Local human impacts interact with geography to drive benthic community depth zonation on contemporary coral reefs

Changes in biophysical conditions and energetic resource supply across depths are predicted to promote or limit the abundance of different coral reef benthic groups. However, the degree to which regional differences in biophysical processes govern and local human activities might alter naturally occurring depth zonation patterns remains unclear. Here we used 2,239 reef surveys conducted between 0-30 m depth around 33 islands (18 unpopulated and 15 populated) across the Pacific Ocean to quantify the percentage cover change of seven broad benthic groups. We tested whether natural depth zonation patterns differed across geographies (using six ecoregions), and whether and how local human impacts might disrupt these natural zonation patterns. We found benthic community depth zonation did not always occur. At the three ecoregions where depth zonation existed, there was no universal ‘natural’ zonation pattern and the benthic groups most responsible for driving patterns of depth zonation differed across geographies. We also found evidence of human-disrupted changes to benthic community depth zonation; patterns were inversed across depths and less distinct at populated compared to unpopulated islands within two ecoregions. We show coral reef communities are naturally highly variable, and that human activities can disrupt natural patterns of ecological organisation in contemporary ecosystems.

Methods
Benthic surveys were conducted at 33 islands across the U.S.-affiliated Pacific as part of the NOAA Coral Reef Conservation Program’s National Coral Reef Monitoring Program (NCRMP). At each site, to assess benthic cover, the mean percentage cover of benthic groups was quantified by taking 30 digital images of the benthos (0.7 – 1 m 2 per image) at 1-m increments along a 30 m transect 1 m above the benthos. At each site, a single depth stratum was surveyed (either shallow, mid, or deep) to maximise spatial coverage around each island while also quantifying depth effects for each island. Coral Point Count with Excel extensions (Kohler and Gill, 2006) was then used to overlay 10 random points onto each image (300 points per transect) and the benthic group under each point assigned to one of the following categories: scleractinian (hard) coral, crustose coralline algae (CCA); turf algae, including highly cropped to thick turf mats; fleshy upright macroalgae, Halimeda spp., (a common upright calcifying macroalga in the  Pacific); fleshy and calcified encrusting macroalgae, and Alcyonacea (soft) coral. Points were pooled to generate site-level percentage cover estimates (~300 points per site).