572 



Fishery Bulletin 100(3) 



Sebasto/obus afascanus 



Mafacocotfus sp 



Micrustomus pacfficus 



Sebastes ateutianus 



Sebastes babcocki 



Sebastes boreafis 



Sebastes a/ufus 



Anopfopoma fimbria 



Myctophidae 



A/batrossfa pectorafrs 



Lyopsetfa exfTrs 



Sebastes zacentrus 



Gfyptocepfia/us zachiws 



Sebastes hetvomacufatus 



BGymnocantftus sp 

 Bafhymaster sfgnatus 

 Trigtops sp. 

 Parophrys vetufus 

 Tncliodon trichodon 

 Mrcrogadus projcrmus 

 MafTofus wTTosus 

 Rajidae unident 

 Theragra chafcogramma 

 Podothecus aclpenserinus 

 Pfeuronectes quadrftubercufatus 

 Pfatfchthys stefTatus 

 Hexagrammos decagrammus 

 Lrmanda aspera 

 Isopsetta fsofepis 

 Gadus macrocephafus 

 Hippogfossus stenofeprs 

 Myoxocephafus sp 

 Hippogtossoides efassodon 

 Hemdepfdotus Jordan! 

 Lepfdopsetta sp 



K t> " >iW 



4   



t£ 



*- 



■♦«► 



¥^ 



if»# Hn .>- 



I I t^ii 



200 400 

 Depth (m) 



1000 2000 

 Alongshore distance (km) 



Figure 8 



Distribution of CPUE by depth and alongshore distance for all species 

 that had a strong positive (A) or negative (Bl association with the first 

 index of species composition derived from an NMDS ordination of abun- 

 dances averaged by strata. Widths of dark bands are proportional to 

 average CPUE of a given species. Average CPUE as a function of depth 

 and alongshore distance was estimated by using a scatterplot smoother 

 (cubic smoothing spline). Degree of smoothing was determmed by cross- 

 validation for each species separately. 



of standardizing the CPUE of each species to a common 

 gear standard prior to analysis because of large uncertain- 

 ties in the estimation of fishing power coefficients (Munro 

 and Hoff, 1995). The estimated differences among gear 

 types agreed qualitatively with results in Munro and Hoff 

 ( 1995 ) for those species for which these authors estimated 

 fishing power. Nevertheless, differences among gear types 

 may have biased some of our results. In spite of these 

 problems, the survey data used in our study provided the 

 best available indicator of relative changes in species com- 

 position and distribution for many species of ecological or 

 commercial interest. 



The observed trends in species richness, diversity, and 

 biomass, as well as in the CPUE of individual species, 

 showed that depth is an important gradient structuring 

 the groundfish community in the GOA, and that the shelf 

 break and upper slope (at 150-300 m) is a particularly 



important depth range. Strong depth-dependent gradi- 

 ents are found in many other demersal fish communities 

 inhabiting shelf and upper slope regions (Colvocoresses 

 and Musick, 1984; Gomes et al., 1992; Blaber et al., 1994; 

 Fujita et al.. 1995; Jay, 1996; Farina et al, 1997; Mahon et 

 al., 1998). Distinct depth preferences of many individual 

 species (Fig. 8) result in a turnover of species along the 

 depth gradient and lead to the observed patterns in rich- 

 ness, diversity, and biomass. Similar to the altitude gra- 

 dient in terrestrial environments (Brown and Lomolino, 

 1998), depth appears to be the major ecological gradient 

 structuring benthic communities in the ocean from shal- 

 low, nearshore areas (Mueter and Norcross, 1999) to the 

 abyssal plain (Merrett, 1992). 



Our results suggest a pronounced peak in species rich- 

 ness, diversity, and total biomass at intermediate depths. 

 Species richness as well as biomass of demersal fish com- 



