By: Henry R. Quinlan
U.S. Fish and Wildlife Service
Ashland National Fish & Wildlife Conservation Office
Ashland, WI

November, 1999

---

CONTENTS

INTRODUCTION
PROJECT OBJECTIVES
METHODS
RESULTS

AGE
GROWTH
WEIGHT - LENGTH RELATION
MORTALITY
POPULATION ESTIMATES
COHORT ANALYSIS
SPAWNING LOCATION AND HABITAT USE
FECUNDITY
STOMACH ANALYSIS

DISCUSSION
RECOMMENDATIONS
REFERENCES
APPENDIX

---

ACKNOWLEDGMENTS

I thank the many individuals who have conducted this work and contributed to preparation of this report. This work was a cooperative effort of the National Park Service, Michigan, Minnesota and Wisconsin Departments of Natural Resources, Keweenaw Bay, Red Cliff and Grand Portage Bands of Lake Superior Chippewa, U.S. Fish and Wildlife Service, U.S. Geological Survey Biological Resources Division, Ontario Ministry of Natural Resources, and Trout Unlimited.

INTRODUCTION

Brook trout, Salvelinus fontinalis, that spend part of their life cycle in one of the Great Lakes have been termed Acoasters@ (Becker 1983). Since 1993, the U.S. Fish and Wildlife Service (Service) has coordinated and led efforts at Isle Royale, Michigan, in Lake Superior (Figure 1) to assess coaster brook trout status, distribution, movement, and abundance, and to collect gametes for broodstock development (Slade 1994, Slade 1995, Newman and Bast 1996, Quinlan 1999). Many individuals from fishery management and research agencies and Isle Royale National Park (ISRO) personnel have cooperated in these efforts. This report describes biological information from gamete collection and assessment activities conducted in fall 1996, spring and fall 1997, and fall 1998 with recommendations for brook trout management at Isle Royale.

PROJECT OBJECTIVES

1) Collect gametes for broodstock development and conservation aquaculture.
2) Gather biological and population structure data during gamete collection efforts.
3) Assess relative and absolute abundance of coaster brook trout.
4) Gather tissue for genetic analysis of coaster brook trout populations.
5) Describe coaster habitat and spawning locations.
6) Determine the contribution of stocked fish to the Siskiwit Bay population.

METHODS

Personnel and equipment traveled to and from ISRO aboard the National Park Service (NPS) vessel, Ranger III of Houghton, Michigan, the Voyaguer II of Grand Portage, Minnesota, and the USGS-BRD research vessel, Siscowet of Bayfield, Wisconsin.

Collection efforts occurred during the fall spawning period from late September to October in 1996, 1997, and 1998, and during June in 1997. Coaster brook trout were captured by fyke net, boat and backpack electrofishers, modified Windermere trap (Edwards et al. 1998), and conventional angling. Gear type used for each project varied depending upon objectives and location.

Within study or collection areas, collection effort occurred throughout the habitats known or suspected to contain coaster brook trout. Effort was concentrated in areas where coasters had previously been reported or captured and at water depths less than 10 feet (Slade 1994, Slade 1995, Newman and Bast 1996) (Figure 2). Capture efforts with gill nets in waters deeper than 12 feet have been unsuccessful in Tobin Harbor, Washington Harbor and Siskiwit Bay (Slade 1994, Newman and Bast 1996).

Fyke nets 4 feet deep x 6 feet wide with 1/4 to 1 inch stretch mesh and 3 feet deep x 6 feet long with 1/4 - 3/8 inch mesh, and modified Windermere traps 2 feet x 4 feet with 1/4 inch stretch mesh were fished overnight. Boat electrofishing was conducted with a Smith Root Incorporated Model GPP pulsed DC electrofisher unit. Control box settings ranged from 175 to 600 volts and 60 or 120 pulses per second, which produced an output between 2 - 3 amps. Boat electrofishing began at dusk and continued for 1-3 hours. Electrofishing took place along the shoreline of the main island and around islands in Tobin and Washington harbors at depths less than 9 feet.

An Advanced Backpack model AbP-3TM electrofishing unit was used to electrofish tributaries. Control box setting for the Big and Little Siskiwit rivers was 175 volts, 75 pps, 15% duty cycle. Tobin Creek setting was 350 volts, 68 pps, 14% duty cycle. A small tributary to Tobin Harbor, Moose Lake Outlet settings were 350-400 volts, 67-68 pps, and 15% duty cycle.

We electrofished Big Siskiwit River in an upstream direction beginning 800 yards upstream from the mouth. Distance sampled was 50 yards in 1997 and 200 yards in 1998. We electrofished Little Siskiwit River from the head of the estuary upstream 100 yards in 1997 and from the head of the estuary upstream to the falls (about 300 yards) in 1998. We electrofished Tobin Creek from the head of the estuary to the foot bridge, a distance of about 50 yards, and Moose Lake Outlet from the mouth to the upstream barrier, a distance of about 50 yards.  We used conventional angling equipment at Siskiwit Bay, Big Siskiwit River and Tobin Harbor.

At each net, trap, and backpack electrofishing location we recorded date, time of day, location, water temperature, depth, effort and species captured. During boat electrofishing, date, time of day, location shocked, water temperature, and effort were recorded.

