Review: Izsak, Price (2001) Measuring β-diversity using a taxonomic similarity index, and its relation to spatial scale. Marine Ecology Progress Series, 215:69-77.

Feature Paper: DOWNLOAD * Izsak, Price (2001) Measuring β-diversity using a taxonomic similarity index, and its relation to spatial scale. Marine Ecology Progress Series, 215:69-77.

Author Abstract: We present a new similarity index, taxonomic similarity (ΔS), which can be used to measure β-diversity. ΔS utilises species presence/absence data, and incorporates both higher taxon richness and evenness concepts. It is derived from the average taxonomic distance (relatedness) of any 2 species from different sites. Therefore ΔS is analogous to taxonomic distinctness recently developed for biodiversity assessment at α- and γ- (landscape or seascape) scales. ΔS is a new index, although its derivation uses a concept similar to the ‘optimal taxonomic mapping statistic’ developed independently for quantifying structural redundancy in marine macrobenthos. Using echinoderm data, we show that ΔS exhibits smoother behaviour and is less influenced by species richness, and hence sampling effort, than the widely used Jaccard coefficient of species similarity. We also believe ΔS to be a more intuitive and comprehensive measure of similarity than Jaccard and other conventional indices based solely on species held in common. Taxonomic similarity between sites is computed for echinoderms examined over 3 different spatial scales: local/small-scale (<10 km), intermediate-scale (10 to 100s km) and province/oceanic-scale (100s to 1000s km). Taxonomic similarity between sites increases progressively with spatial scale, with significantly lower values and higher β-diversity at small spatial scales. The same pattern is evident for species similarity, using the Jaccard coefficient. Possible explanations for this pattern centre on: (1) the large-scale oceanic area examined (Indo-West Pacific), representing a metapopulation of echinoderms for the 2 other, smaller areas examined within (Pula Wé, Sumatra and Lakshadweeps); (2) greater biophysical instability and unpredictability at small spatial scales. Compared with larger spatial scales, these may be characterised by greater likelihood and influence of species migrations and extinctions on a site’s total species composition. Hence, species composition may be highly changeable at small scales, leading to high β-diversity. These findings are based on 1 set of comparative data for 1 faunal group. Any wider conclusions drawn would be premature, although corals may also show greater β-diversity at small spatial scales. The extent to which patterns observed are evident for other marine species groups is not well known.

Note to Readers: Follow links above for author email, full article text, or the publishing scientific journal. Author notes in my review are in quotes.

Review: Today we will look at a then-new similarity index, called "taxonomic similarity) and how it relates to β-diversity, or beta diversity. According to Wikipedia, β-diversity can be defined as "the rate of change in species composition across habitats or among communities. It gives a quantitative measure of diversity of communities that experience changing environments." Therefore, unlike alpha diversity (or pure species counts for an individual location), β-diversity compares the species diversity between locations and looks at change across locales, while gamma diversity looks at β-diversity across a very large regional or global scale. Like Monday's paper (first paper this week) I encourage readers to download the full article to learn the mathematical formulas for the taxonomic similarity index defined. What I'll do is just look at the basics of the paper and interested readers can follow up with the download link to learn more.

The authors use species presence / absence data (checklists) to derive "the average minimum path length between any 2 species in different sites / areas." They use "path lengths" based on taxonomic relatedness, in the following way: "taxonomic path lengths are 0 (same species), 1 (different species but same genus), 2 (different genus but same family), 3 (different family but same order), etc. However, path length here refers to species in different sites / areas, rather than in only one site / area."

A summary of how the authors look at taxonomic similarity can be seen in their figure, below:

In their paper, the authors also showed the effect of increasing sampling effort and sampling area on taxonomic similarity measurements, but I'll leave it to interested readers to follow up, since the authors go into a very extensive analysis and compare their results and index to the Jaccard similarity index.

The authors further point out that while alpha diversity is quite well known in the marine environment (presence / absence checklists are one of the staples of marine biodiversity studies), beta diversity is less well known and that they hope their similarity index (which "appears to be less influenced by sampling effort than [the] Jaccard [index]") helps contribute to scientific progress. Furthermore, their index also seems to be less influenced by difference in area between localities, though common sense would dictate that one should try to compare relatively similarly sized areas… but sometimes only country-level data are available and at least there is some way to compare such areas.

The authors conclude by mentioning that all taxonomic levels ("not just species") should be considered in biodiversity studies, though many authors have shown that at least for the majority of tropical marine organisms, biodiversity is relatively similar across taxonomic levels (though not for all organisms so one must be careful about drawing comparisons between their group of interest and other groups).