Miller and Brodeur: Diets of and trophic relations among dominant marine nekton within the northern California Current ecosystem 



549 



cies (e.g., Brodeur and Pearcy, 1990; Robinson, 2000). 

 During the 1980s, Brodeur et al. (1987) analyzed the 

 diets from a diverse assemblage of pelagic nekton from 

 the NCC, but since that time, the NCC system has ex- 

 hibited major shifts in the abiotic environment and in 

 community composition (Peterson and Schwing, 2003) 

 and a concomitant change in nekton species composition 

 (Emmett and Brodeur, 2000). Observed environmental 

 shifts and changes in the NCC pelagic community are 

 undoubtedly conveyed in-part through trophic inter- 

 actions. Description of the NCC in the context of the 

 current ocean environment and nekton community is 

 therefore important in determining the mechanisms 

 linking climate change and pelagic ecosystem response. 

 In this study we analyzed the diets of 26 nekton species 

 collected during two separate years (2000 and 2002) 

 from coastal northern California to central Oregon, ap- 

 plied cluster analysis to diet data to delineate trophic 

 groups based on dominant prey taxa, and compared 

 the importance of these prey between the two years for 

 several species. 



Materials and methods 



Field collections 



Nekton were collected for diet analyses from Northeast 

 Pacific Global Ocean Ecosystems Dynamics (GLOBEC) 

 cruises during 29 May-18 June and 29 July-12 August 

 2000, and 29 May-18 June and 31 July-19 August 2002. 

 Sampling occurred along nine transects across the shelf 

 between Crescent City, CA (41°54N), and Newport, OR 

 (44°39'N) (for station locations see Brodeur et al. [2004] 

 and Reese and Brodeur [2006]), during daylight hours. 

 At each station, nekton were collected with a Nordic-264 

 (Nor'Eastern Trawl Systems, Bainbridge Island, WA) 

 rope trawl (30 m widex 18 m deep) that was towed for 30 

 minutes. Up to 30 individuals per species were collected 

 per tow for diet analysis. Nekton were immediately 

 frozen on ship (-20°C) after removal from the net and 

 then processed at a later date in the laboratory. 



Laboratory analysis 



Laboratory processing of nekton involved measurement 

 of individuals and extraction of stomachs for diet analy- 

 sis. Lengths offish and squid were measured (±1.0 mm) 

 for either fork length, standard length, or dorsal mantle 

 length. Stomachs were extracted and immediately placed 

 in 10% buffered formalin for 10 days, rinsed with tap 

 water, and transferred to 70% ethanol. Diet analyses 

 were performed by assessing fullness, digestive condi- 

 tion, and identification and quantification of prey taxa in 

 each stomach. Fullness was assessed on a scale of 0-5, 

 with being empty and 5 being distended. Condition of 

 individual digested prey was assessed by using a 0-4 

 scale, with being unrecognizable and 4 being fresh. 

 Prey taxa were identified to the lowest possible taxon, 

 enumerated, and wet weighed (±0.1 mg) after excess 



water was removed with blotting paper. When individual 

 prey items were too numerous to enumerate, individual 

 weights were estimated by obtaining the damp weight of 

 a known number of animals and regressing this number 

 on the total weight of the prey. For Pacific sardine, diets 

 comprised large amounts of phytoplankton mixed with 

 small zooplankton and euphausiid eggs that required a 

 subsampling method for diet analysis (Emmett et al., 

 2005). Chinook salmon {Oncorhynchus tshawytscha) 

 were divided into subyearlings, yearlings, and adults, 

 and coho salmon (O. kisutch) were separated into year- 

 lings and adults (Brodeur et al., 2004). 



Data analysis 



Trophic relationships of nekton were examined by using 

 agglomerative hierarchical cluster analysis (AHCA) 

 to form cluster dendrograms. The percent wet weight 

 contribution of prey to predator diets were arranged 

 in a simple predator (row) x prey (column) matrix. For 

 AHCA we used the Sorensen (Bray-Curtis) distance 

 measure and Ward's linkage method. Trophic groups 

 from cluster analysis were established by choosing a 

 cutoff level with biological meaning while maintaining a 

 reasonable level (at least 40%) of information explained 

 in the cluster dendrogram. The significance of trophic 

 groups was examined using a multiresponse permutation 

 procedure (MRPP), which tests for the null hypothesis 

 of no difference between groups. MRPP gauges within 

 and between group differences using an A-statistic 

 that ranges between and 1, with being no agree- 

 ment within a group and 1 being complete agreement. 

 All AHCA and MRPP analyses were performed with 

 PC-ORD software (version 4, MjM Software, Gleneden 

 Beach, OR). 



Before nonparametric analyses of diets, certain modi- 

 fications to the data were performed. Nekton species 

 that were found in <5'7f of the tows within a cruise 

 were excluded from analyses, although these species 

 were retained in the general description of diets. This 

 level of exclusion was somewhat arbitrary; however 

 exclusion of nekton species that were found in 5-10% of 

 tows would have removed many predator species from 

 analyses. Prey taxa modifications involved removal of 

 rare species and aggregation of certain groups. Larval- 

 juvenile fishes, hyperiid amphipods, brachyuran and 

 decapod larvae, and adult fish prey were combined into 

 higher taxonomic categories and life history stages. 

 This arrangement was required to retain important 

 species groups for multivariate analyses. Adjustments 

 for all data involved the removal of rare prey that were 

 only present in 5lO% of the nekton diets (rows) in the 

 main matrix. Removal of rarer species reduced noise 

 in the data and allowed for comparisons of important 

 prey between nekton predators. 



Trophic relationships were also analyzed by calculat- 

 ing the degree of diet similarity between nekton species 

 pairs by using Schoener's similarity index (Schoener, 

 1974), modified as a percent similarity index (PSD of 

 the diets of paired nekton: 



