A large-scale field comparison of strip transect and stationary point count methods for conducting length-based underwater visual surveys of reef fish populations

Abstract

Monthly visual surveys of reef fish populations in the Florida Keys were conducted from May to September of 1999 and April to December of 2000 using both strip transect and stationary point count sampling methods. Each month divers using SCUBA conducted three 10 m × 30 m transect surveys and three 5 m radius stationary point counts at 26 sites selected using a habitat based, stratified random design procedure. During both sampling procedures divers enumerated and assigned to 5 cm size increments all groupers (Serranidae), snappers (Lutjanidae), hogfish (Labridae), angelfish (Pomacanthidae), butterflyfish (Chaetodontidae), triggerfish (Balistidae), and a few other species of recreational or commercial importance. The rationale behind this sampling design was to assess the differential vulnerability of the various species and size classes to the lower behavioral disturbance/smaller area point count technique as compared to the higher disturbance/larger area transect method, and to assess which method would produce the highest density estimates and census the largest number of individuals per unit of sampling effort for size structure determinations. Although both types of samples required about the same amount of diver effort to complete, approximately 1.5 times more individuals were recorded during transect surveys. However, since transect surveys encompassed 4.3 times as much area as point counts, observed densities averaged about three times higher for point count samples. There was considerable variation among species with respect to this pattern, with total numbers of individuals for the 10 most abundant species ranging from about equal to 4.5 times higher during transect surveys, while densities ranged from nearly equal to about 6 times higher during point counts. Length frequencies within species were generally very similar between the two methods, though a higher proportion of larger individuals was recorded for a few species during transect surveys. Since previous studies have strongly indicated that visual census techniques produce underestimates of true reef fish densities, the higher and presumably more accurate density estimates provided by the point count method, along with other factors such as presumed increases in accuracy of counts and size estimates associated with increased observation time of individual fishes, led us to select this method for our future surveys.