We sorted all fish captured by species and counted the total up to at least 25. We measured total length of all brook trout to the nearest millimeter and checked for previous marks. Our crew weighed brook trout captured in June and October 1997 to the nearest gram. We determined the sex of brook trout captured in fall by expression of gametes. We checked all mature brook trout for spawning condition. We removed scale samples from a representative portion of the catch each year. We did not attempt to collect scale samples from fish that were being held or could possibly be held for spawning in 1998. All brook trout captured and immediately released were given a v-notch clip of either the dorsal, anal or caudal fin upon release. In all years except 1998, all recaptured fish received a second v-notch clip of the anal or caudal fin upon release. We placed brook trout to be spawned in pens until ripening or until our gamete collection effort was completed. We collected an adipose fin tissue sample from all fish spawned and placed it in a vial of buffer solution for future genetic analysis.

I estimated population size in Tobin Harbor using Schnabel and Peterson mark-recapture methods (Van Den Avyle 1993). Upper and lower 95% confidence intervals were calculated using the Poisson distribution (Ricker 1975).

I generated a Petersen mark-recapture estimate for coaster brook trout marked in June 1996 and recaptured in September and October 1996. The fish were marked with internal radio transmitters as part of a tracking study being conducted by the Service. Twenty fish > 13.2 inches were marked by surgical implant of radio transmitters in June. We sampled fish for marks in conjunction with gamete collection efforts in September and October 1996. The minimum growth of a fish marked in June and recaptured in October was 0.2 inch. Therefore, to ensure that only fish from the recapture sample large enough to be part of the June target population were used in the estimation, the minimum length of fish I included in the recapture pool was 13.4 inches.

Several assumptions for valid population estimates were not met. One marked fish grew only 0.2 inches. Data on growth of other coaster brook trout suggests that this is very slow growth for the time period at large. Therefore, it is likely that fish < 13.2 inches in June 1996 grew into the length range >13.4 inches by October 1996 and the assumption of no recruitment was not met. If recruitment to the recapture pool did occur the result would be an overestimation of population size. The assumption of a closed population was also not met. Tracking of radio tagged fish showed that they remain primarily in Tobin Harbor; however, one radio tagged brook trout briefly moved to Rock Harbor before returning to Tobin Harbor in fall 1996 (Lee Newman, Ashland NFWCO, personal communication).

I conducted a second Petersen mark-recapture estimate for coaster brook trout in Tobin Harbor in October 1996. We marked and released a total of sixty fish, the youngest of which was age 1, with an adipose fin clip during gamete collection efforts. The October 1997 survey served as the recapture sample. All fish greater than 9.8 inches, the minimum length of an age 2 fish in our recapture sample, were included in the recapture pool. Due to the duration of time at large the assumption of a closed population was not met.

I estimated the population size for age 1 coaster brook trout in Tobin Harbor in June 1997 using the multiple census Schnabel method (Ricker 1975). We used boat electrofishing gear to sample 6 miles of shoreline. Due to the presence of nesting loons in Tobin Harbor we were limited by ISRO as to the amount of shoreline sampled. The north shoreline was sampled from the mouth of Tobin Creek northeast to a point opposite the sea plane dock. The south shoreline was sampled from the mouth of Tobin Creek to a point 100 hundred yards northeast of Tallman Island. We alternated sampling between each section of shoreline on a daily basis to reduce stress to coasters and to allow mixing of marked fish.

We marked fish on days 1-3 and checked for marks on days 2-4. Catch composition dictated that only the population size of age 1 fish was estimated. Assumptions for population estimates were met.

I attempted to use the same procedure to estimate coaster brook trout population size in Washington and Grace harbors. We sampled nine miles of shoreline in Washington and Grace harbors with boat electrofishing gear. The area sampled included the shoreline of Grace Harbor, Washington Harbor and around Beaver Island. An insufficient number of coasters were captured in this area to estimate population size.

I calculated a Schnabel multiple census mark and recapture population estimate for fish in Tobin Harbor in October 1997. We captured fish in fyke nets and modified Windermere traps and marked them over a two day period with a v-notch anal fin clip. We checked for marked fish in net and trap collections over a two day period. Assumptions for population estimates were met.

We analyzed stomach contents from six brook trout mortalities. Stomach contents of three fish were identified using a Nikon SMZ-U dissecting scope at 10X power and weighed using a Mettler H32 balance. Total stomach weight and weight of individual contents were recorded. Stomach content of three fish were identified visually and recorded.

RESULTS

We captured a total of 863 brook trout at all locations at Isle Royale from October 1996 to October 1998. The majority of fish, 786 or 91%, were caught in Tobin Harbor. In the Siskiwit Bay area, 69 or 8%, were caught (60 in Big Siskiwit River, 8 in Little Siskiwit River and 1 in Siskiwit Bay) and in Washington Harbor, 8 (1%), were caught. Length ranged from 3.3 to 20.0 inches in Tobin Harbor (Figures 3-4), 2.4 to 14.1 inches in Siskiwit Bay area (Figure 5), and 3.8 to 5.9 inches in Washington Harbor. Location, gear type, catch, effort and catch per unit effort for each sample effort are shown in Table 1.

We recaptured 283 fish in October during the same time period. In Tobin Harbor 33% (260 fish) were captured more than once, while in the Siskiwit Bay area and Washington Harbor the percentages were 32% (22 fish) and 12% (1 fish), respectively.

Figure 3.  Coaster brook trout length frequency in Tobin Harbor, fall 1996-1998.

Figure 4.  Coaster brook trout length frequency in Tobin Harbor, June 1997.

