,000
yearling-and-older trout were killed by the event and over 2,000 of these fish
were 12 inches or longer (Table 14).
Species composition and size ranges of
non-trout species killed
MDNR
data clearly show that the 2008 event at the Golden Lotus Inc. dam was severe
enough to kill at least 13 non-trout species of fish, most of which are more
tolerant of degraded water quality than trout species. In addition to trout species, an additional
13 species of fish were collected during searches of three 300-foot sections of
the Pigeon River located within about 2.2 miles downstream of the dam (Figure
3).
A
total of 206 fish representing 9 non-trout species were collected from a 300-foot
section located 0.62 miles downstream of the dam (Table 15). Dead fish collected in this section ranged
from 2 to 16 inches long. White suckers
were the most abundant species in the collection. The release from the dam apparently killed virtually all suckers
from the dam to Sturgeon Valley Road because an electrofishing crew that
surveyed this entire section on June 25, 2008 collected only 10 living white
suckers. A fisheries biologist at the
site observed dead white suckers lodged immediately upstream of the dam
spillway while investigating the event (T. Cwalinski, MDNR Fisheries Biologist,
personal communication). This indicates
that noxious conditions lethal to fish extended into the impoundment, as well as
downstream of the dam as a result of the 2008 event.
At
Sturgeon Valley Road a total of 304 fish representing 10 non-trout species were
collected from a 300-foot section of river centered on the bridge (Table
16). White sucker and creek chubs were
the most common species found in this collection. Non-trout species collected at this site ranged from 1 to 13
inches long. At the third collection
site, located about 2.2 miles downstream of the dam, 364 dead fish representing
12 non-trout species were collected on June 25, 2008 (Table 17). Non-trout species collected here ranged from
1 to 7 inches long. Here, blacknose
dace and creek chub were most frequently found. All dead fish surveys were
severely hampered by high turbidity levels in the river so the counts could not
be used to make quantitative estimates of non-trout species killed.
It
is notable that the total number of dead fish collected in the three 300-foot
sections actually increased with distance downstream from the dam. In fact, the number found closest to the dam
(206 fish) was lower than at Sturgeon Valley Road (304 fish), and this number
was lower than the number found 2.2 miles down from the dam (364 fish) (Tables
15 through 17). Expansion of dead fish
counts in the 300-foot sections that were intensively searched to the first 3
miles of stream immediately downstream of the dam yields an ultraconservative
estimate of over 15,000 dead (non-trout) fish killed here (Table 18).
A
fourth MDNR search for dead fish was conducted concurrently with an
electrofishing survey of the 1.2 miles of river between Sturgeon Valley Road
and the dam on June 25, 2008. During
this survey, the crew picked up dead trout that they observed during the course
of collecting living fish with the survey gear. On this survey the crew picked up 90 dead trout ranging from 5 to
23 inches in length (Table 19). Smaller
dead trout were greatly underrepresented in this dead fish collection because
they were more difficult to see in turbid water than larger individuals. About 2/3 of the sample of dead brown trout
collected upstream of Sturgeon Valley Road were old fish measuring 12 inches or
longer.
Again,
these data attest to the massive mortality of fish that occurred in the Pigeon
River in June 2008. It is also very
likely that some of the dead fish were killed in the impoundment and swept
downstream. If this was the case, it
indicates that the fish kill resulting from faulty operation of the dam
extended both upstream into the pond and downstream of the dam, but MDNR had no
way to estimate the magnitude of the kill within the impoundment.
Analysis of whether trout found refuge
from the release in tributaries or downstream areas
MDNR
conducting the one-pass survey at the Agnes Andreae Nature Preserve, in part,
to ascertain if unexpectedly high densities of trout were present. After the 1984 fish kill Golden Lotus Inc.,
owners of the dam, hypothesized that trout might have escaped mortality from
that event by moving downstream or by finding refuge in tributary streams. The one-pass electrofishing survey of the
Little Pigeon River at Webb Road and the visual survey of Nelson Creek were
conducted to determine if there was any evidence that trout in the main stem
Pigeon River sought refuge in tributaries to escape the event at the Golden
Lotus Dam ( Figure 1, Table 1). Based
on previous survey data, MDNR did not expect to find many trout at the Agnes
Andreae Nature Preserve site on June 27, 2008 unless they had fled there to
escape the noxious conditions resulting from the 2008 event at the dam.
