Distribution of Bocaccio (sebastes Paucispinis) and Cowcod (sebastes Levis) around Oil Platforms and Natural Outcrops off California with Implications for Larval Production

There is increasing evidence that some California oil platforms form important habitats for a number of economically important fishes. We asked to what extent might platforms be important as producers of larvae of several overfished species (bocaccio, Sebastes paucispinis Ayres, 1854 and cowcod, Sebastes levis Eigenmann and Eigenmann, 1889) on a local or regional basis. We compared adult densities and potential larval production of these species at platforms and natural outcrops in California. Densities of mature bocaccio and cowcod were highly variable among survey sites, but were generally very low at both natural reefs and platforms. However, the mean densities for both species were higher around platforms than at natural reefs. Two of the three platforms (Gail and Hidalgo) that harbored mature bocaccio had larger mature individuals than did any natural reef. Platform Gail had by far the highest densities of both mature bocaccio and cowcod of any natural or human-made habitat and the potential larval production of both species at Platform Gail was much higher than at any other site surveyed. We estimated the removal of Platform Gail would be the equivalent of removing 12.57 ha of average-producing natural habitat in southern California for cowcod or 29.24 ha of average-producing natural habitat for bocaccio. These results may have implications for the platform decommissioning process. There are 27 oil and gas platforms off the coast of southern and central California. Located in both state and federal waters, these structures are situated in bottom depths ranging from 11 to 363 m and can have footprints as large as 10,606 m2 (Love et al., 2003). While all of these platforms are currently either pumping oil or gas, or are being used as transfer stations for these products, platforms have a finite economic life span. Once an industrial decision is made to cease oil and gas production, managers must decide what to do with the structure, a process known as decommissioning. Decommissioning is a complex process, involving state and federal agencies, corporate entities, and such stakeholders as recreational and commercial fishermen, and non-consumptive users (Schroeder and Love, 2004). Ultimately, a decommissioned platform could be left in place, removed to some point below the sea surface, toppled to lie on the sea floor, or totally removed (Schroeder and Love, 2004). One issue in the decommissioning process is the role that platforms may play as fish habitat. There is increasing evidence that some California platforms form important habitats for many economically important fishes. This is particularly true of the rockfishes (genus Sebastes) that often comprise over 90% of all fishes observed around platforms in southern and central California waters (Love et al., 2003). Platform habitat may serve at least two functions for these fishes. First, the midwaters of many platforms serve as nursery habitat for a suite of rockfishes and other fish species, often harboring higher densities of these species than do nearby natural outcrops (Love et al., 1999; Carr et al., 2003; Love et al., 2003). Compared to most natural reefs, a platform’s size, structural complexity, and high vertical profile probably BULLETIN OF MARINE SCIENCE, VOL. 77, NO. 3, 2005 398 provide pelagic juvenile rockfishes and larvae of other species with a relatively strong stimulus to trigger settlement. In addition, most platforms have few large fishes in the midwaters and thus predation on young fishes is likely to be low (Schroeder et al., 2000; Love et al., 2003). Platform bottoms, where the jacket and conductor pipes meet the seafloor, may harbor high densities of subadult and adult fishes, again usually comprised primarily of rockfishes (Love et al., 1999; Carr et al., 2003; Love et al., 2003). The high densities of larger fishes at the platform bottoms are due to both acceptable habitat and because some platforms are rarely fished and thus act as de facto marine reserves (Schroeder and Love, 2002; Love et al., 2003). Off southern California, fishing pressure by both recreational and commercial fishermen on natural outcrop species has been very heavy for many decades (Love et al., 1998, 2002). In particular, the adults of economically important species such as bocaccio (Sebastes paucispinis Ayres, 1854) and cowcod (Sebastes levis Eigenmann and Eigenmann, 1889) are now uncommon or even absent on many natural outcrops in southern California (Love et al., 1998b, 2003; Love and Yoklavich, unpubl. data). Given the depleted state of many rockfish species on natural