Additive partitioning of aquatic invertebrate species diversity across multiple spatial scales


Sonja E. S. Stendera, Swedish University of Agriculture, Department of Environmental Assessment, PO Box 7050, SE-750 07 Uppsala, Sweden.


1. Additive partitioning of three measures of diversity (species richness, Shannon's diversity index H and Simpson's diversity D) was used to study the relationship between local and regional diversity of benthic macroinvertebrate communities of boreal lakes (littoral habitats) and streams (riffle habitats) across three spatial scales (sampling sites, ecoregions and biogeographic regions).

2. Alpha (α) and beta (β) diversity are defined as within-habitat and between-habitat diversity, respectively. According to the concept of additive partitioning, diversity can be partitioned across multiple spatial scales such that the total (γ) diversity on one spatial scale becomes within-habitat (α) diversity at the next higher scale. Hence, the total diversity at one scale is determined by the α diversity and the between-habitat diversity (β) at the next lower scale. Consequently, one of the advantages of additive partitioning is that it is possible to study simultaneously β diversity and the regional-local species relationship and the scale dependence of α and β components.

3. For both lakes and streams α diversity was low for sites and ecoregions, whereas β diversity was high, indicating that among-site factors are important in describing the variability among the lakes and streams studied here.

4. Weak, albeit significant, evidence was found for regional and local species saturation patterns. Multiple stepwise regression indicated that local processes might be more important in structuring lake-littoral and stream-riffle species assemblages than regional processes. From these results we conclude that environmental heterogeneity may act as an important factor contributing to species coexistence, resulting in the observed saturation patterns.

5. Our study supports the use of additive partitioning for identifying specific patterns of macroinvertebrate diversity on multiple spatial scales and the underlying processes generating these patterns. This information is needed to improve understanding of the relation between patterns and processes affecting (decreasing) trends in aquatic biodiversity.