Low
numbers of trout were captured at the Agnes Andreae Nature Preserve in both
2003, when there was no documented fish kill, and were even less abundant in
2008 (Table 3). This indicates that
trout did not escape effects of the noxious conditions resulting from the 2008
event by moving downstream.
The
electrofishing survey of the Little Pigeon River at Webb Road conducted on June
27, 2008 indicated that trout did not find refuge from the event in the main
stem Pigeon River by moving to this tributary.
There were over twice as many yearling-and-older trout at Webb Road in
1984 than in 2008 (Table 3). Moreover,
only 2% of trout caught in the 2008 survey of the Little Pigeon River were
brown trout whereas 97% of the trout in the main stem of the Pigeon River near
the mouth of the Little Pigeon River were brown trout or rainbow trout. The McIntosh Landing site sampled in 2008
was just upstream of the mouth of the Little Pigeon River and the Cudlip
property was located just downstream (Figure 2, Table 13). A visual survey of 200 feet of Nelson Creek
conducted on June 27, 2008 immediately upstream from its confluence with the
main stem Pigeon River likewise indicated that trout did not escape the event
by fleeing into this tributary (T. Cwalinski, MDNR Fisheries Biologist,
personal communication). Nelson Creek
is a very small, clear stream where large trout immigrants from the main stem
Pigeon River would have been readily visible.
Comparisons of trout abundance in the
Pigeon River to other trout streams
Overall
density of brook trout and brown trout in the Pigeon River in the 1970s and
1980 compared favorably with median expected densities for trout streams in the
Northern Lower Peninsula but their density was far lower than expected after
the 2008 fish kill. Figure 5 shows that
density of young-of-year brook trout and brown trout during the 1970s and 1980
in all sections of the river from the dam to McIntosh Landing fell between the
25th and 75th percentiles for density at Michigan status
and trends sites. By contrast, in
summer 2008 after the fish kill, densities of young-of-year trout in 7 of 12
sections of the river were well below the 25th percentile for
density (Figure 6). Section B was the
only reach where densities were relatively similar among periods.
Figure
7 shows that density of yearling-and-older brook trout and brown trout during
the 1970s and 1980 from the dam to McIntosh Landing fell between the 25th
and 75th percentiles for contemporary trout streams in the northern
lower peninsula of Michigan. After the
2008 fish kill density at all sites except Tin Bridge fell below the 25th
percentile (Figure 8). Total pounds per
acre of brook trout and brown trout in the 1970s and 1980 were at or above the
median for Michigan status and trends streams at all sites from the dam down to
McIntosh Landing (Figure 9). Following
the 2008 fish kill, pounds per acre was below the 25th percentile at
all sites (Figure 10). These data
demonstrate the severity of the fish kill throughout the reaches surveyed in
2008.
Densities
of rainbow trout in the 1970s and 1980 were not compared to contemporary
abundance at Michigan status and trends sites because rainbow trout were quite
rare in the river until recent times as was discussed earlier in this report
(See Methods). In 2008, density of
young-of year and yearling-and-older rainbow trout at about half of sites
surveyed fell between the 25th and 75th percentiles for
abundance in the 12 status and trends reference streams in the Central Lake
Michigan and Northern Lake Huron management units with rainbow trout
populations (Figures 11 and 12).
Economic valuations of the fish kill
MDNR
valuation of the fish kill includes an estimate of the cost to replace the
trout killed in June 2008 when the event occurred. Using the replacement costs
reported for brown, brook, and rainbow trout in the Great Lakes-Big Rivers, FWS
Region 3 (Southwick and Loftus 2003), MDNR calculated an average cost to
replace a trout for all three species in each of the following inch groupings:
1 to 4.9 inches; 5 to 7.9 inches; 8 to 11.9 inches; and 12 inches and longer
(Table 20). MDNR multiplied these
averaged replacement costs for each species and inch grouping by the estimates
of the total number of trout killed for the same categories (Table 14) to estimate
the replacement costs for the trout killed by the event (Table 21). MDNR estimated a replacement cost for the
trout killed of $218,970 in 2002 dollars.
MDNR estimates the replacement cost for the trout killed at $260,671 after adjusting that figure to represent 2008
dollars.
