Field and simulation analyses of visual methods for sampling coral cover

The consensus view that coral reefs are declining urges the need to test the short-term effects of conservation. Visual sampling methods are a low cost, technically simple, and potentially highly accurate means of monitoring reef status or calibrating larger scale aerial and satellite surveys. We studied the recent literature and found that of the three most commonly used visual methods, the belt transect (BT) method was used twice as much as the line transect (LT) method and 2.5 times as often as the linear point intercept (LPI) method. We tested the repeatability, cost-efficiency, precision, and accuracy of these three methods in the field and in computer simulations. In the field, the observer repeatability was uniformly high among the three methods (R > 0.9). Surprisingly, our study indicates that the least used method (LPI) was the most cost-effective means of measuring coral cover while being at least as precise and accurate as the other two methods. Our simulation study indicated that both the accuracy and precision of the three methods decrease sharply with decreasing coral cover, and that using between 5 and 10 transects over a homogeneous area is an appropriate sample size at either low or high coral cover.