Patchiness of macrobenthic invertebrates in homogenized intertidal habitats: hidden spatial structure at a landscape scale

Many terrestrial habitats, and certainly man-made systems such as woodland and agricultural habitats, are characterised by a mosaic of different habitat types. In contrast, most seafloors have a rather uniform visual appearance which is enhanced by the cryptic nature of many of their inhabitants. The present study aimed to: (1) describe landscape scale spatial patterns of benthic infauna after evaluating 3 methods for analyzing auto correlations (Moran's I, semivariance and fractals), (2) compare the benthic patterns with patterns described for other organisms, and (3) highlight shared characteristics. During 4 consecutive years (2002 to 2005) we assessed spatial structuring of 4 intertidal benthic invertebrates (Cerastoderma edule, Macoma balthica, Nereis diversicolor and Nepthys hombergii) in the Wadden Sea, The Netherlands. We annually sampled similar to 2750 stations based on a 250 m grid, covering an area of ca. 225 km(2). On the basis of simulated spatial distributions, we selected Moran's I to analyze spatial patterns for the following reasons: (1) due to standardization, results can be directly compared, (2) Moran's lis the least difficult to evaluate, since it is related to the familiar Pearson's correlation coefficient, and (3) significance can readily be assessed. The 4 benthic species were all spatially structured at the landscape scale, with spatial features being smaller than the physical structure of the intertidal environment, i.e. the intertidal extent. During the 4 yr, some species changed their distribution, but spatial characteristics, i.e. patch size and amplitude of autocorrelation, remained similar. Higher overall density resulted in stronger autocorrelation with no differences between species. A comparison between spatial structuring of benthic fauna with patterns encountered in other habitats, whether marine or terrestrial, was unsuccessful due to differences in extent and grain. We argue that future research should focus on spatial structure in species' distributions as an ecological relevant parameter.