Skip to Search
Skip to Navigation
Skip to Content


Brook trout (Salvelinus fontinalis) populations in headwater channel networks in Connecticut: riverscape genetics and abundance predictions under climate change

This research was funded primarily by the CT Department of Environmental Protection’s Bureau of Natural Resources through the State and Tribal Wildlife Grants Program, with additional support from the Storrs Agricultural Experiment Station through the Hatch Act, the Weantinoge Heritage Land Trust. The USGS Biological Resources Division’s Conte Anadromous Fish Research Center partnered with this project and provided access to the genetics lab and expert training.

“Cold Water Streams” are recognized as important and imperiled habitat by the Comprehensive Wildlife Conservation Strategy for Connecticut. Brook trout (Salvelinus fontinalis) depend on such coldwater streams, and populations have declined in much of the native range along the east coast. Remaining lotic populations in Southern New England are primarily located in small headwater streams, and such is the case in Connecticut. Land development and resulting habitat alteration and fragmentation have often been associated with population declines, and climate change raises concern for the persistence of this coldwater species in the region.

Despite the current conservation need and many previous scientific studies on biology and natural history of brook trout, relatively little was known about the spatial population ecology of brook trout inhabiting headwater streams. In particular, there is a lack of ecological information at an appropriate spatial scale, so as to understand patterns and processes of local population connectivity and how population persistence might be affected by continuing anthropogenic disturbances.

The purpose of this study was to investigate fine-scale population structure and abundance of brook trout across two headwater channel networks in Connecticut, and infer trout abundance under climate change. Our specific objectives were: (1) quantify local population connectivity and spatial extent of a “population” using highly polymorphic genetic markers (i.e., microsatellite loci), (2) identify how individual genetic differentiation is influenced by riverscape factors such as geographic distance, habitat volume, and stream temperature, and (3) build a linear multilevel regression model to relate trout count to habitat variables in order to make inferences on future trout abundance under climate change.

Brook trout were captured via backpack electrofishing using spatially continuous sampling in two headwater streams in western Connecticut; Kent Falls Brook and Jefferson Hill- Spruce Brook which harbors stream channel network lengths of 4.7 km and 7.7 km, respectively. Anal fin clips were collected from captured trout and eight microsatellite loci were genotyped in the laboratory for 740 individuals (80-140 mm) subsampled in a stratified random design from all 50m-reaches in which trout were captured. A Bayesian clustering method identified that brook trout in Kent Falls Brook comprised a single population, but it revealed a hierarchical structure of genetically distinguishable sub-populations in Jefferson Hill-Spruce Brook. In this study site, such a sub-population structure was present even in the absence of apparent physical barriers, and a group of genetically distinguishable sub-populations was connected with varying degrees of local movement. The degree of local movement was high enough so that genetically distinguishable subpopulations within a watershed will not necessarily show demographic independence. Therefore, the headwater population of brook trout in Connecticut, typically occurring in watersheds of < 15 km2, may be best managed and monitored as a single population.

Landscape genetics analysis revealed that geographic distance was a consistently important variable explaining individual genetic differentiation. A weak but statistically significant isolation-by-distance pattern was observed in both study sites. Trout tagging also indicated that fish dispersal was sufficiently restricted within a single field season, although some degree of fish dispersal and gene flow did occur at the watershed scale. In Jefferson Hill-Spruce Brook, the presence of seasonal barriers, stream gradient and temperature were significant variables for explaining individual genetic differentiation at the watershed scale, although the effect of stream gradient and temperature was confounded by the presence of first-order tributaries that were identified as genetically distinguishable from each other. Further, when comparing individuals collected in a stream segment not separated by seasonal barriers (i.e., Spruce Brook), stream position within a watershed was a significant explanatory variable and trout were less genetically similar in larger, warmer and less steep reaches.

Finally, an exploratory approach was taken to infer future trout abundance under climate change using multilevel regression models. The response variable, trout count per reach, was modeled as a function of three-level habitat variables: reach-scale variables (stream depth and habitat area) nested within segment-scale variables (stream temperature and gradient) that are then nested within streams. The influence of habitat variables differed by trout size class; for example, abundance of the smallest size class (young-of-the-year) was influenced by stream temperature and riffle habitat area, but the largest size class (= 191mm) was affected by stream gradient and maximum depth. Under three climate change scenarios of increasing severity of temperature increase and stream flow reduction, brook trout of all size classes were projected to decrease in abundance, with potentially altered size structure. The effect of such altered abundance and structure on population persistence remains an important question to be answered.

Publications generated from this research:

Kanno, Y., J.C. Vokoun, K.E. Holsinger, and B.H. Letcher. 2012. Estimating size-specific brook trout abundance in continuously-sampled headwater streams using Bayesian mixed models with zero-inflation and overdispersion. Ecology of Freshwater Fish 21:239-251.


Kanno, Y., J.C. Vokoun, and B.H. Letcher. 2011. Sibship reconstruction for inferring mating systems, dispersal and effective population size in headwater brook trout (Salvelinus fontinalis) populations. Conservation Genetics 12:619-628.

Kanno, Y., J.C. Vokoun, and B.H. Letcher. 2011. Fine-scale population structure and riverscape genetics of brook trout (Salvelinus fontinalis) distributed continuously along headwater channel networks. Molecular Ecology 20:3711-3729.

Wildlife and Fisheries Conservation Center
Department of Natural Resources and the Environment
University of Connecticut
1376 Storrs Road, Unit 4087