According
to Southwick and Loftus (2003), it is appropriate to include the cost of
investigating and managing a fish kill.
In FY2008, MDNR, Fisheries Division incurred total direct costs of
$90,467 investigating and managing this fish kill on the Pigeon River (payroll
= $85,210; travel = $2,025; and lease vehicle costs (VTS) = $3,232). FY2009 costs through March 2009 total
$29,424 with the bulk of those costs being investigator payroll while analyzing
data and writing reports. Combining the
FY2008 and FY2009 costs, MDNR estimates a total cost of $119,891 incurred
through March 2009 while investigating and managing this fish kill.
The
2006 National Survey of Fishing, Hunting, and Wildlife-Associated Recreation in
Michigan (U.S. Dept. of the Interior, Fish and Wildlife Service, and U.S. Dept.
of Commerce, and U.S. Census Bureau 2008) estimated that $966,314,000 was spent
on freshwater fishing trips and equipment in Michigan excluding trips taken to
fish on the Great Lakes proper.
Included in that figure were estimates for expenditures on food and
lodging, transportation, equipment and other trip costs. That report also estimated 14,360,000
fishing trips were taken in Michigan, again excluding trips taken to fish on
the Great Lakes proper. Dividing the
estimate for total expenditures by the estimated number of trips yields an
estimate of $67.29 per fishing trip in Michigan.
The
most recent intensive creel census of trout angler effort and trout catches in
the 25 mile section of the Pigeon River impacted by this fish kill was
conducted in 1979 and reported in Alexander and Ryckman (1986). Results reported in that study showed that
4,275 angler trips were taken to this specific stretch of the Pigeon River
annually and would be at risk if angling were to cease. That study also reported that angler use did
in fact cease after a similar event like that being reported on here occurred
at this same dam early during the 1984 fishing season. As reported in Alexander and Ryckman (1986),
it is likely that normal fishing effort ceased for the remainder of the 2008
fishing season after this more recent event and subsequent fish kill.
Multiplying
the estimate of $67.29 spent per fishing trip by the 4,275 angler trips taken
to fish this same stretch of the Pigeon River results in a lost user value in
2008 of $287,674. This is likely a
conservative estimate of the lost user value because it does not take into
account recent increases in costs associated with taking a fishing trip like
the cost of gasoline, and does not account for residual lost angling effort in
coming years while trout populations and the fishery rebuilds over time. It also doesn’t include the indirect impact
a lack of angler spending had after the recent event on the local and regional
economies, or will have in coming years as the fishery rebuilds over time.
Combining
the cost to replace the trout killed as a result of the event ($260,671 in 2008
dollars), the cost to investigate and manage the fish kill through March 2009
($119,891), and the lost user value in 2008 of $287,674, MDNR estimates that
the economic impacts of the fish kill that resulted from from the event that
occurred at the Golden Lotus Dam in June 2008 is at least $668,236.
Discussion
Estimates of the numbers of trout killed
MDNR estimates of brook trout and brown trout killed
were conservative for some river sections.
For example, MDNR estimated that 61 brown trout over 12 inches long were
killed in Section E in the first 1.17 miles down from the dam. However, on June 25, 2008 the electrofishing
crew that surveyed this reach picked up 40 dead brown trout over 12 inches long
here, which is nearly 2/3 of the 62 brown trout of this size present during the
years used as the pre-kill reference period (Table 5). It is certain that the crew detected far
fewer than 65% of dead trout given the turbid conditions at the time and the
fact that most effort was directed at determining what living fish remained in
the stream section. Some of the dead
brown trout collected from Section E may have been killed in the impoundment
and later drifted downstream. This
would indicate that the boundaries of the fish kill caused by operation of the
Golden Lotus Dam extended up into the impoundment as well as downstream of the
dam, although MDNR had no data to estimate numbers of fish killed in the
impoundment.
MDNR examined the possibility that the large
estimates of numbers of juvenile rainbow trout killed resulted because they had
not yet emerged from their redds at the time of the event. MDNR explored this possibility of bias by
estimating swim-up or emergence times for young-of-year steelhead in the Pigeon
River. Water temperature is the major factor
controlling the time required for a steelhead egg laid into a redd in the spring
to develop and emerge from the gravel.