Figure 5.  Coaster brook trout length frequency in Tobin Harbor, October 1997-1998.

Table 1. Description of survey effort and catch of brook trout at Isle Royale National Park, 1996-1998.

Year	Date	Location	Gear Type	Catch	Effort	Catch Rate
1996	Sept. 19 -  Oct. 8	Tobin Harbor	EF boat fyke net backpack EF angling	67 47 1 1	12.2 hrs 151 net nights 0.45 hrs 8 hrs	5.5 0.3 2.2 0.1
1997	June 17-20	Tobin Harbor	EF boat	93	5.8 hrs	16.0
	June 21-23, 25	Washington and Grace Harbors	EF boat	8	4.0 hrs	2.0
	Oct. 2-3	Siskiwit Bay and tributaries	fyke net mod. Windermere trap backpack EF	1 1 1	11 net nights 6 trap nights 0.32 hrs	0.1 0.2 3.1
	Oct. 5-7	Tobin Harbor	fyke net mod. Windermere trap	135 6	24 net nights 18 net nights	5.6 0.3
1998	Oct. 2-21	Tobin Harbor	fyke net angling	433 3	227 net nights 3 hrs	1.9 1.0
	Oct. 5-20	Siskiwit Bay and tributaries	fyke net backpack EF angling	59 7 0	120 net nights 2.3 hrs -	0.5 3.0 -

We collected gametes for broodstock development from coaster brook trout in Tobin Harbor in 1996 and 1998. Gametes were collected from 22 females and 25 males in 1996 and from 15 females and 16 males in 1998. Average length of brook trout spawned is shown in Table 2.

Sixty Tobin Harbor coaster brook trout held for spawning in 1996 were given an adipose fin clip upon release. We recaptured sixteen of these fish during the 1997 survey effort and 11 were caught during the gamete collection effort in 1998.

Despite substantial collection effort we collected no gametes in the Siskiwit Bay area in 1998. Below average precipitation during summer 1998 resulted in low water level and little or no flow in Big and Little Siskiwit rivers. These conditions persisted until mid October near the end of the sample period.

Table 2. Length of Tobin Harbor coaster brook trout contributing gametes to broodstock development.

Year		N	Mean Length (inches)	Max Length (inches)	Min Length (inches)	Std. Dev.
1996	Male	25 22	15.0 16.1	19.1 19.2	8.0 12.3	2.62 1.85
	Female
1998	Male	16 15	13.1 15.0	17.2 17.6	9.8 12.6	2.11 1.42
	Female

Fyke net effort and catch rate varied substantially among years within a location (Table 1). In Tobin Harbor net effort ranged from 24 nights in 1997 to 227 net nights in 1998, while in Siskiwit Bay net effort was 11 nights in 1997 and 120 nights in 1998. Coaster brook trout catch was not proportional to effort. Catch rate in nets was likely affected by the timing and location of net sets and other collection activity. In 1997 nets were set for a three night period which corresponded to a period of active spawning. In 1996 and 1998 the capture effort was extended for more nights over a greater time period which included days of lesser spawning activity. Catch in 1996 and 1998 was also influence by gamete collection efforts in which fish were held in pens for spawning and were not available to recapture.

Figure 6 shows spawning condition of female coaster brook trout by date captured and dates when fish were spawned for gamete collection. The earliest date a ripe female coaster brook trout was captured in Tobin Harbor was September 29. In the Siskiwit Bay area, October 12 was the earliest date a ripe female was captured in 1998. In 1995 females from the Siskiwit Bay area population were spawned on September 30, October 11, 15, and 19 (Newman and Bast 1996).

Figure 6.  Spawning condition by date of female coaster brook trout in Tobin Harbor 1996-1998 and dates spawned.

In Tobin Harbor the proportion of ripe females approached that of green females by October 11. Few spent females were captured. Spent females were only captured in 1997 when no gamete collection took place. While influenced slightly by dates of available transportation off Isle Royale, the majority (76%) of fish spawned for gamete collection were spawned after October 15. Peak spawning of fish being held occurred October 21 and 22.

Four female coaster brook trout with right ventral (pelvic) fin clips were captured in 1998, two in the Big Siskiwit River and 2 in Little Siskiwit River. These were the only ripe female brook trout captured in the Siskiwit Bay area in 1998. Because these fish were products of the gamete collection effort in 1995 no gametes were collected from these fish. The capture of these fish shows that stocked yearlings survived and return to streams where the parental stock originated. These fish appeared ready to spawn and contribute gametes to the Siskiwit Bay area population.

Sex ratio of males to females was 1:0.9 in Tobin Harbor. In Siskiwit Bay the ratio of males to females was 1:1.4. In Tobin Harbor the percentage of fish mature at length and ratio of mature males to females is shown in Figure 7. Age at maturity is shown in Figure 8. Sex determination was not possible for fish collected in Washington Harbor and Tobin Harbor in June 1997.

Figure 7.  Percentage of mature fish at length and ratio of mature males to females from Tobin Harbor, 1996-1998.

We collected adipose fin tissue samples from brook trout in Tobin Harbor 1996 and 1998 and from Washington Harbor in 1997. Samples were sent to USGS -BRD Great Lakes Science Center (GLSC), Ann Arbor, Michigan, for genetic analysis.

