230 



Fishery Bulletin 104(2) 



guild included 40 species or species groups and excluded 

 walleye pollock and Pacific cod (Table 2). The two guilds 

 devised were subjective; however, owing to their dis- 

 tinctive biomass and life history trends, the members 

 of the flatfish guild were examined separately from all 

 other species. 



Biodiversity indexes 



Biodiversity measures were calculated by using Ludwig 

 and Reynolds's (1988) recommendations for species rich- 

 ness and evenness. Richness and evenness are consid- 

 ered robust measures and allow one to use biomass 

 proportions for biodiversity estimations. The richness 

 index used was as follows: 



Richness index = e" 



where H' = ^ 



'iHi) 



A piecewise linear model (Neter et al., 1996) was 

 applied to the biodiversity indexes using S-Plus (vers. 

 6.1, Insightful Corporation, Seattle, WA) to determine 

 distinct breaks or inflection points in the linear data 

 trends. The piecewise model finds a "knot," or inflec- 

 tion point, that breaks the data set into two periods by 

 using the least sum of squared residuals and the high- 

 est i?'- for the two-line model to determine the best fit 

 lines to the data. Linear regression models were then 

 applied to the data sets by using the recommended 

 inflection point as the breaking point for the two lines. 

 The years covered for each best fit line, as well as the 

 accompanying linear statistics for the individual lines 

 and the piecewise linear model for the best-fit model are 

 presented in Table 3. The indexes calculated for survey 

 year 1975 are included for reference (denoted as an X 

 on the diversity index graphs) but were not included in 

 the linear regression models because of the time gap 

 from 1975 to 1979 and the lack of confidence in the 

 standardization of the 1975 survey compared with later 

 surveys (Bakkala, 1993). 



and /?, = the biomass of individuals belonging to the /th 

 of S species in the sample; and 

 n = the total summed biomass for the entire 

 guild for a given year (Ludwig and Reynolds, 



1988). 



The evenness index is the "modified Hill's ratio" (Ala- 

 talo, 1981) which is 



Evermess = 



where A=-;^py, where Pi = N and ?!, is the biomass of the 

 (th species and A^ is the total biomass of all iS (species) 

 in the guild (Ludwig and Reynolds, 1988). 



Biomass estimates obtained from the surveys were 

 used to calculate biodiversity values for both flatfish 

 and roundfish guilds on an annual basis. Biodiversity 

 values calculated were used as an index of the relative 

 proportion of species in each guild and allowed a direct 

 comparison from year to year of the most abundant 

 species. Species richness indicated the effective number 

 of species that are influencing the index (Hill, 1973; 

 Ludwig and Reynolds, 1988). A higher index indicates a 

 larger number of predominant species in the assemblage 

 (i.e., higher diversity). The evenness index determines 

 the distribution of the biomass proportions of the more 

 abundant species among the species in the guild. As the 

 evenness index approaches zero, the biomass proportion 

 of a single species dominates the guild; as the evenness 

 index approaches one, there is less of a single dominant 

 species and the biomass proportions are shared more 

 equally among many species in the guild (Hill, 1973; 

 Ludwig and Reynolds, 1988). 



Results 



Flatfish biomass 



Biomass estimates for the flatfish guild showed an 

 increase from the late 1970s until the early 1990s, and 

 then an overall slight decline between 1999 and 2002 for 

 all domains (Fig. 2). Species that represented a large por- 

 tion that showned an increase in biomass estimates over 

 the study period included northern and southern rock sole 

 (Lepidopsetta polyxystra and L. bilineata, spp. undeter- 

 mined), arrowtooth and Kamchatka flounder (Atheresthes 

 stomias and A. evermanni, spp. undetermined), flathead 

 sole and Bering flounder (Hippoglossoides elassodon and 

 H. robustus, spp. undetermined), and Pacific halibut 

 iHippoglossus stenolepis). Flatfish species with declining 

 biomass estimates included the yellowfin sole (Limanda 

 aspera) and Greenland turbot {Reinhardtius hippoglos- 

 soides), and a decline in lesser abundant species such as 

 longhead dab {Limanda proboscidea) and Alaska plaice 

 (Pleuronectes qiiadrituberciilafus). The inner and middle 

 domains were dominated by northern rock sole, yellowfin 

 sole, Alaska plaice, and flathead sole and Bering flounder 

 (spp. undetermined), whereas main portions of the outer 

 shelf comprised arrowtooth and Kamchatka flounder 

 (spp. undetermined), flathead sole and Bering flounder 

 (spp. undetermined). Pacific halibut, and northern and 

 southern rock sole (spp. undetermined). 



Roundfish biomass 



Roundfish biomass increased from the 1970s to the 1980s 

 and then declined steadily from the early 1990s through 

 2002 for all domains combined (Fig. 2). Among roundfish 

 species there has been a decline since the 1970s and 1980s 

 in many middle domain species, such as yellow Irish lord 

 (Hemilepidotus jordani), butterfly sculpin (H. papilio), 