Most steelhead in Michigan’s Lower Peninsula spawn during the month of
March or April. On April 19, 2005, MDNR
Fisheries Division personnel counted 49 steelhead redds and 2 redd complexes
(multiple redds at one site) between Sturgeon Valley Road and the Golden Lotus
Dam (Section E). No adult steelhead
were observed, which strongly suggests that most, if not all spawning had
already occurred by April 15, 2005.
MDNR made predictions of emergence dates based on
daily water temperature data collected near Elk Hill in the Pigeon River and
made the conservative assumption for modeling purposes that all spawning
occurred on April 15, 2008. Water
temperature data collected near Elk Hill were input into equations relating
average daily water temperatures to hatching and emergence time for rainbow
trout (Crisp 1981, 1988). Hourly
temperature data collected by an electronic thermometer deployed near Elk Hill
were used to compute average daily water temperatures from April 25, 2008
through the end of the emergence period.
Since the thermometer was not deployed at this site until April 25, 2008
MDNR used water temperature data collected in April 2004 for the first 9 days
of egg development (April 16 through April 24). Actual water temperatures in the Pigeon River in 2008 would have
been higher than in 2004 during the first 9 days of egg development because
average air temperatures were 90F warmer in 2008. In other words, using 2004 water temperature
data resulted in a later hatch date prediction.
MDNR’s analysis predicted a median emergence date of
May 31, 2008 for steelhead eggs laid on April 15, 2008. In the hatchery, virtually all steelhead
eggs fertilized on a specific day swim up (emerge) within a week of each other
(J. VanAmBerg, MDNR Hatchery Manager, personal communication). Thus, MDNR concluded that most steelhead
would have emerged from their redds well before the June 2008 event, because
most steelhead in the Pigeon River would have spawned on or before April 15,
2008 and incubation temperatures in the river would have resulted in emergence
of steelhead fry by early June even if they spawned as late as the middle of
April.
MDNR’s
finding that fairly large numbers of rainbow trout survived the fish kill does
not conflict with the large estimates rainbow trout killed if their populations
prior to the event were similar to those in the neighboring Sturgeon
River. MDNR estimated that 376,070
young-of-year rainbow trout and 8,885 yearling-and-older rainbow trout were
killed in the 24.7-mile reach of the Pigeon River between the dam and
M-68. This equates to a kill of 3,055
young-of-year rainbow trout per acre and 72 yearling-and-older rainbow trout
per acre. If contemporary densities of
rainbow trout in the Pigeon River immediately prior to the kill were similar to
those in the Sturgeon River watershed, which lies immediately to the west, then
it can be shown that these kill estimates are reasonable. The Sturgeon River gets spawning runs of
steelhead and brown trout from Burt Lake in the same fashion that the Pigeon
River gets spawning runs from Mullet Lake.
The average density of young-of-year rainbow trout in the Sturgeon River
at Trowbridge road from 2005-07 was 5,081/acre while yearling-and-older density
was 132/acre. In the West Branch
Sturgeon River, a tributary to the main stem Sturgeon River, the densities of
young-of-year and yearling-and-older rainbow trout during the same period were
1,832/acre and 114/acre, respectively.
Thus, average densities of rainbow trout at Sturgeon River watershed
sites were 3,456/acre for young-of-year and 123/acre for yearling-and-older (T.
Wills, MDNR Fisheries Research Biologist, unpublished data). MDNR’s estimate of young-of-year rainbow
trout killed in the Pigeon River divided by average density in the Sturgeon
River watershed [(3,055/3,456)*100) = 88%] is similar to the mortality rates
MDNR observed for young-of-year brook trout and brown trout. Mortality of young-of-year brook trout and
brown trout averaged 79% and ranged up to 96% due to the event (Table 4). Similarly, MDNR’s estimate of
yearling-and-older rainbow trout killed in the Pigeon River divided by density
in the Sturgeon River watershed [(72/114)*100) = 63%] again indicates that the
estimates were reasonable, or perhaps even underestimates, because mortality of
yearling-and-older brook trout and brown trout averaged over 75% in the Pigeon
River (Table 4).
Underestimation of the severity of the
fish kill from dead fish searches
MDNR’s observations of dead, non-trout species as
well as dead trout could not be used to depict the true severity of the fish
kill in 2008. Even when good conditions
for recovery of dead fish such as clear water and relatively shallow depths
exist detection rates may be quite low.
