Ontogenetic and diel variation in stream habitat use by brook trout (Salvelinus fontinalis) in a headwater stream

Abstract

Although considerable information exists on habitat use by stream salmonids, only a small portion has quantitatively examined diurnal and nocturnal habitat variation. We examined diel variation in habitat use by age-0 and age-1+ brook trout (Salvelinus fontinalis) during summer and autumn in a headwater stream in northern Pennsylvania. Habitat variables measured included cover, depth, substrate, and velocity. The most pronounced diel variation occurred in the use of cover during both seasons. Both age-0 brook trout and age-1+ trout were associated with less cover at night. Age-0 brook trout occupied swifter water during the day than at night during both seasons, but the difference was not significant. Increased cover, depth, and substrate size governed the habitat of age-1+ brook trout. Our findings support the need for a better understanding of diel differences in habitat use of stream salmonids when considering habitat enhancement and protection.

Keywords:

• brook trout
• habitat
• diel
• headwaters

Introduction

Brook trout (Salvelinus fontinalis) is the dominant native stream salmonid east of the Rocky Mountains in North America (Scott and Crossman 1973). Because of declining populations, largely associated with habitat loss, several state and federal agencies, along with conservation organizations formed a partnership (i.e., Eastern Brook Trout Joint Venture) to expedite species recovery (Hudy et al. 2008). Toward the southern edge of its native range, the brook trout is increasingly restricted to headwater streams (Hudy et al. 2008). Consequently, it is important to understand brook trout habitat needs in headwater streams, which are generally unproductive (Raleigh 1982) and have a low amount of habitat complexity (Mathews 1998).

The quality and quantity of suitable habitat are important variables governing brook trout populations in streams (Marshal and Crowder 1996). Because preferred habitat can vary by fish size (Johnson and Dropkin 1996), season (Meyer and Gregory 2000), and time of day (Jakober et al. 2000), assessment of preferred habitat must account for all these variables. Failure to understand these differences can adversely impact habitat suitability models for stream salmonids (Jakober et al. 2000) and may greatly reduce the efficacy of habitat modifications designed to enhance stream populations (Bradford et al. 1995). Consequently, the most beneficial studies to help guide habitat restoration and protection efforts for brook trout should provide information on ontogenetic, seasonal, and diel variation in preferred habitat.

Although several studies have examined ontogenetic (Griffith 1972; Johnson and Dropkin 1996) and seasonal differences (Cunjak and Green 1983; Johnson 2008), little information exists on diel habitat use by brook trout (Meyer and Gregory 2000). This is surprising since differences in the diurnal versus nocturnal habitat use of other species of stream salmonids are well documented (Bonneau and Scarnecchia 1998; Roussel and Bardonnet 1999; Jakober et al. 2000). Bradford and Higgins (2001) cautioned that a lack of understanding of habitat variation in juvenile chinook salmon (Oncorhynchus tshawytscha) and steelhead trout (Oncorhynchus mykiss) associated with season, size, and time of day reduces the predictive ability of ecosystem manipulations. The objective of this study was to examine diel habitat use and habitat selection of age-0 and age- ≥ 1 (hereafter age-1+) brook trout in a representative northern Appalachian headwater stream.

Methods

Day and night habitat use by age-0 and age-1+ brook trout was examined during summer and autumn in Straight Run, a 3  km second-order stream that drains 1950 ha of the upper Susquehanna River basin in north-central Pennsylvania (41°78′N, 77°40′W). Straight Run has a fish community composed predominantly of brook trout, slimy sculpin (Cottus cognatus), blacknose dace (Rhinichthys atratulus), and longnose dace (Rhinichthys cataractae) (Johnson et al. 1992). Fish standing crop is 11.4 g/m², of which brook trout make up 44%. The representative study reach, which was established after walking the entire stream, consisted of a 1 km section (33% of the entire stream) that had a 3% gradient and an average width of 4.2 m. Stream discharges during the July and October 2006 sampling seasons were 0.15 and 0.41 m³/s, respectively. This large difference in stream discharge between summer and autumn precluded the comparison of seasonal differences in habitat use by brook trout. Riparian overstory cover (80%) consisted of mixed stands of white oak, (Quercus alba) and maple (Acer spp.). Water temperature averaged 18.5°C in July and 13°C in October.