The parasitic copepoda, Salmincola edwardsii was present on the gill and/or base of fins of coaster brook trout in Tobin Harbor. The number and percentage of brook trout in Tobin Harbor with this parasite increased each year. In 1996, the parasite was detected on one fish (0.9%), in June 1997, four (4.6%), in October 1997, fifty-five (45.1%) and in 1998, on ninety-six (46.4%). After June 1997, a concerted effort was made to check for presence of gill and fin parasites which may have increased the detection rate. In 1996, all fish held for gamete collection were kept in pens at the sea plane dock in Tobin Harbor. In 1998, fish with external parasites were kept in pens at a location removed from the sea plane dock to reduce potential for transfer of parasites to uninfested fish.

AGE

We read scale samples for age determination from 197 brook trout (177 from Tobin Harbor, 15 from Siskiwit Bay area and 5 from Washington Harbor). Removal of scales from the area midway between the dorsal fin and lateral line, the standard location for trout (Nielsen and Johnson 1992), was difficult for males in spawning condition. Severe pressure needed to be applied to remove scales, therefore to avoid injury fewer samples were taken from males in spawning condition. Difficulty removing scales from male brook trout in fall was also reported by Tillma 1997.

A total of 7 age groups (0-6) were present in fall samples from Tobin Harbor. Older age groups (5-6) were represented by few fish. Length frequency histograms show good definition for age 0 and age 1 fish only (Figures 3 and 4). Length frequency modes beyond age 1 are not evident. In the Siskiwit Bay area four age classes were present 1-3, and 5. Ages 3 and 5 were each represented by one individual. Two age classes were present in Washington Harbor. Two were yearlings (based on length) and five were age 2.

An age-length key was used to assign age to fall collected coaster brook trout from Tobin Harbor for mortality and cohort analysis. The age-length key was developed from 139 scale-aged fish collected in fall from Tobin Harbor (Appendix). Several length groups were not represented by scale-aged fish. For these length groups, I used the same percentage as the nearest represented length group.

Sex, condition and age were determined for 82 brook trout from Tobin Harbor. The percentage mature by sex of scale-aged fish from Tobin Harbor is shown in figure 8. Males are shown to mature at a younger age than females. A small percentage of female fish mature at age 2, while the majority of female fish mature at age 3. By age 4 all are mature. All of the scale-aged males had matured as yearlings. The absence of males at age 5 and 6 may be a result of sample size since few samples were taken from large spawning males.

Figure 8.  Percentage of fish mature at ages 0-6 in Tobin Harbor, fall 1996-1998.

Four females stocked in Siskiwit Bay and captured in the Big and Little Siskiwit rivers in 1998 were mature at age 2. An age 3 female was determined to be mature but not ripe on October 2, 1997.

A paired scale and otolith sample was available from one coaster brook trout. Scale age was one year younger (6) than otolith age (7). Otolith samples from 3 other fish were examined. An 18.4 inch male was aged at 6, while the otoliths from the other two were unreadable due to vaterite deposit. Age of the two fish interpreted from otoliths were two of the oldest fish aged by either method.

We captured four known-age fish in the 1998 Siskiwit Bay effort. These fish were stocked as yearlings in 1997 after having been reared at the Keweenaw Bay Indian Community hatchery and Iron River National Fish Hatchery. Scale samples were taken from all four of these fish. Age interpreted from these scales matched known age for one fish and overestimated age for three. Events related to movement between hatcheries with different water supplies (groundwater vs. surface water) and feed conditions or introduction into the wild may have resulted in checks misinterpreted as annuli on these scales.

GROWTH

Figure 9 shows mean length at age for coaster brook trout populations in Tobin Harbor and the Siskiwit Bay area. Mean length at ages 0-2 is similar among all populations and years. The single fish representing age 5 from Siskiwit Bay grew slower than same age fish in Tobin Harbor. Growth decreased at age 6 in both 1996 and 1997. Comparison of mean length at age among years was conducted using the non-parametric Kruskal-Wallis and Mann-Whitney tests. No significant differences in mean length at age were detected at P > 0.05.

Figure 9. Mean length at age for fall caught coaster brook trout from Tobin Harbor 1996-1998, and Siskiwit Bay area 1998.

A direct measure of growth is available for only one fish from Tobin Harbor. An age 1 fish marked with an adipose fin clip in October 1996 measured at 7.9 inches was recaptured in October 1997 at 11.2 inches. The fish grew 3.3 inches in one year.

Indirect measures of growth have been calculated from adipose fin clipped fish. Mean length of adipose fin clipped fish in October 1996 was 15.3 inches (n=60). Adipose fin clipped fish averaged 17.5 inches in both years 1997 (n=16) and 1998 (n=11) (Table 3). From 1996 to 1997 mean length increased 2.2 inches. Mean length of adipose clipped fish caught in 1998 showed no change from 1997. The number of fish and the individual fish captured differed from year to year, so these results provide an inference of growth and not a direct measure.

Tables 4 and 5 show growth as measured by the increase in mean weight (ounces) and length (inches) at age of coaster brook trout caught in Tobin Harbor from June 1997 and October 1997. Increase in mean weight was greatest for age four fish. Increase in mean length of age 1 fish was 2.5 inches. Mean length increased most for age 2 fish.

Table 3. Length comparison of adipose fin clipped coaster brook trout from Tobin Harbor in 1996 and recaptured in 1997 and 1998.

