FISHERY BULLETIN: VOL. 82. NO. 1 



possible loss of kelp (Dean 4 ) on the indigenous fish 

 fauna, also allowed us to extend the comparative 

 approach of Quast and E beling to assess two features 

 of kelp-forest fish faunas and to further evaluate a 

 sampling technique. 



The portion of the kelp forest we examined was 

 located in relatively deep water (15 m) and was 

 anchored on a low-relief cobble bottom. Since it 

 lacked a highly heterogeneous substrate, we were 

 able, by comparison, to further evaluate the effects of 

 kelp per se on nearshore fishes. Because the kelp 

 forest was in deep water, we also had the opportunity 

 to examine the vertical distribution of fishes in 

 greater detail than other workers, by sampling four 

 vertical strata, rather than the two strata (canopy and 

 bottom) sampled by Quast (1968b) and E beling etal. 

 (1980a, b). 



Besides visual transects to sample fish on or near 

 the bottom, we used underwater movies ("cinetran- 

 sects") to estimate the abundance of fishes in the 

 water column above the bottom. Alevizon and 

 Brooks (1975) and Ebeling et al. (1980b) discussed 

 the advantages and disadvantages of cinetransects, 

 but provided only rough estimates of the area 

 sampled in a cinetransect. In this paper we more 

 carefully evaluate cinetransect volume, emphasizing 

 the effect of underwater visibility on cinetransect 

 width. 



Our objectives in this paper are 1) to estimate cine- 

 transect volume as a function of underwater 

 visibility; 2) to examine the vertical distribution of 

 fishes in a deep-water kelp forest; 3) to estimate the 

 overall abundance and biomass of fishes, integrated 

 over depth, in this kelp forest; and 4) to evaluate the 

 importance of kelp to nearshore fishes, by comparing 

 our data from the San Onofre kelp forest with that 

 from an adjacent kelp-depauperate area and with 

 other published data from kelp forests located on 

 more rugose substrates. 



MATERIALS AND METHODS 



Study Areas 



This study was conducted in and near the offshore 

 portion of a giant kelp, Macrocystis pyrifera, forest 

 near the San Onofre Nuclear Generating Station, 

 between San Clemente and Oceanside, Calif. (Fig. 1). 



4 T. A. Dean. 1980. The effects of San Onofre Nuclear Generating 

 Station on the giant kelp, Macrocystis pyrifera. Annual report of the 

 Kelp Ecology Project, January-December 1979, to the Marine Re- 

 view Committee of the California Coastal Commission. Unpubl. 

 rep., 189 p. Kelp Ecology Project, Marine Science Institute, 

 University of California, Santa Barbara, CA 93106. 



San Onofre kelp (SOK) varied in areal extent from 

 <5 to 95 ha during the mid- to late 1970's, and 

 covered about 75 ha during the fall of 1979 (Dean 

 footnote 4). SOK occupied a shallowly sloping, low- 

 relief (< 1 m) cobble and sand substrate between the 

 depths of about 10 and 15 m. Two relatively perma- 

 nent, offshore portions of SOK, and an area with little 

 kelp located <100 m upcoast from SOK, served as 

 our study areas. The upcoast (SOK-U) and 

 downcoast (SOK-D) areas within SOK, and the kelp- 

 depauperate area ("kelpless" cobble), were all about 

 15 m deep and 2-3 km from shore. Because of its 

 depth, low relief, and periodic inundation by sand, 

 the cobble substrate in all areas was relatively bare of 

 understory algae and sessile invertebrates. However, 

 some stands of the 1 m tall laminarian kelp 

 Pterygophora californica were present, especially 

 along the fringes of the Macrocystis forest and 

 throughout the kelpless cobble area. 



Sampling Methods 



Our general sampling plan was to stratify fish cen- 

 suses by depth and to replicate these samples over 

 several dates. In the two kelp-forest areas, we cen- 

 sused each of three, equally spaced strata in the 

 water column, plus a bottom stratum. Only the bot- 

 tom stratum was censused at the kelp-depauperate 

 area, since few kelp-associated fishes were observed 

 above the bottom in this area. Sampling at each 

 stratum was replicated hierarchically: A number of 

 replicate transects were made within an area on a 

 given sampling day, and counts from these transects 

 were averaged. This was repeated on 4 or 5 d at each 

 site. The daily averages at each stratum and area 

 were themselves used as replicates that provided 

 reasonably precise estimates of means per stratum 

 and that allowed estimates of variability due to sam- 

 pling error. Because of time and manpower con- 

 straints, the various study areas were usually 

 sampled on different dates. All three water-column 

 strata in a given area were sampled on the same day; 

 the bottom stratum, however, was usually sampled 

 on a different day. 



All sampling took place from October through 

 December 1979. This time of year offers the most 

 consistently clear and calm water conditions. Since 

 most migratory and transient species were excluded 

 from analysis (see below), our fall study should 

 reasonably characterize the general distribution and 

 abundance of "resident", kelp-associated fishes at 

 SOK. Within this period, sampling was generally 

 limited to dates when horizontal visibility exceeded 3 

 m. 



38 



