|
 |
by: Lee E. Newman U.S. Fish and Wildlife Service Ashland National Fish & Wildlife Conservation Office 2800 Lake Shore Drive Ashland, WI54806 |
|---|
Abstract.- Biotelemetry was employed to determine home ranges and movement patterns of 17 coaster brook trout Salvelinus fontinalis in Lake Superior and in tributary streams on Isle Royale. Transmitters were designed and uniquely programmed for this application, and were surgically implanted in sub-adult or pre-spawn fish ranging from 335 mm (13.3 inches) to 462 mm (18.2 inches) total length. Locations were monitored for 12 months. The maximum distances from release point recorded for individuals ranged from 0.25 km to 6.4 km. (mean=3.59 km). Movement was lowest during high light periods of the day. Results suggest that adults in this population do not enter streams to spawn.
Introduction
In the 19th and the early part of the 20th centuries brook trout Salvelinus fontinalis were a highly sought after game species in Lake Superior and in tributary spawning streams. These lake dwelling brook trout (locally referred to as "coasters"), along with the lake trout Salvelinus namaycush were the only two native salmonid species found in Lake Superior. Becker (1983), defined a "coaster" as any brook trout that spends part of its life in the Great Lakes, a definition that potentially includes both stream and lake spawning populations. The coaster brook trout was once an abundant brook trout stock along much of the shoreline of Lake Superior and provided an excellent fishery (Shiras 1935, Roosevelt 1865). At least 120 streams were probably inhabited by coasters (Newman and DuBois 1997).
During the latter part of the 19th century the entire Lake Superior fishery was subjected to heavy exploitation, and the coaster, which used a nearshore habitat, undoubtedly suffered severely from the intense sportfishing efforts of the day and to a lesser degree, from commercial net fishing. During this period and in later years land use practices such as logging, road development, and mining, and the introduction of Pacific salmonids (coho Onchorynchus kisutch, king, Onchorynchus tshawytsclw, and pink, Onchorynchus gorbuscha, and rainbow trout Onchoryhychus mykiss) and the European brown trout Salmo trutta may have also influenced the decline of the coaster (Newman and DuBois 1997).
By the early 1990's, only a few remnant populations of coasters were still known to exist in Lake Superior (Newman et al. 1999b). Today the few, small remaining populations in Lake Superior occur in remote areas isolated from large human populations, in areas where they are protected by private ownership of spawning streams or where they reproduce in large river systems. One of these known, viable populations occurs in the Tobin Harbor area of Isle Royale, Michigan.
Other than some limited information collected in the Nipigon River and Nipigon Bay in Ontario, there is no data reported concerning the movements and habitat use patterns of native coasters in Lake Superior or in the tributary streams. In the Nipigon River coasters spawn exclusively over upwellings of groundwater (Rob Swainson, Ontario Ministry of Natural Resources, Nipigon, personal communication). Curry and Noakes (1995), have also reported on the coasters' preference for spawning on upwellings of groundwater in Canadian shield waters.
In the past decade interest in, and support for restoration of coaster brook trout in Lake Superior has increased (Newman and DuBois 1997). The rehabilitation plan for brook trout in Lake Superior (Newman, et al. 1999b), explains that fisheries management agencies need to understand the distribution and habitat use patterns of coasters to guide efforts to manage existing remnant stocks and to reintroduce stocks in habitats where they have been extirpated. In 1996, the Ashland National Fish & Wildlife Conservation Office (NFWCO) of the U.S. Fish and Wildlife Service (USFWS) and the Isle Royale National Park (IRNP) teamed with state and tribal agencies to implement a radio telemetry study of the coasters of Tobin Harbor in order to learn more about the movements and home ranges of this important, native population.
Study Area
The study area focused primarily on Tobin Harbor, but included all of the North East end of Isle Royale, MI (Figure 1), and the nearshore waters of Lake Superior. Tobin Harbor is a protected bay that is 7.5 km in length and has a maximum width of 0.8 km. It includes a total area of 280.5 hectares including 18.0 hectares of islands. Water area is 262.5 hectares. Shorelines are mostly steep and rocky with some sand and gravel flats occurring near the inlet of Tobin Creek at the head of the bay. The maximum depth of the bay is about 20 m.
