Baumgartner et at: Cetacean habitats in the northern Gulf of Mexico 
221 
Table 1 
Environmental variables used in the habitat analyses. 
Variable 
Source 
Units 
Depth 
digital bathymetry 
m 
Depth gradient 
digital bathymetry 
m/1.1 km 
Surface temperature 
thermosalinograph 
°C 
Surface temperature standard deviation 
infrared satellite imagery 
°C 
Depth of 15°C isotherm 
CTD and XBT casts 
m 
Surface chlorophyll concentration 
surface samples 
mg/m 3 
Zooplankton biomass 
oblique bongo tows 
cc/100 m :i 
mon measure of bottom relief, 
contour index (Evans, 1975), 
was omitted because it does 
not distinguish between signifi- 
cantly different topographies in 
the northern Gulf of Mexico 
(Baumgartner, 1997). Many 
oceanographic features, such as 
eddies or river discharge, have 
strong sea surface temperature 
signatures, whereas areas where 
different water masses abut 
(frontal zones) are often charac- 
terized as regions of high surface 
temperature variability. Meso- 
scale warm-core eddies in the 
Gulf of Mexico are easily detected in hydrographic tran- 
sects by the deep occurrence of the 15°C isotherm. Finally, 
surface chlorophyll concentration and zooplankton biomass 
represent rough measures of the standing stocks on which 
higher trophic consumers might feed. 
Cetacean surveys were conducted during the spring sea- 
sons of 1992, 1993, and 1994 from NOAA Ship Oregon II 
in the Gulf of Mexico approximately north of a line con- 
necting Brownsville, Texas, and Key West, Florida, and 
primarily in waters deeper than 200 m (Fig. 1). Sighting 
data were collected with 25x binoculars and standard line- 
transect survey methods for cetaceans (e.g. Barlow, 1995; 
Hansen et al. 3 ). Time and the ship’s position were recorded 
automatically every two minutes, and at regular intervals 
the survey team recorded ancillary data, such as sea state, 
sighting conditions, and effort status. These ancillary data 
were appended to the time and position records. Environ- 
mental data were extracted from the appropriate data sets 
(discussed below) and also appended to the time and po- 
sition records. These records comprise the effort data set 
which provides a complete history of the sighting condi- 
tions, survey effort, and environmental observations. The 
cetacean sighting records were also appended with the en- 
vironmental and ancillary data and collectively represent 
the cetacean sighting data set. 
Surface temperature was recorded at one-minute in- 
tervals with a flow-through thermosalinograph (SeaBird 
Electronics, Inc, Bellevue, WA). The temperature measure- 
ments were low-pass filtered to reduce high frequency 
and high wave number variability. The filter was a simple 
5-min running mean which, at an average vessel speed of 
5 m/s ( 10 knots), is equivalent to averaging over 1.5 km. 
Conductivity, temperature, and depth (CTD) or expend- 
able bathythermograph (XBT) casts were conducted every 
55 km (30 nmi) along the survey transect. CTD casts were 
8 Green, G. A., J. J. Brueggeman, R. A. Grotefendt, C. E. 
Bowlby, M. L. Bonnell, and K. C. Balcomb III. 1992. Ceta- 
cean distribution and abundance off Oregon and Washington, 
1989-1990. In Oregon and Washington marine mammal and 
seabird surveys (J. J. Brueggeman, ed.), p. 1-100. U.S. Depart- 
ment of the Interior, Minerals Management Service, contract 
14-12-0001-30426. [Available from National Technical Infor- 
mation Service, U.S. Department of Commerce, 5285 Port Royal 
Road, Springfield, VA 22161.] 
generally made to 500 m or just off the sea floor, whichever 
was shallower. XBT probes capable of operating to depths 
of 200 to 1000 m were used in appropriate depths. Surface 
water samples were collected every 55 km and chlorophyll 
a was measured in these samples by using fluorometric and 
spectrophotometric techniques described in Strickland and 
Parsons (1972) and Jeffery and Humphrey (1975). Plank- 
ton tows were also conducted at 55-km intervals by using a 
61-cm diameter bongo equipped with 0.333-mm mesh nets 
and flowmeters. The nets were towed obliquely from 200 m 
or just off the sea floor, whichever was shallower. Samples 
from one of the bongos were analyzed by the Polish Sorting 
and Identification Center in Szczecin, Poland. Zooplankton 
biomass was computed as the ratio of the displacement vol- 
ume of the sample after organisms larger than 2.5 cm were 
removed (after Smith and Richardson, 1977) to the volume 
of water filtered during the tow. 
Remotely sensed sea surface temperature ( SST) data from 
the advanced very high resolution radiometer (AVHRR) 
carried aboard the National Oceanic and Atmospheric Ad- 
ministration (NOAA) polar orbiting environmental satel- 
lites were acquired from the U.S. National Environmental, 
Satellite and Data Information Service. The raw, level IB 
data from the NOAA 9, 10, and 11 satellites were warped to 
a 0.01° x0.01° linear latitude-longitude projection by using 
the supplied satellite navigation information, coregistered 
to a digital coastline and converted to sea surface tempera- 
tures by using separate day and night multichannel SST 
equations. Because of the lower accuracy and relative pau- 
city of the satellite-derived SST data, the in -situ surface 
temperature from the shipboard thermosalinograph was 
used in the analyses of cetacean habitat. However, these 
remotely sensed data are well suited to detecting horizon- 
tal gradients in SST due to their synoptic coverage. These 
gradients are often resolved by using digital image gradi- 
ent operators (e.g. Sobel, Prewitt, or Roberts operators), 
but we chose another approach after Smith et al. (1986). 
Because horizontal gradients in SST can be measured as 
horizontal variability, we computed the standard deviation 
of the remotely sensed SST within a 10-km radius of each 
transect and sighting position. 
Water depth was extracted from a digital bathymetric da- 
ta set compiled from NAVOCEANO's DBDB5 5-minute x 5 
minute gridded bathymetry. National Ocean Service’s high 
