Open Access
How to translate text using browser tools
1 December 2005 HABITAT SELECTION MODELS FOR PACIFIC SAND LANCE (AMMODYTES HEXAPTERUS) IN PRINCE WILLIAM SOUND, ALASKA
William D. Ostrand, Tracey A. Gotthardt, Shay Howlin, Martin D. Robards
Author Affiliations +
Abstract

We modeled habitat selection by Pacific sand lance (Ammodytes hexapterus) by examining their distribution in relation to water depth, distance to shore, bottom slope, bottom type, distance from sand bottom, and shoreline type. Through both logistic regression and classification tree models, we compared the characteristics of 29 known sand lance locations to 58 randomly selected sites. The best models indicated a strong selection of shallow water by sand lance, with weaker association between sand lance distribution and beach shorelines, sand bottoms, distance to shore, bottom slope, and distance to the nearest sand bottom. We applied an information-theoretic approach to the interpretation of the logistic regression analysis and determined importance values of 0.99, 0.54, 0.52, 0.44, 0.39, and 0.25 for depth, beach shorelines, sand bottom, distance to shore, gradual bottom slope, and distance to the nearest sand bottom, respectively. The classification tree model indicated that sand lance selected shallow-water habitats and remained near sand bottoms when located in habitats with depths between 40 and 60 m. All sand lance locations were at depths <60 m and 93% occurred at depths <40 m. Probable reasons for the modeled relationships between the distribution of sand lance and the independent variables are discussed.

William D. Ostrand, Tracey A. Gotthardt, Shay Howlin, and Martin D. Robards "HABITAT SELECTION MODELS FOR PACIFIC SAND LANCE (AMMODYTES HEXAPTERUS) IN PRINCE WILLIAM SOUND, ALASKA," Northwestern Naturalist 86(3), 131-143, (1 December 2005). https://doi.org/10.1898/1051-1733(2005)086[0131:SMFPSL]2.0.CO;2
Accepted: 26 August 2005; Published: 1 December 2005
KEYWORDS
Alaska
Ammodytes hexapterus
classification tree models
habitat selection
logistic regression
Prince William Sound
sand lance
Back to Top