Brook trout habitat-use assessments were made in July and October during both day (1000–1400 h) and night (2300–0300 h) periods. Habitat-use assessments (one each during the day and night during each season) made within the same month were taken 2 weeks apart to allow brook trout time to reacclimate and disperse into their preferred habitat use. We used the spot-sample electrofishing technique (Bovee 1986) to describe brook trout habitat-use. This method is effective in small shallow streams (<12 cm deep) where water depth is insufficient for snorkeling (Johnson and Kucera 1985; Heggenes et al. 1990). To effectively use this technique, sample sites must be spaced far enough apart (i.e., >3 m) to minimize fish disturbance. At the site of each fish observation, a numbered weighted buoy was placed and the age group of brook trout was recorded. Brook trout were grouped as subyearling (≤70 mm in July, ≤90 mm in October) or overyearling based on size. At the site of each buoy, we recorded water depth, water velocity, substrate composition, and the amount of cover present. Water velocity was measured at a depth of 60% from the surface with a Marsh–McBirney model 201 EM current meter (Orth 1983). Substrate size was estimated visually using a modified Wentworth particle size scale (i.e., 1, detritus; 2, mud; 3, silt; 4, sand; 5, gravel; 6, rubble; 7, boulder; 8, bedrock) (Orth et al. 1981). The amount of total available cover was estimated visually at 5% increments within a radius of approximately four fish lengths (about 24 cm for age-0 brook trout and about 52 cm for age-1+ trout) surrounding the buoy. To determine habitat selection of brook trout, we quantified available habitat within the stream reach and compared these data to brook trout habitat-use data. Available habitat was determined from 25 transects perpendicular to the bank that were established 20 m apart. Six to eight stations were located about 0.5 m apart along each transect, and habitat was quantified at each station. Available habitat was quantified once during the summer and once during autumn.

The frequency distribution of habitat variables for brook trout and available habitat were compared using the Kolmogorov–Smirnov test (Slauson 1998). We used principal component analysis (PCA) with a co-variate matrix to determine the ordination of both brook trout and available habitat variables using Conoco for Windows 4.5 (Wageningen, The Netherlands). A significance level of p ≤ 0.05 was used for all comparisons.

Results

We recorded 584 observations of brook trout habitat usage and 322 observations of available habitat (Table 1). Although higher stream discharges in the fall did not appear to affect two of the habitat variables examined (i.e., cover and substrate), depth and velocity were affected (Figure 1). In summer the average depth of the study reach was 14.7 cm compared to 23.6 cm in fall and about 81% of the available habitat in Straight Run during summer was 0–19 cm in depth, whereas only 3.8% was within this depth range in fall. Mean velocities in the study reach were 26.1 and 43.6 cm/s during summer and fall, respectively, and during summer only 19% of the available habitat had velocities ≥50 cm/s versus 41% during fall.

Figure 1. Diel distribution of age-0 brook trout (BT0) (

, day;

, night), age-1+brook trout (BT1) (

, day;

, night), and available habitat (AH) (

) in Straight Run, Pennsylvania. (A) Depth, (B) Velocity, (C) Cover, and (D) Substrate. Means and standard error are provided for each variable.

Table 1. Number of observations used to quantify brook trout habitat usage and available habitat in Straight Run, Pennsylvania.

	Summer		Autumn
Category	Day	Night	Day	Night
Brook trout 0+	66	57	56	55
Brook trout 1+	94	99	80	77
Available habitat	154	154	168	168

During both diurnal and nocturnal periods in summer, age-0 and age-1+ brook trout selected areas that had significantly more cover, depth, and substrate size than was available on average within the study reach (Figure 1, Table 2). Conversely, both ages of brook trout selected significantly slower velocities than were generally available, within the study reach during both diurnal and nocturnal periods. In the summer, age-0 and age-1+ brook trout occupied areas with significantly less cover at night than during the day. Water depth, velocity, and substrate size in summer did not differ between diurnal and nocturnal periods for age-0 trout. Brook trout occupied faster areas ( x  = 18.0 and 11.3 cm/s) during the night than during the days. The size of substrate used by either age group of brook trout did not differ between diurnal and nocturnal periods of summer.

Table 2. Results of Kolmogorov–Smirnov tests comparing differences in habitat variables used by brook trout with available habitat in Straight Run.

