Abstract

Temperature can be predicted over space and time and can be applied

to help define trout habitat.

Issue

Trout thrive in high quality habitat as do trout fishermen and citizens concerned with the environment. Predicting the availability and location of good trout habitat helps managers determine how to best manage these populations and ascertain what effort needs to take place to improve habitat.

Response

Gardner and Sullivan (2004) examined and modeled the covariance structure of stream temperature data collected through the month of July. Because the data were found to be temporally non-stationary, and the observed changes in the covariance structure were found to be highly dependent on mainstream stream temperatures over time, a non-stationary model was developed. Five nested correlation models were created and compared using Mallows Cp model selection criteria. One model representing large-scale variation in the sill as a linear function of the mainstream temperature for each day was selected as the most parsimonious model. This modeling framework was used to define good, fair, and stressful habitat areas for trout throughout the watershed.

Impact

Agencies, such as the New York State Department of Environmental Conservation, and nongovernmental organizations, such as Trout Unlimited, have been able to identify stream sections in need of habitat restoration with the use of this model.

Funding Sources

• Federal Formula Funds - Research (e.g., Hatch, McIntire-Stennis, Animal Health)

Key Personnel

• Patrick Sullivan, Department of Natural Resources, Cornell University