82 
Fishery Bulletin 108(1 ) 
Table 4 
A forward stepwise logistic regression characterizing variables important in describing bottlenose dolphin ( Tursiops trun- 
catus) foraging sites in Barataria Basin, Louisiana, during from June 1999 to May 2002. Individual variables were entered 
and kept in the model with an a-level of 0.20. Variables are listed below from highest to lowest Wald x 2 statistic. Order of vari- 
able entry into the model is also indicated. Foraging and nonforaging least-square means (±1 standard error [SE] ) were calcu- 
lated for significant continuous variables, and highest and lowest proportions of foraging activity were given for season. 
Order 
Effect 
Wald x 2 
p r > x 2 
Foraging 
(Mean ±1 SE) 
Nonforaging 
(Mean ±1 SE) 
3 
Temperature 
6.78 
<0.01 
24.28 ±0.75 
22.82 ±0.54 
1 
Minimum group size 
5.54 
0.02 
9.27 ±0.72 
5.92 ±0.52 
2 
Turbidity 
4.79 
0.03 
13.22 ±0.78 
11.65 ±0.56 
4 
Season 
2.30 
0.13 
Spring (39.3%) 
Winter (73.5%) 
in factor 3. Distance from shore was greatest in open 
waters areas north of barrier islands where wetland 
areas were more sparsely distributed and also in some 
of the channels and passes opening into the Gulf of 
Mexico. These channels and passes were also typically 
the deepest parts of the study area. 
Minimum group size followed by turbidity, tempera- 
ture, and season were found to be the most important 
variables describing foraging sites according to a for- 
ward stepwise logistic regression (Table 4). The selected 
model was a reasonable fit (Hosmer and Lemeshow 
criterion, %|=5.79, P=0.67). Minimum group size, tur- 
bidity and temperature were all higher during foraging 
observations, and the incidence of foraging was highest 
in spring (39.3%) and lowest in winter (26.5%). Overall, 
foraging behavior was observed during 88 of the 269 
sightings. 
Specific ranges and levels of turbidity, water depth, 
water temperature, distance from shore, and minimum 
group size were found to have higher probabilities of 
foraging activity according to habitat suitability curves. 
Foraging suitability was calculated to be greatest be- 
tween temperatures of 20° and 24°C (Fig. 1). Foraging 
was more commonly observed when dissolved oxygen 
content was around 6 mg/L and declined as values in- 
creased (Fig. 1). Foraging was also more often observed 
in salinity values around 20 psu, turbidity values be- 
tween 20 and 28 NTU, distances between 200 and 500 
m from shore, and water depths between 4 and 6 m 
(Figs. 1 and 2). A positive relationship between in- 
creased number of foraging observations and minimum 
group size was also evident (Fig. 2). 
Discussion 
Variables related to foraging 
Our research into bottlenose dolphin foraging habitat 
in the northern Gulf of Mexico revealed that water 
temperature may be a more informative indicator of dis- 
tribution and foraging activity than season. This asser- 
tion is consistent with documented correlations between 
water temperature fluctuations (associated with La Nina 
events) and short-beaked common dolphin (Delphinus 
delphis) distribution (Neumann, 2001) and distributional 
range limits for some Delphinidae species determined by 
water temperature (Gaskin, 1968). Furthermore, Tershy 
et al. (1990) found seasonal patterns that correlated with 
the presence of fin ( Balaenoptera physalus) and Bryde’s 
(Balaenoptera edeni) whales within the Gulf of Califor- 
nia and that were negatively correlated with increasing 
water temperature. 
However, the importance of other variables (i.e., dis- 
solved oxygen, turbidity, and salinity) is noteworthy. 
These variables also showed significant seasonal dif- 
ferences (Table 2). Salinity and temperature are known 
to be major determinants of coastal and estuarine com- 
munity structure, in part because of salt tolerances 
and adaptations of associated flora and fauna (Day et 
al., 1989). Estuaries are often considered nurseries for 
juvenile fish and invertebrate species and it is likely 
that increased prey densities are related to the noted 
associations between cetaceans and estuarine areas 
(Ballance, 1992; Grigg and Markowitz, 1997; Harzen, 
1998). Selzer and Payne (1988) also documented a sea- 
sonal correlation between sea surface temperatures and 
salinities with Atlantic white-sided (Lagenorhynchus 
acutus) and short-beaked common dolphin ( Delphinus 
delphis) distributions and hypothesized that the in- 
teractions of salinity and temperature with sea floor 
topography and associated upwelling may be related to 
prey aggregations. Studies documenting turbidity and 
dissolved oxygen have been less common. Brager et 
al. (2003) assessed Hector’s dolphin ( Cephalorhynchus 
hectori ) habitat in relation to sea surface temperature, 
water depth, and water clarity and, although relation- 
ships varied by region and season, they found signifi- 
cant relationships between all three variables, for both 
individual variables and variables in combination. We 
also demonstrated that proportionally more foraging oc- 
curred in turbid waters, possibly indicating that higher 
levels of suspended sediments allow bottlenose dolphins 
to forage more effectively on prey that rely on visual 
