222 
Fishery Bulletin 99(2) 
Map of shipboard surveys transects conducted by NOAA Ship Oregon 11 in the spring seasons 
of 1992, 1993, and 1994. Only transects conducted during active searches for cetaceans during 
adequate sighting conditions are shown. The 200-m and 2000-m isobaths are indicated in gray. 
resolution coastal bathymetric data set and Texas A&M 
University’s digitized bathymetric charts (Herring 9 ). This 
depth data set was provided on a 0.01° x 0.01° linear 
latitudedongitude grid with a nominal resolution of 1.1 
km for the entire Gulf of Mexico. Depth gradient or sea 
floor slope was derived from the depth grid by using a 
3x3 pixel Sobel gradient operator. The resulting product 
had the same base resolution and spatial coverage as the 
bathymetry data set. For descriptive purposes, the fol- 
lowing physiographic terms will be used to denote spe- 
cific depth ranges or features: continental shelf (0-200 
m), shelf break (-200 m), continental slope (200-2000 m), 
upper continental slope (200-1000 m), lower continental 
slope ( 1000-2000 m), and deep Gulf (>2000 m). 
A single descriptor of the vertical temperature structure 
in the upper ocean was selected to quantify the influence 
of mesoscale features such as eddies on cetacean distribu- 
tion. Reilly (1990) chose the depth of the 20°C isotherm 
9 Herring, H. J. 1993. A bathymetric and hydrographic cli- 
matological atlas for the Gulf of Mexico (draft report). U.S. 
Department of the Interior, Minerals Management Service, con- 
tract 14-12-0001-30631, 191 p. [Available from National Tech- 
nical Information Service, U.S. Department of Commerce, 5285 
Port Royal Road, Springfield, VA 22161.] 
as an approximate indicator of thermocline depth in his 
study of cetacean habitat in the eastern tropical Pacific. 
We used a similar approach by extracting the depth of the 
15°C isotherm from each CTD and XBT profile. This vari- 
able is not intended to represent the depth of the thermo- 
cline, however. The low-frequency, large-scale temperature 
variability along this isotherm is associated with the me- 
soscale features of interest and it occurs deep enough that 
it never reaches the sea surface during the spring in the 
northern Gulf of Mexico. 
The discrete samples of the depth of the 15°C isotherm, 
surface chlorophyll concentration and zooplankton bio- 
mass from each cruise leg (9-17 days in duration) were 
interpolated on a regular 0.1° x 0.1° linear latitude-longi- 
tude grid by using the kriging method (Golden Software, 
1994). Surface chlorophyll was log-transformed before in- 
terpolation because the observed chlorophyll concentra- 
tions had a log-normal distribution and spanned several 
orders of magnitude (0.02-13.02 mg/m 3 ). The interpola- 
tion method provided consistent results when compared 
with other data sets (e.g. Fig. 2). Because no interpolation 
method will capture the true spatial structure of these 
variables, the accuracy of the interpolated values in the 
effort and sighting datasets is undoubtedly low. Despite 
these errors, however, the horizontal variability associated 
