Estimating Fish Abundance and Community Composition on Rocky Habitats in the San Juan Islands Using a Small Remotely Operated Vehicle

Estimating the abundance of marine fishes living in association with rocky habitats has been a long- standing problem because traditional net surveys are compromised by the nature of the seafloor and direct visual methods, such as scuba or submersibles, are limited or costly. In this study we used a small ROV to survey rocky habitats in the San Juan Islands (SJI) of Washington State to estimate the abundance of rockfishes (Sebastes spp), greenlings (Hexagrammidae), and other northeastern Pacific marine fishes living in nearshore, rocky habitats. The sampling frame was generated by multibeam echosounding surveys (MBES) and geological interpretation and by using charts of known rocky habitats where MBES data were not available. The survey was a stratified-random design with depths less than, or greater than, 36.6 m (120 ft) as the two depth strata. The ROV was deployed from a 12 m survey vessel fitted with an ultra-short baseline tracking system and a clump weight tethered to the ROV during most transects. Twenty-seven sampling days were expended between 29 September and 26 November 2008, during which 207 transects were conducted at depths ranging between 0 (surface) and 234 m. Transect distances were determined via the georeferenced tracking system and widths were determined by using two parallel, forward-facing lasers mounted on the ROV to measure the field of view. Densities were determined for each transect by dividing the number of individuals counted during video review by the transect area. Seafloor and biological features were evaluated throughout each transect and used to identify habitat patches that were subsequently related to mapping precision and fish occurrence. Abundance estimates were made by averaging the species densities among transects and multiplying the mean by the area of the stratum. Coefficients of variation were calculated as the percentage of the square root of the population variance divided by the population estimate.

We found that the habitat map based on the geophysical MBES interpretations always contained some rock and was highly accurate. The map based upon known or charted rocky seafloor only contained rock on 82% of transects. As expected, most rockfish were highly associated with rock, and since more rocky habitats were found in the western portion of the SJI, more rockfish were found there than in the eastern SJI. Species composition differed by depth stratum: kelp greenling, copper rockfish, Puget Sound rockfish, and lingcod were the most abundant species in the shallow stratum while quillback rockfish, Puget Sound rockfish, codfishes, spotted ratfish, lingcod, and yelloweye rockfish were the most abundant species in the deep stratum. Puget Sound rockfish was the most abundant species overall, with an estimated 4.5 million individuals. Copper and quillback rockfish abundance was 546,000 and 440,000 individuals, respectively, and survey precision was high with respective CVs of 14% and 10.5%. There was an estimated 47,000 yelloweye rockfish (25% CV), a population that is a heavily depleted in the SJI and Puget Sound. We found the ROV to be an effective survey tool for rocky habitat species living in semi-protected nearshore waters and determined that by focusing on rocky habitats, high precision was obtained for the most common species.