Pairwise differences in density estimates

Posted 24 October 2020 by Eric Rexstad

Often ecological questions extend beyond simply wanting an estimate of density in a study region. It is common for inference to extend to differences in density over time or space. A function to assess the significance of differences in density estimates exists in our examples library.

Buckland et al. (2001, Section 3.6.5) describes methods for computing t-tests assessing the veracity of the hypothesis that two density estimates do not differ. They examine two situations: where estimates share a common detection function and where estimates have distinct detection functions.

In a new vignette, we present a function that makes this inference non-parametrically, based upon the use of bootstrap replicates, borrowing upon the capability present in the bootdht() function. The function operates on the premise that the user is interested in making pairwise comparisons between stratum-specific density estimates. Those estimates are found in a dsmodel object passed to the function. Because strata are a general construct, the potential exist to assess differences in space (geographic strata) or time (strata represent, say, annual surveys of a given area) or between species (a multispecies survey where strata in the data organisation represent species).

Visit this vignette to see the function and examples of its use.