Year	Sex	Number	Mean length (inches)	Change in mean length from previous year (inches)
19961	male	25	14.9	-
	female	35	15.6	-
	both	60	15.3	-
19972	male	9	17.1	2.2
	female	7	18.0	2.4
	both	16	17.5	2.2
19982	male	3	19.0	1.9
	female	8	17.0	-1.0
	both	11	17.5	0.0

1 Year of adipose fin clip
² Year recaptured

Table 4. Increase in average weight (ounces) at age for Tobin Harbor coaster brook trout in 1997.

Age	Mean weight June 1997 (ounces)	Number sampled	Mean weight October 1997 (ounces)	Number sampled	Increase in mean weight (ounces)
0	-	0	0.5	6	-
1	1.2	31	3.8	20	2.6
2	2.5	1	7.4	12	4.9
3	-	0	14.0	1	-
4	19.7	3	33.9	4	14.2
5	36.5	1	41.1	1	4.6
6	-	0	38.5	1	-

Table 5. Increase in mean length (inches) at age for Tobin Harbor coaster brook trout in 1997.

Age	Mean length June 1997 (inches)	Number sampled	Mean length October 1997 (inches)	Number sampled	Increase in mean length (inches)
0	-	0	4.7	6	-
1	6.0	31	8.5	20	2.5
2	7.0	1	10.7	12	3.7
3	-	0	12.8	1	-
4	15.9	3	16.9	4	1.0
5	18.1	1	19.4	1	1.3
6	-	0	18.9	1	-

WEIGHT - LENGTH RELATION

The length-weight relation of coaster brook trout in Tobin Harbor (Figure 10) was based on all available data (n= 69) from fall samples and includes fish from 1993, 1997, and 1998. The length-weight relation shown here includes both mature and immature fish. Insufficient data are available to examine a separate length-weight relation for male and female fish. The relation likely varies during the year especially for mature females. No spent female fish were included in the analysis. The equation generated by linear regression of the logarithms of length and weight was log W = -2.38 + 3.15 log L, where L = total length in inches and W = weight in ounces.

Figure 10. Length-weight relation of coaster brook trout in Tobin Harbor 1993, 1997 and 1998 and Nipigon River 1987-1990. The curve represents calculated weight at length.

A length-weight relation of coaster brook trout in the Nipigon River was also generated. Data are from angler caught fish from Nipigon River below Alexander dam during the four year period 1987-1990 (Ontario Ministry of Natural Resources, Nipigon District, unpublished data). All available data was used (n=61). The length-weight equation for Nipigon River fish is log W = -2.24 + 3.23 log L, where L = total length in inches and W = weight in ounces. A conversion was developed between fork and total lengths in centimeters for Nipigon Lake coaster brook trout of both sexes (n=271). Converted to inches the regression equation is total length = 0.336 + 1.021 (fork length).

The Nipigon River fish were heavier for a given length than those in Tobin Harbor. However, few data points were available from large fish in Tobin Harbor since we limited the handling and did not weigh fish to be spawned. Conversely, data provided by anglers in the Nipigon River contained few records for smaller fish. The b values greater than 3.0 suggests that coasters become more rotund as length increases.

MORTALITY

A catch curve was calculated by combining 1996-1998 fyke net data from Tobin Harbor (Figure 11). Total annual mortality rate was determined from the descending limb of the catch curve where Y = ln number caught and X = age. Total annual mortality was estimated at 0.43 (Y = 5.8 - 0.56(X)).

Figure 11.  Coaster brook trout catch curve for fyke net captures from Tobin Harbor 1996-1998.  Total annual mortality rate equals 0.44.

Total annual mortality of coaster brook trout in Tobin Harbor was not high compared to other rates described for coaster brook trout in Lake Superior (Tillma 1997) and 12 fast growing wild lentic populations in Algonquin Park, Ontario (Quinn et al. 1994). Tillma found annual mortality rates in Minnesota tributaries to Lake Superior to be 0.84 while those of Algonquin Park ranged from 0.48 to 0.83 with a mean of 0.70 (Quinn et al. 1994).

POPULATION ESTIMATES

I calculated several population estimates for coaster brook trout in Tobin Harbor. Low numbers captured and recaptured in Siskiwit Bay and Washington Harbor preclude estimation of population size in these areas. Four estimates of population size have been made for coaster brook trout in Tobin Harbor (Table 6). Estimates range from 505 (216-1,579) age 1 fish in June 1997 to 228 (143-381) age 2 and older fish in October 1996. High numbers of recaptured fish 229 of 436 (53%) throughout the 20 day capture effort in 1998 support low population estimates.

Due to differences in the target population for each estimate the numbers are not directly comparable to each other. However, the estimates provide a general description, being of similar magnitude, of the number of coaster brook trout inhabiting Tobin Harbor during these years.

Table 6. Population estimates for coaster brook trout in Tobin Harbor. Tobin Harbor is 660 acres in size.

Marking Period	Recapture Period	Target Population	Population Estimate	95% Confidence Limits (Poisson)	Population Density
10/97	10/97	All fish	308	198-509	0.5 / acre
6/97	6/97	Age 1	505	216-1,579	0.8 / acre
9-10/96	10/97	> Age 1	228	143-381	0.4 / acre
6/96	9-10/96	> 13.2"	241	98-603	0.4 / acre

COHORT ANALYSIS

Table 7 shows the number of fish captured by age (scale age and assigned age) and year in Tobin Harbor. No trend is apparent over years or at a particular age. The variation in survey effort and gear type (boat electrofisher, fyke net, modified Windermere trap, backpack electrofisher and angling) between years confounds these results and likely masks trends within a cohort.

