Detecting change in benthic communities at San Cristobal Reef in La Parguera, Puerto Rico, 2007-2009

Jeremy M. Kerr, California State University, Monterey Bay

Thesis (M.S.) Division of Science and Environmental Policy

Abstract

The ecological integrity of coral reef ecosystems around the world and their associated goods and services are threatened by the synergistic effects of natural and anthropogenic disturbances. Benthic community composition at San Cristobal reef in La Parguera, Puerto Rico, was monitored from 2007 to 2009 using SCUBA-based photoquadrat surveys, and assessed for change over time using a model-comparison approach, which accounted for spatial autocorrelation. Additionally, the optimal point densities and optimal photoquadrat totals necessary to detect change reliably in proportional cover were investigated. Benthic community structure was similar at all sampling sites throughout the study; algal and sand classes were the dominant classes with non-scleractinian invertebrate classes having similar or greater proportional cover than scleractinian classes. Detected changes in scleractinian classes, including Acropora and Montastraea classes, were lower than detected changes in other invertebrate classes. To investigate optimal sampling intensity, one-hundred re-samples of the point overlays were drawn for ten point densities, ranging from five points quadrat-1 to 90 points quadrat-1, and seven photoquadrat totals, ranging from two photoquadrats to 50 photoquadrats. The change for estimated proportional benthic cover between 2007 and 2008 for four benthic cover classes (M. annularis, P. astreoides, poriferans, and sand) within each re-sample were assessed for change. The minimum number of points required to detect change reliably depended on the cover class and desired detection sensitivity. Low point densities (e.g., five points photoquadrat-1 or 20 points m-2) were sufficient to detect large changes in classes with high proportional benthic cover while high point densities (e.g., 70 points photoquadrat-1 or 280 points m-2) were needed to detect small changes in benthic cover classes with low proportional cover reliably. Intermediate point densities (e.g., 50 points photoquadrat-1 or 200 points m-2) enabled detection of either small changes in classes with high proportional cover or large changes in classes with low proportional cover. Lower quadrat totals (e.g., 20 photoquadrats 100 m-2) were sufficient to detect large changes, and higher quadrat totals (e.g., 40 photoquadrats 100 m-2) were necessary to detect small changes reliably. Therefore, researchers should be conservative when determining sampling intensity and select higher point densities (e.g., 50-70 points photoqradrat-1) and higher photoquadrat totals because a less intense sampling regime (e.g., five points photoquadrat-1 or 20 photoquadrats) could result in inaccurate proportional cover estimates and unreliable detection of change. Programs monitoring benthic community structure in coral reefs using points overlaid on photoquadrats should determine the appropriate point density and number of photoquadrats per unit area based on proportional cover of the benthic classes, minimum change in cover to be detected, and the spatial autocorrelation distance, and should account for spatial autocorrelation between points when performing statistical analyses.