Taube (1974) was able to locate and retrieve less than 50% of adult coho
salmon introduced into a 1.4-mile stretch of the Platte River in Benzie County
MI after they spawned and died. The
Platte River study reach was slightly larger than the Pigeon River but was
characterized by clear water and mostly shallow depths and adult coho salmon
are very large individuals that would be much more conspicuous than smaller
brown trout. A delay between the times
a fish kill occurs and the time searches for dead fish begin also greatly
reduce the efficiency with which fish can be located. Studies of decomposition rates of dead trout placed in the Pigeon
River during the 1950s revealed that within 72 hours after the dead fish were
introduced they were effectively camouflaged with a coating of silt and
detritus (Bacon 1953, 1955). Clearer
water conditions more conducive to locating dead fish did not exist in the
Pigeon River in 2008 until well after the initial kill event when further
searches would have been futile.
Refugia
MDNR
found no evidence that trout found refuge by fleeing downstream. However, lower mortality of trout in
portions of sections B and A as compared to sites further upstream or
downstream indicates that some river sections contained features that allowed
more trout to survive the event. Survey
crews conducting electrofishing surveys encountered “pockets” of water that
held far more living fish than the majority of water in longer reaches. Inputs of oxygenated spring water or some
other aspect of local microhabitats clearly allowed higher percentages of fish
to survive the event at some locations.
MDNR
likewise found no evidence that significant numbers of trout escaped the event
by moving to tributary streams although if small numbers of individuals found
refuge in tributaries they would not have been detected. Except for the Little Pigeon River,
tributaries to the Pigeon River downstream of the Golden Lotus Dam and upstream
of M-68 are very small and could offer little refuge to fish from pollutants in
the main stem river. The U.S. Fish and
Wildlife Service (USFWS) estimates stream discharge of significant tributaries
to the Pigeon River when they conduct chemical treatments to kill larval sea
lamprey to assure that chemicals are not rendered ineffective due to
dilution. The USFWS sea lamprey
database contains discharge data for only five tributaries to the main stem
Pigeon River. The names and mean
discharge estimates for these tributaries in cubic feet per second (cfs) from
the database were: Little Pigeon River
(19.7 cfs), Tin Bridge Creek (0.7 cfs), Cornwall Creek (2.0 cfs), Nelson Creek
(1.2 cfs) and McIntosh Creek (2.7 cfs).
By contrast, the mean of discharge estimates in the Pigeon River at Tin
Bridge made during sea lamprey treatments was 80.5 cfs. Clearly, the Little Pigeon River was the
only tributary that could offer significant refuge if fish were able to flee
the main stem river, but MDNR found no evidence that it did.
Economic impacts of Golden Lotus Dam
MDNR
estimates the economic impact of the fish kill that resulted from the event
that occurred at the Golden Lotus Dam in June 2008 is at least $668,236. This is likely a conservative estimate of
the monetary damages due the public given the paucity of data available for making
such estimates. Unfortunately, the true
costs associated with this fish kill cannot be fully accounted for using the
methods employed here. The guidelines
used to generate estimated monetary damages, while well accepted, do not
account for differences in the quality of hatchery-raised fish compared to wild
fish or long-term behavioral changes of anglers in response to the fish
kill. In addition, MDNR was unable to
reliably estimate the cost associated with lost nonuse values as a result of
this fish kill. Nor did MDNR address
negative impacts on terrestrial organisms such as mammals, birds, and reptiles
whose food supply was greatly diminished by the noxious conditions in the river
created by the event in 2008.
Of
even greater concern is the possibility for additional fish kills to occur as a
result of operating the Golden Lotus Dam in the future. The negative impact on the fishery and local
economies in 2008 was severe and it remains unclear how long it will take for
both to recover from this kill. It is
also unlikely that economic impacts will lessen in the future given the current
status of the State’s economy and forecasts for a rebound.
Conclusions and Management Implications
·
The June 2008 release
from Golden Lotus Dam and flow fluctuations caused by operations of the dam
killed an estimated 474,486 trout that collectively weighed approximately 5
tons.
o
MDNR data indicated that
the kill extended downstream to at least to M-68, nearly 25 miles down from the
dam.
o
The fish kill in 2008
was much more severe than the kill following the 1984 event.
o
MDNR found no evidence
that trout escaped mortality from the 2008 release by moving downstream or into
tributary streams.
o
Trout abundance in the
Pigeon River before the 1984 and 2008 fish kills compared favorably with
abundance in other trout streams in the Northern Lower Peninsula but after the
kills trout abundance was much lower than expected abundance in similar
streams.