Objectives
To develop baseline data on the movements, ranges, habitat use patterns and spawning behavior of the Tobin Harbor population of coaster brook trout at Isle Royale, MI.
Methods
Capture and tagging of 20 coasters in this study was done on May 19 and 20, 1996 (Table 1), in the Western end of Tobin Harbor (Figure 1). Capture was done by Pulsed DC boat electrofishing at night. Of the initial 20 fish tagged, 3 produced no data after 5 days. These 3 individuals may have died, left the study area or their transmitters may have malfunctioned. All of the remaining 17 fish tagged produced data for slightly over 12 months.
The fish that were tagged ranged in size from 335 mm (13.2 inches) total length (TL), to 462 mm (18.2 inches). The mean length was 396 mm. Fish less than 335 mm were not used because they were thought to be too small to survive tagging and still behave normally. All fish tagged were anesthetized with MS 222 before surgically implanting the tags. When the fish were sufficiently anesthetized, a surgical scalpel was used to make an incision about 15 mm long, down the midline of the belly, about 30 mm behind the pectoral fins. The initial incision went through the skin and partially through the abdominal muscle layer. The tip of a hemostat was then inserted and the back of the jaws used to complete a blunt incision through the remaining muscle layer into the body cavity. The radio transmitter was then inserted into the body cavity. After the transmitter was inserted, 2 or 3 sutures (4/0 monofilament) were used to close the incision with the antenna protruding. The final suture was also wound around the antenna to help "lock" the antenna in place, and to prevent movement within the incision, and to insure retention of the transmitter. Following surgery, the fish were placed in a recovery tank of fresh lake water until they were actively swimming in an upright position. All tagged fish were then returned to the lake at a single release point near the seaplane dock (Figure 1) which is in the area where they were originally captured. One week after tagging, seventeen tagged coasters were actively moving and their transmitters were transmitting recognizable signals.
For this study, it was important to monitor individual fish for a full year, but most radio tags are too heavy or large for use in brook trout as small as 280 mm. Winter (1978), suggested a maximum transmitter weight of 1.25% of the body weight of the fish. To maximize the life span and still keep the total weight under a goal of 9 gm, we selected programmable units (Struthers, et al, 1996) manufactured by Advanced Telemetry Systems (ATS) Inc., Isanti, MN. The units transmit on unique and individually identifiable frequencies (48-50 MHz) separated by 20 kHz. The transmitters were preprogrammed to transmit a signal for 3.5 days on, then 3.5 days off for 182 days, then off for 50 days, then on for 8 hours per day for the remaining battery life. In effect, this provided for tracking from Monday morning at 0800 hr. to Thursday at 2000 hr., from May 20 to November 20. The transmitters were then off until January 10, when they came on again and ran from 0800 hr. to 1600 hr. daily for the remaining battery life. The projected battery life of 120 days of continuous operation was thereby extended to more than 375 calendar days.
During the open water months, the fish were located by boat once a week as consistently as the challenging weather and wave conditions of Lake Superior permitted. Tracking was done with the ATS R2000 receiver and loop antenna, and in some close range applications, a modified "ping pong paddle" or "stub" antenna was used to provide a higher degree of precision to locations. In addition, personnel were able to locate tagged fish and search for outliers by fixed wing aircraft on 3 occasions during the open water season. Signals could usually be received on the ground or water at ranges from 200 to 600 m and at ranges up to 2.5 km by aircraft. Field testing showed that effective location ranges declined with the depth at which the fish was located and that signals were extremely poor when transmitters were in depths more than 30 meters. Boat surveys only covered the area of Tobin Harbor. Aircraft surveys covered Tobin Harbor, Rock Harbor and Duncan Bay (Figure 1). On one occasion (January 29, 1997) locations were made from snowmobile in Tobin Harbor.
