Nichol and Somerton: Tidal stream transport of Lepidopsetta polyxystra in the eastern Bering Sea 
223 
Lauth, 2008). The two recovered fish, both captured by 
commercial trawlers, were a 34-cm-total-length (TL) (at 
release) female at liberty for 314 days (northern fish) and 
a 40-cm-TL (at release) female at liberty for 667 days 
(southern fish). Both fish were assumed to be mature 
because they were larger than the reported mean size at 
maturity of 32.8 cm (Stark and Somerton, 2002). 
The fish were tagged with Lotek wireless LTD-1100 
(St. John’s, NF, Canada) data storage tags. Tags were 
attached to the eyed-side, just below the anterior end 
of the dorsal fin with a 0.5-mm diameter stainless- 
steel wire. The wire was inserted through two points 
on the tag, through the epaxial musculature above the 
pectoral fin, and affixed on the blind side of the fish 
by using oval plastic backing. The two wire ends were 
fastened on the outside of the backing with a crimped 
connector sleeve. 
Tag data, including depth (pressure) and tempera- 
ture, were recorded at 0.5-h or 1-h time intervals, total- 
ing 12,015 and 16,346 data pairs for the northern and 
southern fish, respectively. Two sampling intervals were 
used because, as a memory management function of the 
tags, the frequency of recordings decreased with the 
time at liberty. Depth had a resolution of 0.58 m when 
fish remained at depths less than 150 m, and 1.2 m if 
the fish exceeded 150 m; temperature had an accuracy 
of ±0.3°C. The northern tag recorded for the entire 314 
days the fish was at liberty, whereas the southern tag 
recorded for 620 of 667 days at liberty before the bat- 
tery died. 
Tide prediction 
Tidal height and current speed and direction were esti- 
mated at the midpoint location between fish release and 
recovery (northern fish: 58°18'N, 167°02'W; southern 
fish: 54°55'N, 164°31'W) for each depth measurement 
using the OTIS Tidal Inversion Software (Oregon State 
University, Corvallis, OR) which was created with solu- 
tions specifically for the eastern Bering Sea (Egbert 
et al., 1994; Egbert and Erofeeva, 2002). To test the 
accuracy of the tide model, speed and direction were 
also estimated at the site of an oceanographic moor- 
ing maintained by the Pacific Marine Environmental 
Laboratory (NOAA, Seattle, WA) near each fish (Fig. 
1) and compared to the measured bottom current data. 
The northern mooring was located approximately 118 
km west of the northern fish location and collected data 
from October 2004 to April 2005; the southern mooring 
was located approximately 66 km north of the southern 
fish location and collected data from March 1995 to Sep- 
tember 1995. Each data set consisted of hourly current 
velocity vectors (u=east-west component, v=north-south 
component) over a period of 193 days. 
Identification of vertical excursions 
Time intervals during which the fish were off bottom 
were identified as follows. Measured tag depths were 
first corrected for tide height variation by subtracting 
the tide height predicted by the tide model. Because 
along-slope movements can be confused with off-bottom 
movements, the difference in bathymetric complexity 
between northern and southern tag release locations 
dictated differences in the subsequent analysis. 
For the northern fish, distance off bottom was calcu- 
lated as bottom depth minus tag depth where bottom 
depth was estimated in two stages. First, daily bottom 
depth was chosen as the maximum tag depth during 
each 24-h period, on the assumption that northern rock 
sole contact the bottom at least once a day. Second, bot- 
tom depths for each 0.5-h or 1-h recording were estimat- 
ed by linearly interpolating between the times of daily 
maxima (proc Expand; SAS, vers. 8.02, SAS Inst., Inc., 
Cary, NC). Times of vertical excursions were identified 
as those when the off-bottom distance exceeded 2 m. 
Discrete excursions away from the bottom were defined 
as groups of successive off-bottom time recordings. 
For the southern fish, which resided in steeper, more 
rugged terrain, distance off bottom was calculated 
similarly by using 6-h rather than 24-h time windows. 
In addition, during times when horizontal movements 
appeared to be occurring in steep terrain, off-bottom 
distance was calculated by using estimates of bottom 
depth for still more frequent intervals, assuming that 
bottom depth was identical to tag depth during each 
tag recording if tidal fluctuations were obvious in the 
original tag depth data (i.e., not corrected for tide). The 
assumption is that the fish was on bottom when tidal 
fluctuations were recognized. The southern fish was con- 
sidered off bottom when off-bottom distances exceeded 
3 m. Following the criteria used for the northern fish, 
each 0.5-h or 1-h depth value was designated as either 
on or off bottom, and discrete excursions were identi- 
fied. Analyses of excursions for the southern fish were 
limited to summary statistics because of the difficulty 
in accurately identifying off-bottom periods (see Discus- 
sion section). 
Diel and tidal influence on vertical excursions 
To determine whether the likelihood of excursions dif- 
fered between day and night, each 0.5-h or 1-h record 
collected by a tag was first designated as daytime or 
nighttime based on the predicted times of sunrise and 
sunset at the midpoint location between fish release and 
recovery. Daily sunrise and sunset times were calculated 
by using an algorithm obtained from the U.S. Naval 
Observatory, Astronomical Applications Department 
(Washington, DC). The percentage of off-bottom time 
recordings occurring during the day and night were then 
calculated for each fish, and the timing and duration of 
excursions were plotted with respect to day and night. 
To determine if vertical excursions of the northern 
fish were selectively made with respect to tidal cur- 
rent direction, patterns in tidal current direction were 
examined using compass plots (function Compass (x,y); 
Matlab, vers. 7.5.0.342, The MathWorks, Inc., Natick, 
MA); plots of current direction during hours when the 
fish was off bottom were compared with plots of all 