·
The 2008 event killed
large numbers of at least 13 non-trout fish species in addition to killing
brook trout, brown trout, and rainbow trout.
o
MDNR could not
accurately estimate how many non-trout were killed because turbidity that
persisted after the initial fish kill hindered searches for dead fish and dead
fish quickly disappeared. Moreover,
there were no pre-kill data on abundance of non-trout species to use for
comparisons.
o
It is certain that the
number of non-trout killed was very large and that the total kill was probably
well over a half million fish.
·
Even absent future “accidents”
at Golden Lotus Dam it will be a long time before the trout population,
particularly larger individuals, can fully recover since trout 12 inches or
longer are generally 3 or more years old.
·
It is very likely that
unnatural flow fluctuations below the dam reduce reproductive success of all
trout species and hence the Pigeon River does not produce and sustain nearly as
many trout as it could under natural flow conditions.
o
It is virtually certain
that less severe fish kills occur on a regular basis, particularly kills of
young trout, but MDNR can not readily detect such events.
o
If lengthy periods of
extreme low flows such as the one that was induced in September 2003 had
happened during hot weather, temperatures in the river might have rapidly risen
to lethal levels for trout
·
Up to 16 river miles of
thermally optimal habitat might be created for trout if the dam was removed
·
There were vast losses
of recreational value and opportunity due to this fish kill
o
A conservative estimate
of the economic impact of the fish kill resulting from the event at the Golden
Lotus Dam in June 2008 is $668,236
o
This estimate minimally
accounts for the economic impact of the fish kill on local businesses.
o
It is unclear how long
it will take for the fishery to recover but it is safe to say it will be
several years.
o
It is unclear what the
eventual impact will be on the local economy but it is safe to say trout
anglers contribute significant dollars to local economies when they travel to
the region to fish and that fishing in the impacted stretch of the Pigeon River
was seriously compromised in 2008 and will remain below normal for some time to
come until trout populations rebuild.
·
The risk of fish kills
such as the one that occurred in June 2008 will increase over time since the
impoundment is presently nearly full of sediment. In the present case during a mere 6-hour period, human errors and
mechanical failures were sufficient to result in a massive fish kill
o
Six hour periods of very
unnatural flows are not uncommon downstream of Golden Lotus Dam.
·
The only way to assure
that similar fish kills do not continue to occur in the future is to remove the
dam.
·
MDNR recommends that the
Golden Lotus Dam be removed to restore the natural flow regime, to restore
natural rates of sediment transport downstream, to reduce summer water
temperatures downstream, and to restore natural free upstream and downstream
passage of fish in the Pigeon River.
References
Alexander, G.R. and J.R. Ryckman. 1986.
Effect of an abnormal discharge of sediment from the Lansing Club
Impoundment on the trout population in the Pigeon River, Otsego and Cheboygan
Counties, Michigan. Michigan Department
of Natural Resources Technical Report 86-8, Ann Arbor.
Bacon, E. H.
1953. Studies on the
disappearance of dead trout and creek chubs in the Pigeon River. Michigan Department of Conservation,
Institute for Fisheries Research Report 1392, Ann Arbor.
Bacon, E. H.
1955. Studies on the
disappearance of dead trout in the Pigeon River. Michigan Department of Conservation, Institute for Fisheries
Research Report 1442, Ann Arbor.
Clark, R. D. Jr., and G. R. Alexander. 1992.
Evaluation of catch-and-release trout fishing regulations on the South
Branch of the Au Sable River, Michigan.
Institute for Fisheries Research Report 1987, Ann Arbor.
Crisp, D. T.
1981. A desk study of the
relationship between temperature and hatching time for the eggs of five species
of salmonid fishes. Freshwater Biology
11:361-368.
Crisp, D. T. 1988.
Prediction, from temperature, of eyeing, hatching and ‘swim-up’ times
for salmonid embryos. Freshwater
Biology 19:41-48.
Enk, M. D.
