Graduation date: 2007
For many species of marine turtle the characteristics that define pelagic habitat
have yet to be fully identified. A better understanding of these habitat characteristics is
critical to reduce high seas fisheries interactions with turtles, especially as the status of
many turtle populations has placed them on the threatened or endangered species list.
The combination of high-resolution satellite-tracking data with remotely sensed
oceanographic data makes it possible to identify habitat for loggerhead turtles by
analyzing the behavior of individual animals. Bycatch of loggerhead turtles in longline
fisheries can also be examined using the same high-resolution oceanographic data to
determine if there are identifiable habitat differences in high- and low- bycatch areas.
I analyzed the tracks of ten loggerhead turtles tagged in the spring and fall of
1998 near Madeira, Portugal in relation to the marine environment they occupied. To
determine the relationship between an individual turtle and its environment, some
measure of behavior was necessary. I calculated the straightness index (SI), the ratio
of the displacement of the animal to the total distance traveled, for individual weekly
segments of the ten tracks as a measure of individual behavior. I then extracted
information about the chlorophyll, sea-surface temperature (SST), bathymetry, and
geostrophic current of the ocean in a 20km buffer surrounding the tracks, and
examined the relationship between the straightness index and those characteristics
using logistic regression. Chlorophyll a value, bathymetry, and movement of the turtle
with geostrophic currents were consistently related to the straightness index of the
tracks of all ten animals (two-sided p-value from Wald's test: 0.005, 0.0017, and
0.0018, respectively). Tracks were less straight in high chlorophyll regions and in
shallower ocean areas, and animals were more likely to be moving with prevailing
geostrophic currents during straighter track segments. These results confirm
comparable analyses of loggerhead tracks in the Pacific, and indicate that sea turtles
alter their behavior (likely representing a shift from traveling to foraging) when they
encounter high-chlorophyll regions.
Turtles with highly sinuous tracks spend more time in a given area or habitat
than those who pass straight through, and therefore may be more susceptible to
incidental capture by fisheries operating in those habitats. To address the fisheries
bycatch/ habitat interactions I analyzed longline bycatch data to determine whether the
marine environmental variables identified in the first part of my study were related to
the probability of catching a turtle on a given longline set. I performed a logistic
regression analysis using bycatch of turtles as the response variable, and bathymetry,
SST, SST gradient (indicative of frontal activity), chlorophyll a, and chlorophyll a
gradient as the independent variables. I also included the location and the date of the
longline sets as potential predictor variables. I found that the most important variables
predicting the odds that a turtle would be caught on a given set were chlorophyll a
value in the area of the haul ( Wald's test, p=0.009) and the latitude at the beginning
of the haul (Wald's test, p=0.0005). Turtles were more likely to be caught on sets in
lower chlorophyll regions and in higher latitude regions of the data set, and there was
no indication of important effects of bathymetry. These results disagree with my
predictions from the tracking analysis, either because the fisheries-dependent bycatch
data set did not provide enough contrast of habitat types, or because bycatch
probability is not related to turtle behavior.
My results indicate a difference between the critical variables selected as
predictors of turtle habitat using the bycatch data and those selected using the behavior
of individual tracked animals. While bycatch information is important, the distribution
of fisheries data is highly biased towards frontal zones and regions of historic high
catch. Judgments about turtle behavior based on only fisheries interactions could lead
to incorrect conclusions about where animals spend the majority of their time.
Assuming that animals are more likely to have an increased probability of interaction
with longlines in areas where they spend more time foraging, fishing pressure should
be reduced in those areas of high-use for pelagic loggerheads. It is crucial to base
fisheries time-area closures and the design of marine protected areas on the behavior
of tracked animals, and not just on fisheries bycatch data.