964 



Fishery Bulletin 97(4), 1999 



Field collections 



We used beach seines to sample nearshore fish com- 

 munities. These nets effectively and nonselectively 

 sample shallow, inshore waters with sandy or smooth 

 bottoms (Cailliet et al., 1986). In Kachemak Bay, 

 samples were obtained during 16 June-26 July 1995 

 and 8 February-10 December 1996. Weather and sea 

 conditions prevented sampling in January and No- 

 vember 1996. Sampling at Chisik Island and the 

 Barren Islands occurred during 3 July-17 August, 

 and 26 June-8 September 1996, respectively. 



Our variable-mesh nets were 44 m long. The 

 middle 15.3-m-long, 4-m-deep section was made of 

 3-mm knotless nylon stretch mesh (sm), and the 

 wings tapered to a depth of 2.3 m were constructed 

 of 13-mm knotted sm nylon. Thirty meters of rope 

 were attached to the ends for deployment. Blackburn 

 et al. (1980) used similar-size seines to sample 

 Kachemak Bay in 1976. Nets were set parallel to 

 shore about 25 m from the beach as described by 

 Cailliet et al. (1986). Samples were collected about 

 every two weeks in Kachemak Bay during May-Sep- 

 tember, and once a month throughout the winter. 

 Samples from Chisik Island and the Barren Islands 

 were collected every two weeks for the duration of 

 their field seasons. 



We sampled 38 Kachemak Bay sites that were 

 seined by Blackburn et al. (1980) in 1976. At these 

 sites, we made 60 and 245 sets during May-Septem- 

 ber of 1995 and 1996, respectively. In comparison, 

 Blackburn et al. made 131 sets during May-Septem- 

 ber 1976. Raw data from the 1976 study were obtained 

 from J. Blackburn, Alaska Department of Fish and 

 Game (ADF&G), Kodiak, Alaska. 



Sets were made at eight sites on the west side of 

 Chisik Island. Twenty-three of these sets were made 

 at high tide and seven at low tide (low tide sets were 

 limited by deep mud in the subtidal zone). At the 

 Barren Islands, 40 sets (18 at high tide and 22 at 

 low tide) were made in Amatuli Cove. 



Beach seining was conducted within one-hour win- 

 dows on either side of high and low tides. To assess 

 tidal influence on catch rates, seven sites in 

 Kachemak Bay were seined regularly on consecutive 

 high and low tides during periods of maximum tidal 

 oscillation (about every 15 days). A single set usu- 

 ally provides good representation of species richness 

 and dominant species rank (Allen et al., 1992). How- 

 ever, to assess variability of beach seine catches, two 

 consecutive sets were made adjacent to each other 

 at the seven sites intensively studied at each tidal 

 stage. As a result, these sites were sampled four times 

 on each day that seining occurred. Fish were sorted 

 and counted by species. Owing to the key trophic role 



of sand lance in the Gulf of Alaska (Blackburn and 

 Anderson, 1997), these fish were also separated into 

 adult (age groups >1) and juvenile (age group 0) cat- 

 egories on the basis of otolith interpretations. 



Midwater trawls were used to sample offshore shelf 

 environments by using the ADF&G 22-m stern 

 trawler RV Pandalus in July 1996. Forage fish were 

 located with a Biosonics DT4000 digital echosounder 

 (120 kHz), and significant targets were fished by 

 using a modified herring trawl with a SO-m^ open- 

 ing. Mesh sizes diminished stepwise from about 5 cm 

 in the wings to 1 cm at the codend. The codend was 

 lined with 3-mm mesh and contained a collecting 

 bucket. Trawls were monitored with a Furuno net- 

 sounding system. Tow duration ranged from 20 to 

 60 minutes depending on fish concentrations. 

 Midwater trawls were not directly comparable to 

 beach seines; trawls selectively sampled the pelagic 

 zone, whereas beach seines unselectively sampled 

 both pelagic and demersal zones. 



Analyses 



We calculated four indices to assess broad differences 

 between sites and sampling periods. The Shannon- 

 Wiener index ( H' ) indicates diversity; it increases as 

 both the species number (richness) and equitability 

 of species abundance (evenness) increase (Pielou, 

 1977). To measure species "richness" in diversity, we 

 used Margalefs index (D; Margalef 1968), and to 

 assess equitability of species abundance, we calcu- 

 lated Pielou's evenness function (J'; Pielou, 1977). 

 Similarity between species lists from different sam- 

 pling periods was tested by using Jaccard's similar- 

 ity coefficient for presence and absence data ( Boesch, 

 1977). All diversity calculations were based on num- 

 bers of individuals and were made with natural logs 

 (log,,). Species assemblages were compared statisti- 

 cally with the Mann-Whitney rank sum test. 



Results 



Interdecadal comparison of beach seine catches in 

 Kachemak Bay: 1976 versus 1995-96 



Beach seines were an efficient method of catching 

 nearshore fish. Of the 305 sets made in Kachemak 

 Bay during May-September in 1995 and 1996, only 

 four failed to yield fish.A total of 155,991 fish compris- 

 ing 50 species were caught in these seines (Table 1). Of 

 these species, 35 were primarily represented by ju- 

 venile stages. All species found during winter were 

 also present during summer Sand lance were most 

 numerous; they represented 71% of total individuals. 