Locations of individual fish were recorded as GPS coordinates (Eagle Accu-Nav Sport hand held units) for all tracking. GPS coordinates were mapped and movements were measured by GPS computations to describe movements and ranges. Accuracy of the GPS units used is rated as within 100 m. Field personnel believe that boat or snowmobile locations were commonly accurate to within 50m and consistently to 150 m. In some cases where individual fish used cover such as very large rocks or docks, more precise locations could be made by using the shielding effect of the cover. Locations made by aircraft were the least accurate because of the speed of the aircraft (approximately 90 knots). However, the aircraft provided an excellent means of providing general locations of individuals that had strayed furthest. Aircraft locations are not employed as data in this report, except that in the final survey (May 29, 1997) it was noted that all 17 functioning transmitters were located within the confines of Tobin Harbor.
In addition to the weekly surveys, one survey was conducted on August 10, 1996 to describe diel movement. A group of 6 tagged fish in one area of Tobin Harbor was selected based on being a manageable set from a logistic viewpoint (all were within about 2 km of the boat dock.). These fish were then located by boat at 4 hour intervals at approximately 0200 hr., 0600 hr., 1000 hr., 1400 hr., 1800 hr, and 2200 hr. The movements found in the 24 hour study are presented separately in Table 2 and Figure 2.
Results and Discussion
Throughout the study period, the maximum distance that any of the 17 tagged coasters moved from the point where it was released after tagging was 6.4 km, the minimum was 0.25 km. The mean maximum distance moved for all 17 tagged was 3.59 km. Only 1 tagged fish left Tobin Harbor, and that one moved only 2.5 km around Scovill Point to Snug Harbor. After a 3 week stay, it returned to Tobin Harbor. The ranges found here are very small compared to the wide ranging movement patterns reported for other salmonids in the Great Lakes and even smaller than the ranges reported for a reintroduced population of Lake Nipigon Strain coasters at Grand Portage, MN (Newmar et al. 1999a). In that study the mean maximum distance moved per individual was 6.06 km and the maximum for an individual was 31.58 km.
Both weekly and diel locations of coaster in Lake Superior show a distribution pattern strictly defined by association with shoreline. A total of 348 weekly (mid-day) and 36 diel locations, all occurred within 150 m of shore. As some fish could not always be located, this study does not rule out the possibility that on occasion, tagged fish moved deeper or further from shore than we observed. However the results do support the hypothesis that coasters often use nearshore habitats and shallow depths, and that their home ranges are small.
Tagged coaster movement was much greater through the sunset /sunrise period (Table 2) than during the highest light periods. Tracking 6 randomly selected coasters at 4 hour intervals showed that the distance moved between 1000 hr. to 1400 hr. and from!400 hr. to 1800 hr. averaged only 70.5 m. Movement during periods from 1800 hr. to 2200 hr. and 2200 hr. to 0200 hr. and 0200 hr. to 0600 hr. ranged from 371 m to 517 m. and averaged 436 m. Newman et al. (1999a) also noted this pattern of increased movement during the low light periods. The range of movement exhibited by individual coasters during the 24 hour period varied widely. The highest total movement of an individual was 6 km (6000 m) while the lowest range was less than 100 m. Considering the accuracy limits of the locations, this individual may not have moved from a specific site for the whole 24 hour period (it did move in subsequent weeks).
Coasters in this study were not aged, but assuming that their age, length and size at sexual maturity was similar to that found for a Lake Superior population by Newman et al. (1999), then all of the tagged fish should have reached spawning condition in the fall of 1996. It has been generally assumed that most, if not all coasters in Lake Superior spawn in streams. The Tobin Harbor population was thought to spawn either in Tobin Creek (a stream which does support some brook trout in the headwater areas) or possibly in the outlet stream of Moose Lake (Figure 1).
Weekly tracking in this study continued through October 10, 1996 which apparently included most of the spawning period for this population (Quinlan 1999). None of the tagged fish entered either of the suspected spawning streams during the month of October. One tagged fish did enter Hidden Lake and stayed through the spawning period. A final tracking effort (by aircraft) on May 29, 1997 found all 17 tagged fish in Tobin Harbor.
