thumbnail

Classification tree models for predicting distributions of michigan stream fish from landscape variables

Transactions of the American Fisheries Society

By:
, , ,
DOI: 10.1577/T07-119.1

Links

Abstract

Traditionally, fish habitat requirements have been described from local-scale environmental variables. However, recent studies have shown that studying landscape-scale processes improves our understanding of what drives species assemblages and distribution patterns across the landscape. Our goal was to learn more about constraints on the distribution of Michigan stream fish by examining landscape-scale habitat variables. We used classification trees and landscape-scale habitat variables to create and validate presence-absence models and relative abundance models for Michigan stream fishes. We developed 93 presence-absence models that on average were 72% correct in making predictions for an independent data set, and we developed 46 relative abundance models that were 76% correct in making predictions for independent data. The models were used to create statewide predictive distribution and abundance maps that have the potential to be used for a variety of conservation and scientific purposes. ?? Copyright by the American Fisheries Society 2008.

Additional Publication Details

Publication type:
Article
Publication Subtype:
Journal Article
Title:
Classification tree models for predicting distributions of michigan stream fish from landscape variables
Series title:
Transactions of the American Fisheries Society
DOI:
10.1577/T07-119.1
Volume
137
Issue:
4
Year Published:
2008
Language:
English
Larger Work Type:
Article
Larger Work Subtype:
Journal Article
Larger Work Title:
Transactions of the American Fisheries Society
First page:
976
Last page:
996