A COMPARISON OF DISTANCE‐BASED AND CLASSIFICATION‐BASED ANALYSES OF HABITAT USE

• Published in: Ecology (Durham) 2003-02, Vol.84 (2), p.526-531
• Main Authors: Conner, L. Mike, Smith, Mark D., Burger, L. Wesley
• Format: Article
• Language: English
• Subjects: Analysis; Animal and plant ecology; Animal, plant and microbial ecology; Animals; Autoecology; Aves; Biological and medical sciences; Birds; compositional analysis; Euclidean distance; Fundamental and applied biological sciences. Psychology; habitat analysis; habitat selection; habitat use; Habitats; Multivariate analysis; Northern Bobwhite; radio telemetry; telemetry bias; triangulation error; Vertebrata
• ISSN: 0012-9658; 1939-9170
• DOI: 10.1890/0012-9658(2003)084[0526:ACODBA]2.0.CO;2

Quantifying habitat use is vital to understanding animal ecology. Herein, we contrast classification‐based (i.e., animal locations are placed into habitat categories for subsequent analyses) and distance‐based (i.e., distance between animal locations and habitat features are used in subsequent analyses) approaches for analyzing habitat use data. Compositional analysis (CA) and a distance‐based analysis (DA) were used to quantify habitat selection of Northern Bobwhite (Colinus virginianus). We qualitatively compared the outcome of these different procedures to illustrate advantages of the DA approach. The DA approach identified edges as an important habitat feature, and location error did not alter conclusions from DA. In contrast, CA did not detect the importance of edge, and presence of location errors altered conclusions. Moreover, modeling the distribution of location error did not effectively reduce sensitivity to error within CA. Distance‐based approaches to habitat analyses are not restricted to linear or point habitat features, require no explicit error handling, and permit extraction of more information from the data than classification‐based analyses alone. Corresponding Editor: J. M. Fryxell