reefs (seven species have been declared overfished by NOAA Fisheries), and the relative abundance of some of these species on some platforms, we asked to what extent might platforms be important as producers of larvae on a local, or even regional, basis. That is, how important are adult populations of rockfishes on platforms? To help answer that question, we conducted a pilot study that focused on cowcod and bocaccio (Fig. 1), two species declared overfished by NOAA Fisheries and the two overfished species that are most abundant as adults around California platforms. We compared adult densities and potential larval export of these species at both offshore platforms and natural outcrops in central and southern California. A knowledge of the relative importance of these human-made structures as fish habitat could play a role in decommissioning decisions. MATERIALS AND METHODS FISH SURVEYS.—Between 1995 and 2002, we surveyed platforms sited over a wide range of bottom depths, ranging between 29 and 224 m, and sited from north of Point Arguello to off Long Beach, southern California. Most of our platform surveys were conducted at seven structures (Platforms Irene, Hidalgo, Harvest, Hermosa, Holly, Grace, and Gail) located in the Santa Barbara Channel and Santa Maria Basin. In addition, we surveyed over 80 deeper-water outcrops (many in the vicinity of platforms) in waters between 30 and 360 m deep. Most of these deeper-water natural sites were visited once, a few were surveyed during as many as 4 yrs and, one outcrop, North Reef near Platform Hidalgo, was sampled annually. We surveyed fish assemblages using the DELTA submersible, a 4.6-m, two-person vessel, operated by Delta Oceanographics of Oxnard, California. Aboard the DELTA, we conducted belt transects about 2 m from the substrata, while the submersible maintained a speed of about 0.5 kts. At the platforms, transects were made around the bottom of the platform and around each set of crossbeams to a minimum depth of 20–30 m below the surface (e.g., midwater habitat). The belt transect was also used to sample the shell mounds surrounding the platforms and natural rock outcrops. The shell mounds and outcrops were sampled in consistently the same fashion as the platform method described above. Submersible surveys were conducted during daylight hours between 1 hr after sunrise and 2 hrs before sunset. During each transect, observations were taken from one viewing port on the starboard side of the submersible. An externally mounted hi-8 mm or digital video camera with associated lights filmed the same viewing fields as seen by the observers. The LOVE ET AL.: DISTRIBUTION OF BOCACCIO AND COWCOD OFF CALIFORNIA 399 observer identified, counted, and estimated the lengths of all fishes and verbally recorded those data on the video. All fishes within 2 m of the submarine were counted. All fish larger than the size of first maturity were observed within 2 m of the bottom (none were seen in midwater transects). Thus, densities were calculated as fish m−2. Fish lengths were estimated using a pair of parallel lasers mounted on either side of the external video camera. The projected reference points were 20 cm apart and were visible both to the observer and the video camera. An environmental monitoring system aboard the submarine continuously recorded date and time, depth, and altitude of the vessel above the sea floor. The environmental data were overlaid on the original videotape upon completion of each survey. Transect videos were reviewed aboard the research vessel or in the laboratory. Field observations were transcribed into a database. For each fish, we recorded the following information: 1) species (if known), 2) its estimated total length (TL), and 3) in the surveys over natural reefs, the habitat it occupied (e.g., rock, sand, mud, cobble, or boulder). Many years of experience along the Pacific Coast have shown that if the Delta is moving at a constant and slow rate of speed, as in these surveys, there is very little obvious effect on demersal rockfishes (M. Love, V. O’Connell, Alaska Department of Fish and Game, and M. Yoklavich, NOAA Fisheries, pers. obs.). Certainly, we noticed virtually no movement at all from most of the fishes in this study as the research submersible passed by. DATA ANALYSES.