Table 7. Catch at age and catch rate of coaster brook trout in fall samples from Tobin Harbor 1996-1998. Age based on scale interpretation and age-length key (Appendix). Recaptured fish not included.

Tobin Harbor fyke net captures Catch at Age (years)

Year	0	1	2	3	4	5	6	Total
1996	16	11	2	7	8	1	0	45
1997	15	49	33	3	5	4	7	116
1998	8	32	77	35	41	7	6	206
Total	39	93	112	45	58	12	8	367

Tobin Harbor fyke net effort

Year	Net Nights
1996	151
1997	24
1998	221

Tobin Harbor Fyke Net Catch Rate Catch Rate (number/night) at Age

Year	0	1	2	3	4	5	6	Catch Rate
1996	0.10	0.07	0.01	0.05	0.05	0.01	0	0.30
1997	0.63	2.04	1.38	0.13	0.21	0.17	0.29	4.83
1998	0.04	0.14	0.35	0.16	0.19	0.03	0.03	0.93

Tobin Harbor Catch by all Gear Types (fyke net, electrofishing, angling, modified Windermere trap)
Catch at Age (years)

Year	0	1	2	3	4	5	6	Total
1996	16	27	12	15	27	9	4	110
1997	20	50	33	3	5	4	7	122
1998	8	32	77	35	42	7	6	207
Total	44	109	122	53	74	20	17	439

SPAWNING LOCATION AND HABITAT USE

We captured coaster brook trout over a wide variety of substrate at all locations. Substrate type included boulder, rubble, cobble, gravel, sand, clay and silt. While several locations produced a large number of fish in our nets, substrate does not appear to be the determinant. In Tobin Harbor high catches seem more likely related to movement from deep water (>10 feet) to nearshore areas.

In all years 1996-98 female coasters captured in fyke nets in Tobin Harbor were ripe and released eggs upon handling. A pair of spawning size adults was observed on a redd of sand, pea gravel, and sticks in October 1997. The redd was located about 5 feet (1.5 m) from the shore at a depth of 1.6 feet (0.5 m). Four to six smaller coasters, presumably young males, were in the immediate vicinity of the redd. The sighting of a pair of mature coasters on a redd in Tobin Harbor and the capture of ripe females suggests that lake spawning is occurring. We observed no spawning activity or ripe fish in tributaries to Tobin Harbor.

Water temperature (50 ^o F) and conductivity (100.6 mico Siemens/cm) measured in October 1998 within the area of the 1997 redd was the same as that of nearby and deeper (>10 feet) waters. Spring fed streams were detected along the north shore of Tobin Harbor. Water temperature in Moose Lake outlet and an unnamed tributary in SW 1/4, SE 1/4, Sec. 33, T. 67N, R. 33W, was 46^o F on October 4, while surface water temperature in Tobin Harbor was 54^o F at the same time and day. Nets set in these areas caught coaster brook trout, although no spawning activity was observed.

No redds or spawning activity were observed in the Siskiwit Bay area. A ripe male was captured in Siskiwit Bay.

FECUNDITY

Fecundity information exists for a single female, length 16.3 inches and age 6. This fish contained 1,530 eggs after having been spawned for gamete collection. Measured egg take during spawning was estimated at 300 eggs. Total egg count was about 1,800.

STOMACH ANALYSIS

We examined stomach contents from six coaster brook trout in Tobin Harbor. We brought three stomachs back to Ashland NFWCO for analysis and examined three in the field. Fish size and stomach content information is shown in Table 8. The presence of salmonine eggs indicates that some predation by coaster brook trout on their own eggs may be occurring.

Table 8. Biological information and stomach contents of coaster brook trout in Tobin Harbor.

Fish Length (inches)	Fish Weight (ounces)	Stomach Contents (number)	Content Weight (ounces)
8.2	2.5	Brook stickleback (1) Miscellaneous fish parts Corixidae (4)	0.0036 0.0117 0.0019
7.7	2.4	Trichoptera (1) Miscellaneous fish parts	0.0014 0.0145
18.4	36.2	Salmonine eggs (8)	0.0013
8.6	-	Trichoptera case (1) Corixidae (several)	- -
9.7	-	Plecoptera (1) Snail (4) Unidentified insect (1)	- - -
9.6	-	Amphipod (1) Corixidae (several) Miscellaneous fish parts	- - -

 

DISCUSSION

Based upon preliminary genetic analysis (Mary Burnham-Curtis, USGS-BRD, Great Lakes Science Center, personal communication), apparent life history differences (shoreline vs. stream spawning), the limited range of radio tagged fish at Isle Royale (Lee Newman, Ashland NFWCO personnel communication), and no evidence of interchange between stocks which are separated by a minimum distance of 25 miles of shoreline, the three Lake Superior coaster brook trout populations at Isle Royale are considered discrete.

We have found that coaster brook trout in the Siskiwit Bay area are anadromous, entering the Big and Little Siskiwit rivers to spawn. Siskiwit Bay is exposed to winds from the north, west and east and high energy wave action make shoal or shoreline spawning unlikely. The Tobin Harbor stock shows evidence of shoreline spawning through capture of ripe females and observation of adult fish in a redd along shore. Electrofishing tributaries and nets set at the mouth of three tributaries to Tobin Harbor have produced no evidence of anandromous coaster brook trout. The tributaries to Tobin Harbor have little suitable brook trout spawning habitat. The substrate in Tobin Creek, the largest tributary, is composed entirely of clay, silt and sand. The outlets of Moose Lake and Hidden Lake are short, less than 50 yards, and substrate is predominantly rock, rubble and boulder. It is unknown whether coaster brook trout in Washington Harbor are riverine or lacustrine spawners. Two tributaries, Washington Creek and Grace Creek, are known to support riverine brook trout populations and have been reported, but not confirmed, to support runs of anandromous coaster brook trout (Lagler and Goldman 1982 and Slade 1995).