1977. Instream overhead bank
cover and trout abundance in two Michigan Streams. Master’s thesis. Michigan
State University, East Lansing.
Hayes, D. B., J. Bence, T. Kwak, and B. Thompson. 2007.
Abundance, biomass and production.
Pages 327-374 in Guy, C., and M. Brown (eds.) Analysis and
interpretation of freshwater fisheries data.
American Fisheries Society, Bethesda, MD.
Latta, W. C.
1961a. The eleventh annual creel
census Pigeon River Trout Research Station 1959. Michigan Department of Conservation, Institute for Fisheries
Research Report 1611, Ann Arbor.
Latta, W. C.
1961b. The twelfth annual creel
census and progress report, Pigeon River Trout Research Station, 1960. Michigan Department of Conservation,
Institute for Fisheries Research Report 1632, Ann Arbor.
Latta, W. C.
1962a. Post-season population
estimates of wild trout in experimental sections of the Pigeon River,
1949-58. Michigan Department of
Conservation, Institute for Fisheries Research Report 1636, Ann Arbor.
Latta, W. C.
1962b. The thirteenth annual
creel census and progress report, Pigeon River Trout Research Station,
1961. Michigan Department of
Conservation, Institute for Fisheries Research Report 1647, Ann Arbor.
Latta, W. C.
1963. The fourteenth annual
creel census and progress report, Pigeon River Trout Research Station,
1962. Michigan Department of
Conservation, Institute for Fisheries Research Report 1676, Ann Arbor.
Latta, W. C.
1964. The fifteenth annual creel
census and progress report, Pigeon River Trout Research Station, 1963. Michigan Department of Conservation,
Institute for Fisheries Research Report 1695, Ann Arbor.
Latta, W. C.
1965. The sixteenth annual creel
census and progress report, Pigeon River Trout Research Station, 1964. Michigan Department of Conservation,
Institute for Fisheries Research Report 1707, Ann Arbor.
Latta, W. C.
1967. The seventeenth annual
creel census and progress report, Pigeon River Trout Research Station,
1965. Michigan Department of
Conservation, Institute for Fisheries Research Report 1734, Ann Arbor.
NorthStar Economics, Inc. 2008. The economic impact
of recreational trout angling in the driftless area. Report Published April NorthStar Economics, Inc.
Ricker, W. E.
1975. Computation and
interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of
Canada 191.
Smith, G.
1989. Guide to nongame fishes of
Michigan. Museum of Zoology, University
of Michigan, Ann Arbor.
Southwick, R. I., and A. J. Loftus, editors. 2003.
Investigation and monetary values of fish and freshwater mussel
kills. American Fisheries Society,
Special Publication 30, Bethesda, Maryland.
Sullivan, C.
1956. The importance of size
grouping in population estimates employing electric shockers. The Progressive Fish –Culturist.
18(4):188-190.
Taube, C. M.
1974. Disappearance of dead fish
and assessment of mortality. Michigan
Department of Natural Resources, Fisheries Research Report 1812, Ann Arbor.
U.S. Department of the Interior, Fish and Wildlife Service,
and U.S. Department of Commerce, U.S. Census Bureau. 2008. The 2006 National Survey of Fishing,
Hunting, and Wildlife-Associated Recreation (FHW/06-MI).
Waters, T. F. 1960.
The 1957 Pigeon River flood.
Michigan Department of Conservation, Institute for Fisheries Research
Report 1591, Ann Arbor.
Wills, T. C.
2006. Comparative abundance,
survival, and growth of one wild and two domestic brown trout strains stocked
in Michigan rivers. North American
Journal of Fisheries Management 26:535-544.
Wills, T. C., T. G. Zorn, and A. J. Nuhfer. 2006.
Stream Status and Trends Program sampling protocols. Chapter 26 in Schneider, James C. (ed.) 2000.
Manual of fisheries survey methods II: with periodic updates. Michigan Department of Natural Resources,
Fisheries Special Report 25, Ann Arbor.
Zorn, T. G., and A. J. Nuhfer. 2007.
Influences on brown trout and brook trout population dynamics in a
Michigan River. Transactions of the
American Fisheries Society 136:691-705.