Comparison	Depth (cm)	Velocity (cm/s)	Cover (%)	Substrate
Summer
BT0+ – AH (D)	18.5 (1.0)–14.7 (0.6)*	17.2 (1.9)–26.1 (1.9)*	28.9 (2.1)–12.4 (0.9)*	5.0 (0.05)–4.9 (0.03)*
BT0+ – AH (N)	19.1 (1.3)–14.7 (0.6)*	13.4 (3.9)–26.1 (1.9)*	22.6 (2.0)–12.4 (0.9)*	5.0 (0.06)–4.9 (0.03)*
BT1+ – AH (D)	22.7 (0.8)–14.7 (0.6)*	11.3 (1.1)–26.1 (1.9)*	33.9 (2.0)–12.4 (0.9)*	5.2 (0.02)–4.9 (0.03)*
BT1+ – AH (N)	26.0 (1.2)–14.7 (0.6)*	18.1 (2.2)–26.1 (1.9)*	33.5 (1.5)–12.4 (0.9)*	5.1 (0.04)–4.9 (0.03)*
BT0+ – BT1+ (D)	18.5 (1.0)–22.7 (0.8)*	17.2 (1.9)–11.3 (1.1)*	28.9 (2.1)–33.9 (2.0)*	5.0 (0.05)–5.2 (0.02)*
BT0+ – BT1+ (N)	19.1 (1.0)–26.0 (1.2)*	13.4 (3.9)–18.1 (2.2)*	22.6 (2.0)–33.5 (1.5)*	5.0 (0.06)–5.1 (0.04)
BT0+ – BT0+ (D–N)	18.5 (1.0)–19.1 (1.3)	17.2 (1.9)–13.4 (3.9)*	28.9 (2.1)–22.6 (2.0)*	5.0 (0.05)–5.0 (0.06)
BT1+ – BT1+ (D–N)	22.7 (0.8)–26.0 (1.2)*	11.3 (1.1)–18.1 (2.2)*	33.9 (2.0)–29.5 (1.5)	5.2 (0.02)–5.1 (0.04)
Autumn
BT0+ – AH (D)	21.2 (1.1)–23.6 (0.8)	12.9 (2.9)–43.6 (3.0)*	30.0 (2.0)–15.8 (1.1)*	5.3 (0.05)–5.0 (0.03)*
BT0+ – AH (N)	22.2 (1.0)–23.6 (0.8)	9.5 (2.6)–43.6 (3.0)*	28.5 (2.0)–15.8 (1.1)*	5.1 (0.05)–5.0 (0.03)*
BT1+ – AH (D)	28.5 (1.5)–23.6 (0.8)*	21.8 (3.5)–43.6 (3.0)*	44.2 (2.5)–15.8 (1.1)*	5.3 (0.05)–5.0 (0.03)*
BT1+ – AH (N)	27.9 (1.1)–23.6 (0.8)*	11.0 (2.0)–43.6 (3.0)*	36.3 (1.6)–15.8 (1.1)*	5.2 (0.05)–5.0 (0.03)*
BT0+ – BT1+ (D)	21.2 (1.1)–28.5 (1.5)*	12.9 (2.9)–21.8 (3.5)*	30.0 (2.0)–44.2 (2.5)*	5.3 (0.05)–5.3 (0.05)
BT0+ – BT1+ (N)	22.2 (1.0)–27.9 (1.1)*	9.5 (2.6)–11.0 (2.0)	28.5 (2.0)–36.3 (1.6)*	5.1 (0.05)–5.2 (0.06)
BT0+ – BT0+ (D–N)	21.2 (1.1)–22.2 (1.0)	12.9 (2.9)–9.5 (2.6)	30.0 (2.0)–28.5 (2.0)	5.3 (0.05)–5.1 (0.05)*
BT1+ – BT1+ (D–N)	28.5 (1.5)–27.9 (1.1)	21.8 (3.5)–11.0 (2.0)*	44.2 (2.5)–36.3 (1.6)*	5.3 (0.05)–5.2 (0.06)
Notes: BT0+, subyearling brook trout; BT1+, overyearling brook trout; AH, available habitat; D, day; N, night.
Values presented are means with standard error within parenthesis.
*Denotes significant difference at % = 0.05.

Age-1+ brook trout occupied significantly deeper areas and areas with more cover than age-0 trout during both diurnal and nocturnal periods in summer (Table 2). Although the velocities used by both groups of brook trout in summer were significantly different at both time periods, age-0 trout occupied significantly higher velocities in the day and lower velocities at night than age-1+ trout. Age-1+ brook trout used larger sized substrate materials than age-0 trout during the day in summer.

