Latitudinal trends of species diversity in rocky intertidal herbivore assemblages: spatial scale and the relationship between local and regional species richness

In this study we assessed the impact of the spatial scale of analysis on patterns of latitudinal diversity of rocky intertidal invertebrates and on the relationship between local and regional species richness. Fifty-three wave-exposed sites were sampled along the coast of Chile between 18degrees40'S and 42degrees35'S, covering a range of over 2600 km and 25degrees of latitude. Three spatial scales were defined to analyze latitudinal trends of diversity: (1) regional, based on species range limits across the entire region (100s of km), (2) site, corresponding to hundreds of square meters sampled at each location, (3) sampling unit scale, corresponding roughly with a square meter of rocky shore. The analysis showed that spatial patterns of species richness and species turnover varied according to the scale used. At a regional scale, species richness showed a mid-latitudinal peak (i.e. around 30 to 32degreesS), decreasing toward northern and southern latitudes. No clear latitudinal trends in diversity were detected at site and especially at sampling unit scales. Despite the fact that at regional scale species turnover was low and geographically uniform, at the site scale beta diversity showed the existence of 2 zones of higher species turnover (19 to 20degreesS and 30 to 32 S), which may be attributed to changes in ecological and oceanographical regimes. The relationship between regional and local species richness changed depending on the 'local' scale used; unsaturated patterns (linear positive relationship between local and regional diversity) were evident at site scale, whereas saturated patterns (quadratic trend between local and regional diversity) were observed when the local scale was defined in terms of the sampling unit. We suggest that different regulating processes, operating at different spatial scales, may explain the latitudinal trends in diversity at site and sampling unit scales. While regional processes may set an upper limit in species richness at site scale, species interactions (e.g. competition) may control species richness at sampling unit scales. These findings question the current idea of the disproportionate prevalence of regional factors over local ones in determining diversity of local assemblages. They also suggest that regulation of local diversity results from the interplay of regional and local processes.