190 



Fishery Bulletin 105(2) 



seasonally with the growth and senescence 

 of annual kelps, changes in the kelp com- 

 munities were correlated with associated 

 fish abundances. 



Materials and methods 



Study sites 



This study was conducted in Kachemak 

 Bay, the southernmost inlet on the western 

 shore of the Kenai Peninsula, in south-cen- 

 tral Alaska. Ten sites were chosen based 

 on their structural characteristics and the 

 presence of kelp communities (Fig. 1). All 

 sites contained understory kelp, providing 

 varying degrees of macroalgal cover, and 

 five sites contained the canopy-forming kelp 

 Nereocystis. A distance of at least 200 m 

 (predominantly sandy bottom) separated all 

 sites from each other. Sites were situated at 

 a water depth of approximately 7 m. 



Study design 



At each site, transects (77 = 3) were surveyed 

 monthly to quantify kelp densities and fish 

 presence between May 2002 and Septem- 

 ber 2003. A haphazard starting point was 

 selected for each transect from which a 

 random direction was taken. Although 

 visibility varied among sampling periods, 

 transects were surveyed when visibility 

 was at least a transect width (2 m on each 

 side) or more and therefore such visibility 

 was not included in the analyses. Because 

 of turbidity and poor visibility at this site, 

 MacDonald Spit was not sampled in July 

 and November 2002 and Anisom Point 

 was not sampled in October 2002. Each 

 survey had two components (a kelp and a 

 fish survey), which were conducted concur- 

 rently by two separate divers. 



Kachemak Bay 



TUT I 



SPI 



JAK LJA LTL^ "  j,^(,3 e^j, 



N 



A 



OJA 



* ' ^-Little Tutka Bay 



' Uttle Jakolof Bay 

 < lakolof Bay 



Figure 1 



Location of Kachemak Bay and study sites. Sites characterized by 

 high structural complexity of the substratum are denoted with a circle 

 and low structural complexity sites are denoted with a square. 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; SPI = MacDonald Spit; and TUT=Tutka Bay. 



Physical habitat variables 



Physical habitat variables (rugosity, substrate size, 

 and verticality) were measured once for each site in 

 September 2003. Rugosity and substrate size were 

 measured for every quadrat at all sites during Sep- 

 tember 2003. Rugosity provides a measure of habitat 

 complexity on a small spatial scale and is defined as 

 the ratio of the true distance contour along the bottom 

 to a one-meter horizontal distance (Leum and Cheat, 

 1980). Rugosity was measured by using a 1-m bar 

 with a series of 5-mm links attached at one end. The 

 bar was held horizontally with the link end resting on 

 the substrate. The links were then draped along the 

 substrate beneath the bar. These links were counted 



and a rugosity measure was calculated for each quad- 

 rat and averaged per transect. Substrate size was 

 determined by measuring the diameter of samples 

 of the bottom relief (e.g., sand, cobble, bedrock) that 

 composed greater than 50% of the quadrat (Garcia- 

 Charton and Perez-Ruzafa, 2001). When no substrate 

 type dominated, the percentages and sizes of each 

 substrate type were noted. These measurements were 

 categorized from one (sand or silt) to five (bedrock) and 

 an average value was calculated for each site. Verti- 

 cality, a subjective measure ranging from one (for low 

 structural relief) to five (high), was assigned to each 

 site (Bodkin, 1986). Monthly water temperature was 

 also measured at each site. 



