192 



Fishery Bulletin 105(2) 



greatest densities in sites with higher values for rugos- 

 ity and verticality, although, curiously, not for substrate 

 size (Table 1). Five sites (ANI, HER, HES, OJA, and 

 SPI) contained Nereocystis in 2002, compared to two 



80 



A 



o 

 o 



-- 80 



P 



i 



100 



60 



sad Ija Itu |ak OJA HER TUT HES SPI ANI 



sad OJA Itu lak SPI HER TUT l|a HES ANI 



c 



sad Itu l|a OJA lak HER TUT HES SPI ANI 



Figure 2 



Cluster dendrograms based on percent dissimilari- 

 ties of variables among sites. (A) Structural habitat 

 descriptors (rugosity, substrate size, and verticality); 

 (B) Kelp communities; and (C) Fish communities. Site 

 abbreviations indicate relative structural complexity 

 (lower case = low complexity; upper case = high complex- 

 ity. Study sites are abbreviated as follows: ANI=Anisom 

 Point; HER = Herring Islands; HES = Hesketh Island; 

 JAK= Jakolof Bay; LJA=Little Jakolof Bay; LTU=Little 

 Tutka Bay; OJA=Outside Jakolof Bay; SAD = Sadie Cove; 

 SPUMacDonald Spit; and TUT=Tutka Bay. 



sites (HER and SPI) in 2003. Canopy kelp persisted 

 throughout the winter at one site (HER). The great- 

 est number of canopy kelp individuals was observed in 

 October 2002, the lowest densities (fewer than 5 Nereo- 

 cystis/120 m- transect) were observed from November 

 through April, and no canopy kelp was observed in 

 May 2003 (Fig. 4). Canopy kelp was more abundant 

 during months with higher water temperatures and 

 in sites with larger substrate and greater vertical 

 relief, but, unlike the understory kelp, was negatively 

 correlated with rugosity (Table 1). 



The presence of some fishes was associated with 

 season, year, physical habitat characteristics, and kelp. 

 Four hundred twenty-two fishes representing 15 spe- 

 cies from eight families were sighted on 34% (n = 171) 

 of transects surveyed. Three families (Hexagrammidae 

 [greenlings], Scorpaenidae [rockfishes], and Gadidae 

 [codfishes]) each composed at least 20% of the total 

 abundance and together accounted for more than 80% 

 of all fishes sighted. Infrequently sighted fishes in- 

 cluded those in the families Pholidae (gunnels, 6%), 

 Cottidae (sculpins, 3%), Pleuronectidae (flatfishes, 

 2%), and others (including ronquils, searchers, and un- 

 identified fishes, 5%). Fish presence varied over time 

 (Table 2); more fishes were sighted in 2003 (2.59 ±6.92 

 fish/transect) than in 2002 (0.63 ±0.96 fish/transect). 

 More fish (considering all fish species pooled across 

 months, years, and sites) were seen during periods 

 with higher temperatures and in sites characterized 

 by larger substrate and greater densities of annual 

 understory and canopy kelp (Table 2). Greenlings (pri- 

 marily kelp greenling [Hexagrammos decagrammus]) 

 accounted for the majority of sightings (35% of total 

 abundance) and their presence did not differ among 

 months (Table 2). Greenlings were most commonly 

 seen in sites with low rugosity values and larger sub- 

 strate (i.e., boulder to bedrock), and during periods 

 with warmer water temperatures and higher densities 

 of annual understory kelps. 



Schooling species, such as rockfishes (primarily the 

 black rockfish [S. mela?iops]) and adult Pacific cod 

 (Gadus macrocephalus), were observed infrequently. 

 However, these groups accounted for the greatest num- 

 ber of fish seen on any one transect and exhibited the 

 greatest variability in sightings per month in the fish 

 groups, primarily during summer 2003 (Fig. 5). There 

 was no difference in the presence of rockfishes among 

 months, but there were significant temporal differences 

 for codfishes (Table 2). Considering the major families 

 observed in this study, only the presence of rockfish- 

 es showed significant annual variability (Table 2) — 

 more in 2003 (1.23 ±5.83 fish/transect) than in 2002 

 (0.08 ±0.42 fish/transect). Both rockfishes and cod- 

 fishes were most commonly seen during periods with 

 higher water temperatures (Table 2). Although these 

 results are based only on sightings of adult fishes, 

 large schools (thousands of individuals) of juvenile 

 codfishes (predominately G. macrocephalus) were ob- 

 served at all sites during August and September 2002. 

 The juvenile codfish schools observed in summer 2003 



