State of the Greater Fundy Ecosystem

GREATER FUNDY ECOSYSTEM RESEARCH PROJECT

UNB Faculty of Forestry and Environmental Management

State of the Greater Fundy Ecosystem

Population and Movement of Brook Trout
in a Small Forest Stream

Douglas Clay and Shirley Butland
Fundy National Park
P.O. Box 40, Alma, N.B.
E0A 1B0

Brook Trout are the most common salmonid in many small streams of the Maritime Acadian forest ecoregion of eastern North America, particularly those higher elevation tributaries emptying to the Bay of Fundy. These fish play a key role in cycling aquatic nutrients from forest leaf litter via insects and fungi back to mammalian and avian predators. The movement of Brook Trout has been affected by forest harvesting activities, including the obstruction of fish movement through road construction. The degree of stress imposed upon the population by such conditions will be determined by the requirement for movement. Movement or migration strategies may be adaptations to different environments (Heggens et al., 1991) and thus may change over time. Gowen et al. (1994) suggested movement may be more common in variable or harsh systems and less common in more constant or benign ones such as spring fed streams or large rivers.

Brook Trout populations can have two life-history strategies in many rivers, where both the smaller sized resident fish and larger anadromous fish can coexist. Although salmonid anadromy is best known, stream residency is more common, especially among trout populations in small streams. Whether the selection of this trait is genetic or environmentally determined is still debated. Smith and Saunders (1958) considered anadromy to be less common in the Bay of Fundy region compared to waters flowing to the Gulf of St. Lawrence or the Atlantic. This may be due to the small size of many of the streams and the high energy, steep gradient, nature of many of these rivers which inhibits anadromous movement.

GOALS

This study was designed to investigate the densities and movement of resident Brook Trout in a small forest stream in FNP. Movement was inferred from tagged fish and used to assess site fidelity during the snow-free period. Stream discharge and temperature were monitored to determine if relationships between environmental variables and movements could be identified.

STUDY AREA

This investigation took place on a 1,025 m stretch of McManus Brook, a first order headwater stream of the Pollett River, a major tributary of the Petitcodiac River (Figure 1). McManus Brook is a multi-spring fed stream that flows about 1.2 km through the northwest corner of FNP. After leaving the park the stream flows about another 2 km to the Pollett River. The gradient of this stream in the study area was low (15 m/km). The stream originates in an Alder wetland (tall tree swamp) and then flows through a poorly drained Spruce-Fir-Alder forest with large individual Tamarack trees. The stream was 2.26 m wide and had a dense growth of intertwined, overhanging Alders growing along both banks. Live and dead Alder stems combined in the brook to slow water flow and increase habitat complexity. This made electrofishing, and even walking, difficult and reduced the efficiency of capture.

Figure 1. Location of McManus Brook (tributary of the
Petitcodiac River) in the northwest corner of
Fundy National Park.

The study area was bisected by Highway 114. The culvert under the road did not pose a physical obstacle to fish movement. For 110 m upstream of the highway the stream flowed through the roadside ditch where riparian vegetation has been mowed to the water's edge. The lower end of the study area was bounded by a Beaver dam and an impassable road culvert. The upper end was a semi-permanent wetland that can support fish during wet periods of the year.

The climate of the area is characterized by cold winters with significant snowfall accumulations and cool, moist summers. Precipitation is relatively uniform throughout the year with a slight increase during the winter months. During early winter the stream freezes over and in some areas, during times of low snow accumulation, it can freeze to the bottom. However, once significant snowfall has accumulated, a snow bridge forms, the stream thaws, and then runs freely under the snow bridge for the remainder of winter.

METHODS

Flow was estimated from a removable V-notch weir that was placed in the stream during periods of different flow rates. At the same time, water depth in the highway culvert was measured to develop a stream discharge to culvert-depth relationship. This was then used to estimate flow through the season from culvert-depths. Hourly water temperature was recorded on a MINILOG data logger. Atmospheric weather data were available from an automated climate station located in the watershed. Stream width was calculated by measuring the bank full width at the start of each plot and at one random location within each plot.

Two fish sampling methods were used during this investigation. A counting fence was established on Aug. 28, 1995, and was in place until Nov. 12, 1995. It was re-established May 25, 1996 to Dec.16, 1996. In addition, in 1996, the stream was divided into 21 plots, approximately 50 m each, and electrofishing was conducted from early June to early December at approx. one month intervals. Five to seven single sweep electrofishing surveys were conducted in each of the 21 plots.

All fish over 8 cm were tagged (except occasional escapees) with individually numbered polyethylene streamer tags. Earlier work indicated fish of 7 cm and less tended to be carried by the force of the current on the tag. The tagged fish were measured for fork length and weight. Sex and maturity were identified externally whenever possible. Population estimates were based on the multiple census technique of Schabel as described in Ricker (1975). The 95% confidence limits were calculated following his technique, assuming the recaptures follow a Poisson distribution.

Example of a fish tag being used in the Point Wolfe and Upper Salmon River
drainage basins (actual size)

Site fidelity was estimated using Cunjak and Randell's (1993) methodology for the frequency of occurrence of marked and unmarked individuals. Due to the difficult electrofishing habitat and the resultant low capture rate, we applied their correction factor to account for the unmarked members of the original population. To test the effect of climate between sample periods in 1996, each sample plot/period was considered an experiment and only those sites with greater than or equal to five marked fish from the previous period were used to calculate site fidelity.