In autumn, during both day and night periods, age-0 brook trout occupied areas that had significantly less velocity, more cover, and larger size substrate than was available (Figure 1, Table 2). Age-1+ brook trout in autumn occurred in areas with significantly more depth, slower velocities, more cover, and larger substrate size than was available during both day and night. Age-1+ brook trout used areas that were significantly deeper and faster and had more cover than age-0 trout during both time periods in autumn. The only habitat parameter that was significantly different for age-0 brook trout between day and night periods in autumn was substrate size. Conversely, age-1+ brook trout occupied significantly slower areas with less cover at night in autumn.

Axis 1 explained 72.6% and axis 2 explained 25.8% of the variation in the data using PCA (Figure 2). PCA showed that diel variation in habitat use by age-0 brook trout was more pronounced in summer, whereas diel variation in habitat use of age-1+ brook trout was greater in autumn. Increased cover, depth, and substrate size governed habitat use of age-1+ brook trout. Available habitat centroids diverged the most from brook trout habitat centroids in autumn suggesting that habitat selection of both age groups of trout was perhaps greatest during high-flow conditions.

Figure 2. First two axes of habitat data ordination using PCA. BT0+, age-0 brook trout; BT1+, age-1 brook trout; AH, available habitat; S, summer; F, fall; D, day; and N, night.

Discussion

Of the four habitat variables examined, the amount of cover used varied the most between day and night periods for both age groups of brook trout in Straight Run. The amount of cover used by brook trout during the day was always more than at night and only in one instance (age-0 in fall) were these differences not significant for age-0 and age-1+ trout. Since the size of substrate materials was a main component of instream cover, brook trout were most often associated with larger size substrate during the day. Diel variation in the amount of cover used by bull trout (Salvelinus confluentus) and cutthroat trout (Oncorhynchus clarki) has been observed (Bonneau and Scarnecchia 1998; Jakober et al. 2000). Both of these studies found more use of cover in the day. Increased use of cover during the day is thought to be associated with daytime predation risk from diurnally active terrestrial and avian predators (Jakober et al. 2000).

During both seasons, subyearling brook trout occupied swifter water during the day than at night in Straight Run. Similar diel variations in water velocities used by juvenile trout were reported for brown trout (Salmo trutta) (Roussel and Bardonnet 1999) and juvenile bull trout and cutthroat trout (Bonneau and Scarnecchia 1998). Roussel and Bardonnet (1999) found that subyearling brown trout occupied deep areas during the day, but Bonneau and Scarnecchia (1998) reported that juvenile bull trout and cutthroat trout utilized deeper areas at night. In Straight Run, there was no difference in the depth of water used by subyearling brook trout in either season. However, overyearling brook trout occupied deeper water at night in summer but not under higher stream-discharge conditions in autumn.

Less information is available on diel variation in habitat use of age-1+ salmonids in streams compared to subyearlings. Campbell and Neuner (1985) found that rainbow trout moved to shallower and slower areas at night. Banish et al. (2008) also found that bull trout occupied slower areas at night, but these areas were deeper than those used during the day. Large (≥100 mm) bull trout and cutthroat trout occupied faster and deeper areas that were farther from the substrate and cover than smaller trout during both day and night. In Straight Run, age-1+ brook trout occupied slower areas at night in autumn but were found in swifter areas at night during the summer. Nocturnal movement to deeper areas by age-1+ brook trout only occurred in summer in Straight Run.

In other studies that examined diel variation in habitat use of juvenile salmonids, observed differences were consistent among seasons (Bonneau and Scarnecchia 1998; Jakober et al. 2000). However, diel variation in habitat use of age-0 brook trout was more pronounced in the summer than in autumn in Straight Run and more pronounced in fall for age-1+ trout. It is possible that high stream discharge in autumn (2.7 × summer) influenced each age group of brook trout differently in Straight Run. If this happened, it could mask diel habitat differences of these age groups of brook trout that may occur under normal flows in autumn, perhaps illustrating the importance of stream discharge in governing habitat use by stream salmonids.

Diel variation in the use of some habitat variables occurred for both age groups of brook trout in Straight Run. This variation seems less pronounced than has been observed for stream salmonids elsewhere and may be due to the lack of habitat complexity in a small second-order stream such as Straight Run. Both age groups of brook trout selected areas with more cover during both day and night in summer and in autumn than was available. Consequently, day cover may be a critical habitat variable influencing brook trout populations in headwater streams.

Acknowledgements

The authors thank Bill Lellis, Kathy Lellis, and Dianne Franco for assistance in the field and Marc Chalupnicki for help with data analysis. This article is Contribution 1622 of the US Geological Survey Great Lakes Science Center.