Table
1 .–Description of the locations downstream of the dam where one pass
electrofishing surveys were conducted from June 24, 2008 through July 1,
2008. Sites within the table are
arranged from upstream (top rows) to downstream. Latitude and longitude are given in decimal degrees. The Little Pigeon River and Nelson Creek are
tributaries to the main stem Pigeon River.
|
|
|
|
Upstream end
|
Downstream end
|
|
Site
|
Length (feet)
|
Date surveyed
|
Latitude North
|
Longitude
West
|
Latitude North
|
Longitude West
|
|
Main stem
Pigeon River sites
|
|
Section E
|
6,178
|
6/25/08
|
45.14598
|
84.47286
|
45.15627
|
84.46771
|
|
Power line upstream
|
1,200
|
6/25/08
|
45.16500
|
84.46202
|
45.16698
|
84.45883
|
|
Behind PRC Headquarters
|
600
|
6/25/08
|
45.17179
|
84.43963
|
45.17076
|
84.43850
|
|
Tubes Campground
|
300
|
6/24/08
|
Not
|
recorded
|
45.17709
|
84.42652
|
|
Tubes Campground
|
1,200
|
6/26/08
|
45.17657
|
84.43060
|
45.17709
|
84.42652
|
|
Webb Road
|
1,200
|
6/26/08
|
45.27208
|
84.45992
|
45.27291
|
84.46164
|
|
Pigeon River Road Bridge
|
1,000
|
7/01/08
|
45.32937
|
84.49445
|
45.33167
|
84.49410
|
|
Agnes Andreae Preserve
|
1,200
|
6/27/08
|
45.39328
|
84.53187
|
45.39575
|
84.53191
|
|
|
|
|
|
|
|
|
|
Tributary
streams
|
|
Little Pigeon River, Webb
Road
|
500
|
6/27/08
|
45.27730
|
84.49730
|
45.27910
|
84.49722
|
|
Nelson Creek (visual
survey)
|
200
|
6/27/08
|
Not
recorded
|
45.29113
|
84.47118
|
Table
2.–Description of the locations on the Pigeon River where mark and recapture
electrofishing surveys were conducted from June 30, 2008 through July 29,
2008. Sites within the table are
arranged from upstream (top rows) to downstream. Latitude and longitude are given in decimal degrees.
|
|
|
|
Upstream end
|
Downstream end
|
|
Station description
|
Date of marking
run
|
Station length feet
|
Latitude
North
|
Longitude
West
|
Latitude
North
|
Longitude
West
|
|
Pigeon River at Old Vanderbilt Rd.
|
06/30/2008
|
1,000
|
45.12627
|
84.50646
|
45.13161
|
84.50368
|
|
Pigeon River at PAS
|
07/16/2008
|
1,000
|
45.13162
|
84.50089
|
45.13170
|
84.49832
|
|
Section E
|
07/16/2008
|
6,178
|
45.14655
|
84.47299
|
45.15659
|
84.46718
|
|
Section D
|
07/14/2008
|
6,230
|
45.15659
|
84.46718
|
45.16738
|
84.45859
|
|
Section C
|
07/14/2008
|
5,966
|
45.16738
|
84.45859
|
45.17260
|
84.44005
|
|
Section B
|
07/14/2008
|
6,283
|
45.17260
|
84.43993
|
45.17769
|
84.42606
|
|
Section A
|
06/30/2008
|
6,917
|
45.17769
|
84.42606
|
45.19067
|
84.43478
|
|
Tin Bridge
|
07/16/2008
|
2,000
|
45.21945
|
84.43207
|
45.22417
|
84.42886
|
|
Trail 34 Elk Meadows
|
07/29/2008
|
2,000
|
45.23206
|
84.43020
|
45.23476
|
84.43539
|
|
Pine Grove Campground
|
07/16/2008
|
2,000
|
45.24269
|
84.44414
|
45.24459
|
84.44936
|
|
Webb Road Bridge
|
07/08/2008
|
2,000
|
45.26944
|
84.45882
|
45.27240
|
84.46152
|
|
McIntosh Landing
|
07/17/2008
|
2,000
|
45.28656
|
84.46729
|
45.28920
|
84.47102
|
|
Cudlip property
|
07/28/2008
|
1,500
|
45.30379
|
84.48633
|
45.30614
|
84.48807
|
|
Pigeon R. Road
|
07/17/2008
|
1,000
|
45.32937
|
84.49445
|
45.33167
|
84.49410
|
