368 
Fishery Bulletin 106(4) 
Table 3 
Correlations between longline catch rates of giant grenadier ( Albatrossia pectoralis ) (grenadier), sablefish ( Anoplopoma fimbria), 
and rockfish (shortraker [Sebastes borealis ] and rougheye rockfish [Sebastes aleutianus] combined) caught during the National 
Oceanic and Atmospheric Administration Alaska Fisheries Science Center annual longline surveys in the six management 
areas, 1979-2003. The Pearson’s correlation coefficient (r) and the P-value associated with the significance of the correlation (P) 
is shown as well as the sample size (n). 
Grenadier- Sablefish Rockfish-Sablefish 
Area 
r 
P 
n 
r 
P 
n 
Bering Sea 
-0.23 
0.0006 
243 
-0.31 
<0.0001 
243 
Aleutian Islands 
-0.37 
<0.0001 
264 
0.03 
0.6626 
264 
Western Gulf 
-0.38 
<0.0001 
243 
-0.27 
<0.0001 
243 
Central Gulf 
-0.36 
<0.0001 
375 
-0.22 
<0.0001 
368 
West Yakutat 
-0.38 
<0.0001 
200 
-0.28 
<0.0001 
200 
East Yakutat and Southeast 
-0.46 
<0.0001 
269 
-0.43 
<0.0001 
400 
Table 4 
Strongest correlations between giant grenadier ( Albatrossia pectoralis) (grenadier) and sablefish (Anoplopoma fimbria) catch 
rates and between rockfish (shortraker [Sebastes borealis ] and rougheye [Sebastes aleutianus ] combined) and sablefish catch 
rates by 50-m depth increments. Data were collected during the National Oceanic and Atmospheric Administration Alaska Fish- 
eries Science Center annual longline surveys, 1979-2003. The Pearson’s correlation coefficient (r), the P-value associated with 
the significance of the correlation (P), and the sample size in) are reported, n is often larger for rockfish because in some years 
the deeper areas, where grenadier reside, were not sampled at each station. Catch rates from the Bering Sea for giant grenadier 
and sablefish could not be transformed to fit a normal distribution; therefore, no data are presented. 
Area 
Grenadier-Sablefish 
Rockfish-Sablefish 
Depth (m) 
r 
P 
n 
Depth (m) 
r 
P 
n 
Bering Sea 
350-400 
-0.35 
<0.0001 
199 
Aleutian Islands 
750-800 
-0.44 
<0.0001 
218 
400-450 
-0.09 
0.1751 
218 
Western Gulf 
450-500 
-0.51 
<0.0001 
239 
250-300 
-0.31 
<0.0001 
244 
Central Gulf 
600-650 
-0.51 
<0.0001 
356 
300-350 
-0.40 
<0.0001 
366 
West Yakutat 
700-750 
-0.50 
<0.0001 
187 
350-400 
-0.35 
<0.0001 
199 
East Yakutat and Southeast 
600-650 
-0.44 
<0.0001 
341 
400-450 
-0.53 
<0.0001 
393 
at some stations in some years, gear was not set in the 
preferred depth range for grenadier or rockfish. For 
example, in the East Yakutat and Southeast area there 
were 269 station/year combinations for grenadier and 
400 for rockfish because gear was not set deep enough for 
the preferred grenadier depth range at some stations in 
some years. At certain 50-m depth intervals, within the 
preferred depth range, correlations were more strongly 
negative than catch rate correlations for the entire pre- 
ferred depth range (Table 4). 
When sablefish catch rates were high, grenadier and 
rockfish catch rates were low, and vice versa on the 
longline surveys. To illustrate an example of these 
negative correlations, we plotted the transformed catch 
rates in the East Yakutat and Southeast area (Fig. 
2). The 90% density ellipses demonstrated that the 
trend in the data was negative in both the sablefish 
and grenadier and sablefish and rockfish scatter plots. 
Raw, untransformed catch rates also illustrated that 
the trend between the catch rates was negative. As an 
example, grenadier and rockfish catch rates were plot- 
ted against sablefish catch rates in the East Yakutat 
and Southeast area (Fig. 3). In both figures, sablefish 
catch rates were high when grenadier or rockfish were 
low and vice versa. 
When the catch rates of all the stations were av- 
eraged within a management area, the negative cor- 
relation was evident in the time series. For example, 
the time series of the average catch rates in the East 
Yakutat and Southeast area showed that grenadier and 
rockfish catch rates were higher than average when 
sablefish catch rates were lower than average and vice 
versa (Fig. 4). This trend was evident in all areas where 
there was a negative correlation. 
