Spatial patterns in indicator taxa: Effects of scale of analysis, landscape unit and data aggregation.

Abstract

Successful ecosystem management depends upon accurate assessment of ecological conditions. Limited biological information for most ecosystems means that ecological indicators must serve as surrogates for other ecological characteristics in monitoring programs. The diversity of one or more taxa is a common ecological indicator whose use is based upon the assumption that species diversity in one taxa, covaries with other taxa inhabiting common landscape units. This approach also assumes that indicator species richness is predictably related to ecological characteristics of interest. Both indicator-indicator and indicator-ecosystem relationships are also assumed to remain consistent: (1) across spatial scales; (2) in different geographic regions; and (3) regardless of the size of landscape units used to sample spatial data. The influence of these factors on observed patterns of stream fish and breeding bird species richness was examined across a series of watersheds in lower Michigan using simple linear and multiple regression statistical models. Residuals from species area curves were used to examine spatial correspondence in the species richness of indicator taxa across a series of landscape units. Residuals provided a measure of indicator diversity for comparison while controlling for effects of landscape unit size on species richness. The consistency of statistical relationships between taxa richness and ecosystem characteristics was evaluated by comparing statistics from applications of a multiple regression model consisting of six landcover characteristics and area regressed against breeding bird and stream fish species richness. Comparisons of species area residuals did not reveal a clear spatial correspondence between the two indicator taxa. Correspondence in species richness across watersheds. Relationships between indicators and landcover characteristics were affected by changing the scale of analysis, landscape unit and geographic region, both quantitatively and qualitatively. A correlation in indicator-landcover relationships was seen between taxa (e.g., taxa richness responded similarly to individual landcover characteristics) but was overshadowed by large variations in taxa sensitivity to landcover characteristics. Bird richness was more sensitive then fish richness to different landcover characteristics in all comparisons except within river basins. These results suggest caution in development and use of ecological indicators for assessment of species diversity.