Resource Selection: Tests and Estimation using Null Models
Resource selection (usually either habitat choice or food preference) is modelled using Manly’s alpha (Manly 1974; Manly, McDonald and Thomas, 1993; Chesson 1978, 1983). This measure uses ratios to allow for different availabilities of the resources. Resource selection models lead to measures of niche overlap between taxa, which is used in the description and analysis of biological community structure.
The taxa in this report are groups of organisms. They may be species, functional groups, size groups, or some other classification.
In this report, we construct null models (Gotelli and Graves 1996, Gotelli 2000) to evaluate and compare resource selection, using a unified approach which is appropriate for different kinds of raw data. The Bray-Curtis distance (an overlap statistic, see Schmid and Schmidt, 2006) is the measurement of distance used throughout, for comparing two statistical distributions, or comparing observed with expected values.
Definitions and notation are given in Sections 2.1 and 2.2. Sections 2.3 to 2.5 provide tests under the assumption that usage occurs as count data, and Section 3 gives adaptations of the tests for continuous data. Section 4 has a speciallydeveloped test for two different types of resources which are measured on different scales, and a brief discussion of the tests is given in Section 5.