USGS Home
Contact USGS
Search USGS

Development and application of mechanistic models of stream fish population persistence:

Our goal is to understand the mechanisms responsible for population dynamics in flowing waters and to provide managers with tools to evaluate alternative management scenarios, effects of habitat fragmentation, changes in stream flow, and likely effects of climate change on persistence of populations. This kind of analysis can best be done with detailed mechanistic models, but these models require extensive data and advanced estimation procedures. With the assistance of numerous students, post-docs and colleagues we have been transforming our very large data sets into mechanistic models of persistence. This activity has attracted substantial attention (and research funds), particularly from The Nature Conservancy, the USFWS, NOAA Fisheries, the USFS and the USGS climate change program and Eastern Region.

Our approach to developing the mechanistic models using our data involves three steps of increasing complexity: 1) Simple, multi-state models that simultaneously estimate body growth and movement in the survival estimates. These simplified models are important because they give an indication of strengths and directions of effects and because they can be used as demonstration models that managers can use to set up preliminary management tools. 2) Detailed models of each process (body growth, survival and movement) as a function of environmental variables (see Figure below). This second stage of modeling provides the parameter estimates that will form the core of the detailed management tool. 3) Bayesian parameter estimation. With the modeling in step two, the interactions between processes are not made explicit (e.g. size-dependent interaction between growth and movement). Bayesian models of survival, growth and movement can explicitly account for these interactions, resulting in more realistic models. For each of the three models of estimation complexity, we have developed (or are developing) spatially-explicit individual-based models that form the core of the management tool. Our flexible modeling framework is written in R and, once, finalized, will be made available to the public as an R package.

Primary study areas:

West Brook, MA (1996-present) The focus of this project has evolved over time, but the primary emphasis has been on seasonal collection of body size, location and life history data of individually tagged fish. Every captured fish is also fin-clipped for subsequent genotyping. Genotypes are used to examine trout spatial population structure, for specific experimental designs with stocked salmon and for estimating fitness of trout. The primary study area is a 1-km section of the West Brook, divided into 47 20-m sampling sections. In 1997, we tagged Atlantic salmon and collected data on the two trout species (brook and brown trout) in the West Brook mainstem. In 2000, we added trout tagging and in 2002 we started tagging fish in the three tributaries to the West Brook (300-m reaches). To test community responses to the addition of salmon to the system (they are stocked as 25-mm fry in the spring), we suspended Atlantic salmon stocking in 2004. Assuming continued funding, we plan to reintroduce salmon in 2012 for a complete addition-subtraction-addition test design. Movements within the stream network are monitored continuously with stationary PIT tag antennas at the confluence of each tributary and at the top and bottom of the mainstem study area. We have also conducted portable PIT tag antenna surveys to estimate small scale (space and time) movements and survival.

Shorey Brook, ME (1999-2006) We conducted this study to obtain data from a population of Atlantic salmon listed as endangered by the federal government and to provide data specific to Maine. The project was run by Gregg Horton, a Ph.D. student at the University of Massachusetts.

Catamaran Brook, New Brunswick, Canada (1999-present) The experimental design and objective of this study are similar to those for the West Brook, but the study has been conducted in Catamaran Brook, a tributary of the Miramichi River in New Brunswick. Data collected in this study supplement the large existing data set on the salmon and environmental data collected over the last 15 years in Catamaran Brook. The collection and analysis of the 15 year study has been supervised by Dr. Rick Cunjak (University of New Brunswick), who is the co-supervisor of Doug Sigourney, the Ph.D. student (through the University of Massachusetts) running the project. In addition to adding to the Catamaran Brook data (data on individually-tagged fish had not been collected previously), data collected in this project will form a critical component in model development.

Stanley Brook, ME (2006-present) This project is focused on the ecology of sea-run brook trout. The goal is to identify how many fish use marine resources and how plastic marine migrations are. PIT tag antennas at the mouth of the river identify emigrants and we are using the early history of tagged fish to test whether there are consistent life history patterns (growth, movement) among migrants.

Major collaborative studies:

Atlantic salmon:

Three Rivers near Trondheim, Norway (2002-2005, collaborators: Torbjorn Forseth and Ole Ugedal, Norwegian Institute for Nature Research). This project tested habitat-specific Atlantic salmon growth and survival across three rivers. The data formed the basis for a scaled-up model of salmon production in Norway.

Tana River, Norway. (2006-present, collaborators: Morten Johansen, Audun Rikardsen, and Per-Arne Amundsen). This project examines tributary use of Atlantic salmon parr in one of the largest and most productive undammed rivers in the world. The extreme environment (71 N) provides special challenges to the salmon, and one tactic is to seek refuge in small tributaries. Habitat use is examined with both stationary and portable PIT tag antennas.

Brook trout (resident):

Nash Stream, NH. (2005-present, collaborators: John Magee, State of New Hampshire and many USFWS personnel). This large project combines culvert replacement evaluation with detailed growth, movement and survival analyses of brook trout in three streams of the Nash Stream network. We provided oversight of study design, designed, built and installed PIT antennas and sampled and tagged fish.

Brandon Brook, VT (2003-2005, collaborators: Steve Roy and Keith Nislow, USFS). This project used PIT tags to evaluate a culvert repair. Results indicated that the repair provided improved passage for brook trout.

Brook trout (searun):

Quashnet River, MA. (2005-present, collaborators: Brandon Annett, NOAA; Steve Hurley, State of Massachusetts). This project examines the extent of marine use by tagged brook trout. A PIT tag antenna at the mouth of the river records movements into and from the estuary and frequent portable antenna surveys provide data for estimating movement and survival. Our role is study design, PIT tag tagging and antenna setup, and data analysis.

Red Brook, MA. (2007-present, collaborators: Steve Hurley, State of Massachusetts; SE MA Trout Unlimited). This project is similar to the Quashnet River study and is also on Cape Cod. It is another sea-run brook trout stream, but with very different habitat compared with the Quashnet. As in the Quashnet, a PIT tag antenna at the mouth of the river records movements into and from the estuary and frequent portable antenna surveys provide data for estimating movement and survival. Our role is study design, PIT tag tagging and antenna setup, and data analysis.

Cove Brook, ME. (2007-present, collaborators: Joe Zydlewski, Maine Cooperative Fish and Wildlife Research Unit (UMaine-Orono)). This sea-run brook trout project has similar goals and approaches as the other sea-run projects. The twist here is that the brook is also part of the federally endangered Atlantic salmon distinct population segment, so combined with our community interaction models (trout and salmon) information on the brook trout will be key to effective salmon management under ESA.