Using hierarchical sampling to understand scales of spatial variation in early coral recruitment

Abstract

Ecological patterns are created by processes acting over multiple spatial and temporal scales. By com- bining spatially explicit sampling with variance compo- nents models, the relative importance of spatial scale to overall variability can be determined. We used a spatially structured experimental design in the Mombasa Marine National Park in Kenya to quantify variation in coral recruitment across four spatial scales (*1-1,000 m) and to generate hypotheses about processes affecting recruitment and potential sources of post-settlement mortality during early life history. For the dominant recruiting corals (Pocillopora spp.), variation in recruitment on surfaces protected from fish grazing was greatest at the largest spatial scale examined (1,000 m). We hypothesize that recruitment on protected surfaces varies mainly with larval delivery due to different lagoonal circulation and water flow between sites. Conversely, variation on surfaces exposed to fishes was greatest at the smallest spatial scale (1 m). We hypothesize that recruitment on exposed sur- faces mainly reflects local differences in the scale and intensity of fish grazing, which may obscure larval delivery patterns. Spatial variation in recruitment can affect many ecological processes and factors, including growth, sur- vival to maturity, the distribution of habitat, and variation in species interaction strengths. This study demonstrates how spatially explicit sampling, followed by variance components modeling to partition variance across scales, can help to identify potential drivers of patterns at each relevant scale.