Species diversity in subtidal landscapes: Maintenance by physical processes and larval recruitment

Heterogeneous patterns of species diversity are rarely linked to the proces ses that maintain them on spatial scales larger than tens of meters. In thi s study, subtidal landscapes of a New Zealand fjord were used to study the spatial patterns of epifaunal invertebrate species diversity and to test hy potheses about the mechanisms maintaining the patterns. Patterns of species diversity were quantified along 1000-m(2) sections of vertical rock wall h abitat at three sites separated by 3-6 km of horizontal distance during Apr il-August 1993. Species diversity data from random 0.25-m(2) quadrats at fo ur depth strata (3, 5, 10, and 18 m) within a 20 x 50 m area were contoured to reveal spatial patterns at each site. General patterns consisted of loc alized patches of 300-660 m(2) of high diversity (i.e., H' [Shannon-Weiner diversity index] of 2.0-3.2/0.25 m(2) and S [species richness] of 28-32 spe cies/0.25 m(2)) centered at 10 m depth. High-diversity patches, dominated b y bryozoans, sponges, and ascidians, were bounded vertically by low-diversi ty mussel assemblages and horizontally by low-diversity assemblages dominat ed by encrusting calcareous algae with areas of bare rock. Vertical pattern s of diversity were attributed to physical stress from a low-salinity surfa ce layer impinging on shallow areas (e.g., 0-5 m depth) of the landscapes, while horizontal patterns could be partly attributed to large-scale landsli des and severe grazing by sea urchins. The hypothesis that the areas of high diversity could be maintained by larv al recruitment was tested by deploying an array of recruitment tiles with p redator exclusion treatments. Experiments were conducted on the same spatia l scale as the documented patterns of diversity (e.g., at 3, 5, 10, and 18 m depth), with tiles placed in and out of the high-diversity patches. Recru itment densities of calcareous polychaetes, bryozoans, and hydroids showed significant depth and habitat effects (i.e., in vs. out of the high-diversi ty areas) after 3 mo, but no significant effects of predator exclusion trea tment. A positive linear relationship between species diversity and number of recruit species in adjacent areas of wall explained 27-57% of the varian ce. These data also suggested that the size of the high-diversity patches a nd location within the fjord are important factors affecting this relations hip. We suggest that the localized patches of high diversity over hundreds of me ters of continuous habitat can be maintained on temporal scales of months e ither by spatially limited recruitment of short-lived larvae from the patch es or by recruitment near established conspecifics. Additionally, there may be a feedback mechanism involving the interaction of increased biogenic st ructure provided by the high-diversity patches enhancing recruitment. This feedback would then serve to maintain localized areas of high diversity wit hin the extensive areas of low diversity on scales of hundreds to thousands of meters.