Other locations around Isle Royale are presumed to support small populations of coaster brook trout. ISRO staff occasionally receive reports from anglers of coaster brook trout being caught in other embayments. The presence, size and structure of other remnant populations has not yet been examined.

During the spawning season crews have observed that in both natural and hatchery environments, that the body of female brook trout tended to have a cylindrical shape, silvery color, and a smaller proportioned head than males, which had a more compressed dorso-ventral shape, larger head and jaws, and darker coloration.

It has been found that males mature at a younger age and at smaller lengths than females. Therefore, the percentage of males in the population should exceed females rather than the near 1:1 sex ratio observed. Possible explanations are that males grow faster or are more aggressive than females and therefore make up a disproportionate percentage of angler harvest.

Reasons for the observed decline in mean length at age 6 shown in Figure 9 are speculative. Angler harvest may be selective for fish about 20 inches. The harvest of these larger fish results in survival of only the slower growing fish at ages 5 and 6. It is also possible that bias in scale-age interpretation is greater at older ages and aging error could be a cause.

Comparison of mean length at age of Isle Royale coaster brook trout to Nipigon River coaster brook trout indicates that length at age is similar for all populations (Table 9 and Figure 9).

Table 9. Mean length (inches) at age of coaster brook trout at 3 locations in Lake Superior. (T.H. = Tobin Harbor, 
S.B. = Siskiwit Bay area and N.R. = Nipigon River).

	T. H. 1998			T.H. 1997			T. H. 1996			S. B. 1998			N. R. 1987-1990*
AGE	Mean	N	SD	Mean	N	SD	Mean	N	SD	Mean	N	SD	Mean	N	SD
0	-	-	-	4.6	8	0.5	4.4	14	0.4	-	-	-	-	-	-
1	7.7	3	0.8	8.5	20	1.0	8.7	19	0.9	8.0	6	1.0	8.8	1	-
2	10.0	13	0.7	10.6	14	0.9	11.5	5	1.2	11.2	6	1.0	14.1	10	2.8
3	11.4	6	0.2	13.4	2	3.1	12.9	8	0.8	-	-	-	15.5	10	2.5
4	-	-	-	17.8	4	1.0	15.8	12	1.2	-	-	-	16.4	2	2.0
5	-	-	-	19.4	1	-	17.5	5	1.1	14.1	1	-	19.8	2	3.8
6	17.1	3	0.7	18.4	1	-	16.6	2	0.6	-	-	-	-	-	-

*Measurements converted to total length from fork length (total length = 0.336 + 1.021 (fork length)).

Few spent females have been captured at Isle Royale (Figure 5). The low number of spent females may in part be due to our gamete collection efforts. Nearly all ripe females captured are held for spawning. Fish held in pens may release their eggs later than fish at large. Most fish held in pens were spawned near the end of, or after capture efforts had ceased (Table 1 and Figure 5) and were therefore not available to capture after being released and spawning in the wild. In Tobin Harbor peak spawning of fish being held occurred October 21 and 22. Capture efforts ceased two weeks prior to peak manual spawning in 1996. In 1998, capture efforts continued until the last day of manual spawning. If spawning activity is normally distributed around the peak date, it is possible that late spawning fish have not been sampled.

Neither scales nor otoliths have been validated for age determination of coaster brook trout. Scale checks as a result of emigration from the river to lake environment or changes in forage may confound scale-age of coaster brook trout. Until wild known age fish have been marked and recaptured, I anticipate difficulty in accurate scale-age analysis. There is debate in the literature regarding accuracy of scale-age interpretation for brook trout (Power 1980).

Age four fish in Tobin Harbor showed the greatest increase in average weight from June to October 1997 (Table 4). This may correspond to maturation and gonad development which is likely greatest for age four fish, most of which are entering their second spawning season (Figure 8).

The lower mortality rate in Tobin Harbor (0.43) compared to Minnesota tributaries to Lake Superior (0.84) may be due to environmental conditions. Environmental stability afforded by the confines of Tobin Harbor and relatively stable water levels of Lake Superior contrast sharply with the dynamic nature of Minnesota=s tributaries to Lake Superior. The relative stability of Tobin Harbor likely reduces morality due to natural causes as compared to riverine environs where natural mortality is typically high. At present no data exists to partition rates of natural or fishing mortality in Tobin Harbor.

The presence and increased incidence of the copepod parasite Salmincola edwardsii on coaster brook trout in Tobin Harbor is a concern. While infested fish captured appear healthy and vigorous these parasites are known to ultimately kill the host (Becky Lasee, LaCrosse Fish Health Center, personnel communication). At present the impact of this parasite on individual fish or the population is unknown.

Low water and lack of flow prevented adequate assessment of the Siskiwit Bay coaster brook trout population in 1998. The mouth of Big Siskiwit River was closed by sand and little flow was present in Little Siskiwit River until heavy rain raised water levels on October 17. Based on the dates of the observed spawning run in 1995, when 90% of the fish were spawned prior to October 17 (Newman and Bast 1996), low water conditions during the 1998 spawning run may have prevented coaster brook trout from entering these rivers to spawn.