By observation, it appears that several tagged fish did tend to "cluster" in specific areas during October (Figure 1), where clean sand or gravel substrates existed (a common condition in Tobin Harbor) at suitable depths for spawning. Specific sites where clusters occurred included both shorelines of Tobin Harbor around Tallman Island and toward the mouth of Tobin Creek, and in, or near the outlet of Hidden Lake. Tracking of tagged fish in this project did provide useful information to guide a concurrent effort to capture mature coasters for the purpose of a gamete harvest (Quinlan 1999) to establish a broodstock from this population. A majority of the fish captured in the gamete collection came from the areas where clusters of tagged fish were found in September and October.
The results of this study showed no evidence of stream spawning and suggested that tagged coasters here were spawning at those lake shore locations where "clusters" were noted during September and October. In 1997, subsequent to completion of field work for this study, coaster spawning activity was visually confirmed (Quinlan 1999) at one of the lake shore sites where tagged fish clustered during the spawning season during this study (Figure 1).
The small ranges noted in this study, and by Newman et al 1999a, in Minnesota, could vary to some degree among other coaster stocks in Lake Superior. However, these studies suggest that historically, many local populations of coasters existed in Lake Superior and that they may have been largely discrete. If true, this would suggest that future management of coasters will need to be based on consideration of this population structure.
The movement patterns and the close and consistent association of coasters with the shoreline and shallow water areas found in both of these studies may also suggest that commercial and subsistence gillnet fisheries targeting lake trout, herring and whitefish could usually be made compatible with coaster management by avoiding the nearshore and shallow waters where native coaster populations exist or where reintroduced stocks are present.
RECOMMENDATIONS
As one of the remaining few remnant stocks of coaster in Lake Superior, and particularly because it is the only population in which lake spawning has been observed, the Tobin Harbor population can provide more important data to guide coaster rehabilitation efforts in the basin. A high priority should be given to studies that:
- Identify and describe the exact spawning locations and habitats of the stock.
- Describe and measure the population structure and dynamics of the stock, especially the age, growth and maturation cycle and the mortality rates.
ACKNOWLEDGMENTS
The author would like to thank the many agencies and individuals who made this project possible. The personnel, time and equipment donated by the Isle Royale unit of the National Park Service (especially Jack Oelfke, Larry Kangas and Dave Solem), the Michigan Department of Natural Resources, and the Bay Mills, Keweenaw Bay, Fond du Lac, Grand Portage and Red Cliff Tribal Natural Resources Departments were essential to the completion of this study.
Table 1. Tagging Data Summary
Tag
Number
Fish Length (mm)
Number of Observations
Final Observation
Maximum distance (km) from release
Mean distance (km) from release/number locations*
11
335
26
1/29/97
4.40
2.58/25
32
340
25
1/29/27
4.85
2.27/24
52
358
24
1/29/97
4.13
3.93/22
72
411
24
1/29/27
3.75
3.00/23
91
447
21
8/1/96
4.45
1.48/20
111
439
8
7/11/96
2.30
1.81/7
131
368
7
1/29/97
6.00
2.29/6
151
391
32
1/29/27
1.82
1.35/31
171
437
22
10/2/96
1.45
0.61/21
191
462
27
10/2/96
4.5
2.03/26
211
401
25
10/2/96
6.40
1.88/24
232
434
3
6/7/96
4.85
2.77/2
251
406
22
1/29/97
1.96
1.76/21
291
376
11
1/29/97
5.1
4.30/10
311
381
24
1/29/97
3.15
2.51/23
332
376
24
10/9/96
1.75
1.58/22
372
338
19
1/29/97
0.25
1.76/16
*Indicates the actual number of locations used to compute the mean distance from the release point.
Table 2. Diel movement of six tagged fish in Tobin Harbor,
July 31-August 1,1996 Average movement at 4 hour intervals.
Time interval
Average distance
traveled (m)0600 hr. To 1000 hr.
419
1000 hr. To 1400 hr.
71
1400hr. Tol800hr.
70
1800hr. To 2200 hr.
517
2200 hr. To 0200 hr.
371
