240 
Fishery Bulletin 96(2), 1998 
920820 920821 920822 920823 920824 
Figure 2 
Time-depth sections of 38-kHz and 150-kHz volume scattering strength (S , dB) and temperature during five 
days of PODS92 surveys. Contour lines are isotherms (°C), from XBT (expendable bathythermograph) profiles 
at tick marks along top of section. 
of three observers using 25-power binoculars (Man- 
gels and Gerrodette, 1994a, 1994b). Local dolphin 
abundance was indexed by the number of sightings 
of spotted ( Stenella attenuata), spinner (S. longi- 
rostris), striped ( S . coeruleoalba), common (Delphi- 
nus delphis and D. capensis), bottlenose ( Tursiops 
truncatus), Risso’s (Grampus griseus), rough-toothed 
( Steno bredanensis ), Fraser’s ( Lagenodelphis hosei), 
Pacific white-sided (Lagenorhynchus obliquidens), 
and unidentified dolphins. Daily numbers of sightings 
were standardized to sighting rates of schools or in- 
dividuals per 100 km in optimum sea-state condi- 
tions (Gerrodette 1 ). Sighting rates were log-trans- 
formed to normalize distributions before testing cor- 
relation with various time and depth means of S v (z,t). 
We report correlations with 150 kHz S v here because 
data were collected on about 25% more days than 
those for 38kHz S v . 
Acoustic and other data were gridded by kriging 
with the software SURFER (Golden Software, 1995). 
Variogram models were fitted to the data and the 
best fit was selected by using the original code based 
1 Gerrodette, T. 1996. On estimating relative abundance. 
Unpubl. manuscript. 
on Pannatier (1996). Contour maps were also gener- 
ated with SURFER . 
Results 
The five-day time-depth sections in Figure 2 illus- 
trate features observed in all sections: 1) data at one 
or both frequencies were occasionally missing; 2) the 
150-kHz signal was lost in the noise at depths >200 
m; and 3) bottom depth was <400 m as on 24 August 
after about 1700 hours. All time-depth sections of S v 
were dominated by a deep-scattering layer (DSL) at 
300^100 m depth during daylight hours and 0-100 m 
during the night. The DSL migrated about 200 m in 
1-2 hours near dawn and dusk. Morning descent was 
generally more rapid than evening ascent. During 
some days, the DSL was split into two distinct lay- 
ers at depth, separated by up to 100 m (cf. Fig. 2, 38 
kHz section on 22 and 23 August). 
The DSL was detected at both 38 and 150 kHz. 
However, the 150-kHz return from the DSL at its 
daytime depth was barely above background noise 
levels. The smaller effective depth range of the ADCP 
was due to 1) greater attenuation of sound at the 
higher frequency (a=0.039 dB/m at 150 kHz com- 