Quality coaster brook trout habitat is present in all three locations sampled at Isle Royale. The threat to this habitat is minimal due to National Park status. To date only one redd has been located in Tobin Harbor. Coaster brook trout in the Nipigon River and Bay have been found to spawn in very localized areas over groundwater upwellings (Rob Swainson, Ontario Minister of Natural Resources, Nipigon District, personal communication). Spawning areas in Nipigon River were detected through telemetry. One attempt at telemetry for detection of spawning sites at Isle Royale was unsuccessful.

Tobin Harbor holds the largest known population of coaster brook trout at Isle Royale, and in U.S. waters of Lake Superior. Stocks sizes at Siskiwit Bay and Washington Harbor have not been determined mainly due to too few fish captured. In Tobin Harbor the population is self-sustaining. Young of the year coaster brook trout were captured in all years. Age structure of coaster brook trout in Tobin Harbor shows 5-7 age classes present. This population exhibits the range of age classes recommended of future rehabilitated stocks in Lake Superior (Newman, Dubois, and Halpern 1999).

While conditions in the Big and Little Siskiwit rivers appeared unfavorable in 1998 there is evidence of successful reproduction from 1995-1997 based on the presence of young-of-the-year, age 1 and age 2 fish. Four year classes were present in fall 1995 (Newman and Bast 1996) and 1998. This population does not exhibit the range of age classes recommended in the Brook Trout Rehabilitation Plan for Lake Superior (Newman, Dubois, and Halpern 1999).

At present the Siskiwit Bay area coaster brook trout population appears to be depressed and smaller than the Tobin Harbor population. Anecdotal reports from anglers on staff at ISRO and Wisconsin DNR, and others indicate that large (about 17 inches and greater) coaster brook trout were present, and perhaps abundant, in near shore waters of Siskiwit Bay as recent as 5 - 20 years ago. Habitat conditions have changed little during this time period, but fishing pressure has increased as a result of improved boat docking facilities at Siskiwit Bay campground near the mouth of Big Siskiwit River and Hay Bay campground near the Little Siskiwit River. Regulation changes described below may reverse what appears to be a declining trend in the Siskiwit Bay area stock.

Capture of yearling and age two coaster brook trout in Washington Harbor in June 1997 indicates that natural reproduction is occurring in the harbor or that brook trout are emigrating from Washington or Grace creeks to Lake Superior as young-of-the-year or spring yearlings. We captured no adult fish in the Washington Harbor area in June 1997. However, the low number of adult fish caught in Tobin Harbor in June 1997 and their presence in fall 1997 suggests that adult fish were not vulnerable to boat electrofishing gear in June 1997. At least one adult coaster was present in Washington Harbor in spring 1998, when an angler caught a 25 inch coaster brook trout (Larry Kangas, Isle Royale National Park, personal communication).

From 1994 to 1999 regulatory changes by ISRO and Michigan DNR have increased the minimum size limit, reduced daily bag limits and closed fishing during spawning season. Current Michigan DNR regulations at Isle Royale in waters of Lake Superior include an 18 inch minimum size limit, a daily bag limit of 1 per day and closure of the fishing season from the day after Labor Day to April 30. ISRO regulations for tributaries for to Lake Superior are minimum size limit of 7 inches, daily bag limit of 2 of which only 1 can be greater than 12 inches and season closure from August 15 to the last Saturday in April.

Based upon mean length at age (Figure 9) and age at maturity (Figure 8) these regulatory changes will allow coaster brook trout the opportunity to spawn more than once prior to recruiting to the sport fishery as recommended in the Brook Trout Rehabilitation Plan for Lake Superior (Newman, Dubois, and Halpern 1999). Current Michigan DNR and ISRO sport fishing regulations appear appropriate to ensure the long term viability of coaster brook trout at Isle Royale.

RECOMMENDATIONS

Since Fish and Wildlife Service led surveys began in 1993, much has been learned about coaster brook trout populations at Isle Royale. However, questions remain, and progress has been slow in some areas. The following list of recommendations identifies actions needed to further our understanding and management of coaster brook trout at Isle Royale.

Document the presence or absence of other remnant stocks around Isle Royale and determine their size and structure.

Identify, describe, quantify and protect coaster brook trout spawning locations in rivers and embayments.

Identify groundwater discharge areas at Isle Royale to determine if coaster brook trout are seeking out these areas for spawning.

Develop monitoring plans to evaluate changes in the size and structure of coaster brook trout populations in Tobin Harbor, Siskiwit Bay and Washington Harbor.

Rehabilitate the Siskiwit Bay population through stocking of Siskiwit Bay strain fingerlings and yearlings. Consistent with the Rehabilitation Plan for brook trout in Lake Superior, the primary goal for this population is to ensure its long term sustainability through the presence of 6 or more age groups (0-5) and a spawning population that exhibits densities sufficient to ensure a viable gene pool.

Describe and quantify nursery, juvenile and adult habitat at Isle Royale.

Assess the abundance and life history characteristics of the coaster brook trout population in Washington and Grace harbors.

Conduct angler education and encourage practice of catch and release angling to assist rehabilitation efforts and ensure the long term viability of coasters at Isle Royale.

Mark coaster brook trout which will provide data on movement and growth of coaster brook trout and allow validation of scale-age interpretation.

REFERENCES (click here)

APPENDIX (click here)

---