RESULTS

Summer flow in McManus Brook was closely related to recent precipitation and ranged from 1 to 35 litres/second (July - Dec 1996). Peak spring snowmelt discharge was over 1000 litres/second. Mean daily water temperature in 1996 ranged from -0.2 to 17.5 oC. Cumulative precipitation measures show that the driest period occurred in August 1996. Total precipitation from May 1 to Dec 31, 1996, was 1618 mm. The 1990-95 mean was 1600 mm. The 1995 precipitation total was 1440 mm. Mean stream width was 2.26 m.

During 1995, 384 Brook Trout were caught in the trap. Of these, 179 were tagged and released and 127 recapture events occurred. Many fish were recaptured multiple times. In the June 1996 electrofishing survey, two of the 1995 tagged fish were recaptured. No other 1995 tagged fish were seen in 1996, however, 36 fish with healed tag scars were caught, indicating high tag loss rather than tag mortality.

During the 1996 operation of the counting fence, 117 fish were caught and 103 tagged. Of the 24 tag recaptures, only four had been tagged at the trap.

The 1996 electrofishing surveys caught 1449 fish. Of these, 449 were tagged and 134 of these tagged fish were recaptured at least once. Considering multiple recapture-releases there were 653 tagging events and 204 recapture events. The recaptures by time-at-large provide an indications of the period of tag retention. The return rate over the first three survey periods (months) in the 1996 season averaged 14.7% dropping to 3.2% in the fourth month after tagging.

The counting fence exhibited a similar size distribution pattern as the stream with the highest catches occurring about October 20 of each year. However, the 1995 catch was nearly 10 times higher than the 1996 catch.

Population estimates were calculated for each section of forest stream for 1996. The mean numbers of fish (greater than or equal to 8 cm) per linear meter of stream ranged from 2.08 to 5.24 in most of the study area. After aggregating the plots into approximately 250 m blocks, the population ( ³ 8 cm) was estimated to be 1843 for the study area.

Many tagged fish were never recaptured. To see if these fish were moving downstream out of the study area, a 75 m reach below the beaver pond and culvert that bounded the study area was fished in November 1996. No tags were found although 46 fish ³ 8 cm were captured.

The distance that recaptured tagged fish moved in 1996 from their capture site is skewed strongly to little or no movement. Over half of the recaptured fish did not move from their capture site and nearly 80% of the fish moved 75 m or less. The total distance moved by all recaptured fish during the study period was -1,025 m, indicating an upstream movement bias in 1996.

Home range for these fish in 1996 was small, as over 50% of the recaptured fish did not move out of their 50 m capture interval and as they were released mid-interval, with a stream width of 2.26 m, their maximum home range would be approx. 56 m2.

Site fidelity, calculated by plot (50 m reach) using the cumulative data for all marked fish in a plot, ranged from 4% to 53%. Using each site/time period as a separate experiment, the site fidelity ranged from 10% to 54%. Multiple regression analysis was used to examine the correlation between site fidelity, total precipitation (flow) and mean water temperature during the sample period. Though the correlation coefficient was low (R = 0.39) the beta prime values derived from the regression analysis indicate that both flow (negative correlation) and temperature (positive correlation) were contributing to fish movement during the 1996 study period.

DISCUSSION

Although the 50 m interval size and high degree of habitat complexity make it impossible to associate individual fish with specific habitat type. Brook Trout in McManus Brook tended to congregate in areas of cover, both under Alder branches and logs as well as under overhanging banks. This is probably to increase their chance of avoiding overhead predators such as Belted Kingfisher.

The population density of about 2 fish (greater than or equal to 8 cm) per linear m (assuming a stream width of 2.26 m) is not high when compared to that found in streams with larger flows and multispecies communities (Decker and Erman, 1992). However, when viewed in mid-summer as little more than a small drainage channel, this population density could be regarded as high.

Fish can move for many reasons, among them: habitat, food availability and distribution, reproductive sites, territoriality, and predator avoidance. Movement for any of these reasons can be seriously impaired by habitat fragmentation which can be caused by blockages (either natural or anthropogenic) or habitat alteration (e.g. ditching or riparian clearing).

Movement was affected by climate (precipitation and temperature) and could therefore vary seasonally as well as annually. Low flow results from low precipitation which in summer can be related to higher temperatures. This variable movement can result in misinterpretation of the results of a single year research study which may indicate little or no movement in an aquatic system during a particular study. Although no apparent mixing may occur in a single year that appears normal in most measured respects, the mixing that does occur in other years could be necessary to the long term survival of the stock. Movement allows fish to recolonize empty areas and to maintain the reproductive mixing that alleviates the risk of genetic bottlenecking.

IMPLICATIONS FOR MANAGEMENT

Movement is essential to promote genetic diversity and could be hampered by habitat fragmentation due to ditching and erosion from road building and forest harvest operations as well as clearing of riparian zones. Movement is controlled by many factors and habitat stress, brought about by reduced flows, could be one. Thus fish may have localized movement when environmental conditions are good with larger scale movements more common as habitat stress increases.

Jones et al. (1996) found lower heterozygosity in several small lakes within FNP than was found in their inlet streams. They suggested the lower genetic variation found in the lake populations may be the result of low numbers of founders and (or) population bottlenecks possibly due to high angling or other types of mortality. Infrequent movement in small forest streams with complex highly diverse habitats could maintain higher heterozygosity compared to the more uniform lentic environments. If low discharge, leading to harsher environmental conditions is a factor encouraging movement of Brook Trout, then geographic regions with more stable annual precipitation (or runoff) patterns may tend to have populations with higher levels of heterozygosity.

These preliminary data indicate small forest streams can have significant salmonid populations, that movement can vary within one population from year to year, and that movement occurs more often in years of climate stress - less often in years when environmental factors are stable. Thus single year environmental impact studies are probably inadequate to describe population movements within such a salmonid population.

- Table of Contents

- View other case studies

- Bibliography