—While our surveys estimated size for most observed fish, we were not able to estimate sex ratios. Cowcod are not sexually dimorphic and it seemed reasonable to assume an equal female to male ratio. Bocaccio are sexually dimorphic and sex ratios of commercial catches are not 1:1 and are related to size. We obtained data to estimate length-specific sex ratios for bocaccio from Mark Wilkins (NOAA Fisheries, AFSC, Seattle) in the form of length compositions by sex from the triennial bottom trawl survey conducted by the AFSC. The trawl survey uses finer meshed nets than commercial operations and all captured fish are sampled for length compositions. Commercial operations also discard some fish due to market demand or regulations. Our submersible survey estimated fish size to the nearest 5 cm. Observers were not able to estimate size for a small proportion of fish (0.006 for bocaccio and 0.048 for cowcod). These fish were assumed to be smaller than the size-at-maturity and not included in the estimates. The trawl survey measured size to the nearest 2 cm. We converted the trawl survey length compositions to 5 cm intervals. NOAA Fisheries estimated the total abundance of bocaccio off southern and central California in numbers of fish by size and sex for each year of the survey. The estimates of total numbers of fish varied considerably among years. Since growth is sexually dimorphic for bocaccio and there is considerable variation in year-class strength, we decided it would be best to calculate the ratios of females to total fish at each year and then average the ratios, in order to give equal weight to each year’s estimates. The results indicated that ratios varied without trend through 60 cm and then rapidly increased to one (all females) at 70 cm. We concluded that because females grow faster and to larger sizes than males, it Figure 1. A) Adult bocaccio (Sebastes paucispinis) and B) cowcod (Sebastes levis) at the bottom of Platform Gail, eastern Santa Barbara Channel. Note the undercut area below the bottom-most cross beam, which provides shelter for both species. BULLETIN OF MARINE SCIENCE, VOL. 77, NO. 3, 2005 400 would be reasonable to apply the average female to total fish ratio, 0.44, between 25 and 60 cm, the observed value, 0.82, for 65 cm, and 1.00 for larger fish (Table 1). We used estimates of maturity schedules for bocaccio and cowcod obtained from the study area (M. Love, unpubl. data). We believe it was more appropriate to use these data than other published results from fish that were collected from more northern waters. These data are summarized in Love et al. (1990), but the published results were not in sufficient detail for the purpose of the present study. We converted to maturity data in 1–5 cm intervals. We estimated maturity schedules for combined sexes by sum for each sex of proportion mature weighted by proportion for that sex (Table 2). We used the size fecundity estimates of Love et al. (1990) to estimate number of eggs of mature females, as these were derived from fishes collected from the study area. The bocaccio estimates are similar to those estimated by Phillips (1964) from fish that were collected from more northern waters. For bocaccio, Fecundity = 1.15 (L3.2696), and for cowcod, Fecundity = 1.702(L3.1542), where L is TL in cm. To calculate the number of larvae that could be produced by resident bocaccio and cowcod at a given area, we first segregated all observed fish during a survey into size classes. For bocaccio, we estimated the number of females in each size class using the sex ratio data in Table 1 (cowcod are assumed to have a 1:1 ratio). Next, for each species in each size class, we used maturity schedules (Table 2) to estimate the proportion of mature females in each class. Length-fecundity equations were then used to determine the number of larvae each female could produce. Larval production is therefore linked to fish density and rates are scaled appropriately across sites. We estimated area surveyed by multiplying transect distance by two, because survey width was 2 m, and then calculated densities per area for each species and size. Transect lengths were estimated using a pair of parallel lasers mounted on either side of the external video camera. The projected reference points were 20 cm apart and were visible both to the observer and the video camera. Transect lengths were computed by counting the number of 20 cm laser segments in 15 s subsamples (1 min−1) throughout the transect, calculating speed based on those counts and averaging them over the whole transect, and multiplying that average speed by the transect duration. The 15 s subsamples were made during the first 15 s of each minute of the transect in which the laser points were visible. We first summed all data for a dive. Sometimes there was more than one dive for a station in a year. In this case, we averaged the results of the dives for estimates of the station-year. In many cases a station was surveyed in more than 1 yr and we averaged the results over years for each of those stations. We also grouped stations that were adjacent to each other. Table 1. Estimated ratio of female to total bocaccio, in 5 cm lenth intervals, in southern California.