Sea Turtle Population Ecology
293
periodicity in ocean conditions, sampling error, and the underlying population trend
that is a cumulative result of density-dependent and density-independent forces that
drive survival, growth, and reproduction. Nesting beach numbers are our primary
response variable for assessing changes in sea turtle population size. Because female
turtles exhibit strong site fidelity, a nesting beach survey can give a good assessment
of changes in the adult female population, provided that the study is sufficiently long,
and effort and methods are standardized (Meylan, 1995; Gerrodette and Brandon,
2000; Reina et al., 2002). However, nest and nesting female abundance is often highly
variable from year to year because of environmental factors that affect female condition
and the intrinsic variability caused by the variable remigration intervals of individual
females (Meylan, 1995; Hays, 2000; Chaloupka, 2001). Simple regression analysis is
generally insufficient for estimating trends because of high variability or cycling in
nest abundance. Error in abundance estimation is likely when methods and effort are
not standardized. Increased effort to locate nests, either by biologists or through
improved public awareness, can result in apparent population growth or mask population decline. Surveys of index beaches over fixed time intervals can improve trend
estimation; however, female nesting activity may exhibit shifts in time and space,
making index surveys problematic (Godley et al., 2001). Trend analysis can be made
more accurate by incorporating mark–recapture information in remigration probability
(Kerr et al., 1999) and by statistical methods that incorporate uncertainty in extrapolation methods (Bjorndal et al., 1999; Reina et al., 2002).
A fundamental problem with nesting beach surveys is that they are unlikely to
reflect changes in the entire population. This is because of the long time lag to
maturity and the relatively small proportion of females on a nesting beach that are
reproducing for the first time, at least in populations with high adult survival rates.
Unknown adult and juvenile sex ratios also prevent extrapolation from nesting female
abundance to population estimates (Meylan, 1995). A decrease in pelagic juvenile
or benthic immature survival rates may be masked by the natural variability in nesting
female numbers and the slow response of adult abundance to changes in recruitment
to the adult population (Chaloupka and Limpus, 2001). Figure 11.2 illustrates how
a hypothetical nesting population might respond to a decrease in pelagic juvenile
survival. When random variability or a 5-year cycle of remigration rate is added to
the simulation, it takes many more years to detect the change in recruitment (females
nesting for the first time, a measure of cohort strength) on the nesting beach. Holmes
(2001) found that extinction risk estimates based on time series of nesting females
can be strongly biased, and advocated a weighted running sum method to reduce
the variance in time series caused by stage-specific counts.
There are many ways to assess population size and trends in abundance beyond
nesting beach surveys (Table 11.3). An alternative for assessment and monitoring is
to combine beach surveys with in-water surveys and absolute abundance estimates
with analyses of changes in survival rates (mark–recapture) and size or age distributions (e.g., Chaloupka and Limpus, 2001). All capture and census methodologies
include biases, and many are highly variable (Table 11.3). However, using a variety
of methods to assess changes in populations should improve our ability to detect
problems that may lead to population declines. Likewise, we may be able to eliminate
alternative hypotheses for our observations. For example, loggerhead turtles in North
1123 book.book Page 293 Tuesday, November 12, 2002 7:43 AM
293
periodicity in ocean conditions, sampling error, and the underlying population trend
that is a cumulative result of density-dependent and density-independent forces that
drive survival, growth, and reproduction. Nesting beach numbers are our primary
response variable for assessing changes in sea turtle population size. Because female
turtles exhibit strong site fidelity, a nesting beach survey can give a good assessment
of changes in the adult female population, provided that the study is sufficiently long,
and effort and methods are standardized (Meylan, 1995; Gerrodette and Brandon,
2000; Reina et al., 2002). However, nest and nesting female abundance is often highly
variable from year to year because of environmental factors that affect female condition
and the intrinsic variability caused by the variable remigration intervals of individual
females (Meylan, 1995; Hays, 2000; Chaloupka, 2001). Simple regression analysis is
generally insufficient for estimating trends because of high variability or cycling in
nest abundance. Error in abundance estimation is likely when methods and effort are
not standardized. Increased effort to locate nests, either by biologists or through
improved public awareness, can result in apparent population growth or mask population decline. Surveys of index beaches over fixed time intervals can improve trend
estimation; however, female nesting activity may exhibit shifts in time and space,
making index surveys problematic (Godley et al., 2001). Trend analysis can be made
more accurate by incorporating mark–recapture information in remigration probability
(Kerr et al., 1999) and by statistical methods that incorporate uncertainty in extrapolation methods (Bjorndal et al., 1999; Reina et al., 2002).
A fundamental problem with nesting beach surveys is that they are unlikely to
reflect changes in the entire population. This is because of the long time lag to
maturity and the relatively small proportion of females on a nesting beach that are
reproducing for the first time, at least in populations with high adult survival rates.
Unknown adult and juvenile sex ratios also prevent extrapolation from nesting female
abundance to population estimates (Meylan, 1995). A decrease in pelagic juvenile
or benthic immature survival rates may be masked by the natural variability in nesting
female numbers and the slow response of adult abundance to changes in recruitment
to the adult population (Chaloupka and Limpus, 2001). Figure 11.2 illustrates how
a hypothetical nesting population might respond to a decrease in pelagic juvenile
survival. When random variability or a 5-year cycle of remigration rate is added to
the simulation, it takes many more years to detect the change in recruitment (females
nesting for the first time, a measure of cohort strength) on the nesting beach. Holmes
(2001) found that extinction risk estimates based on time series of nesting females
can be strongly biased, and advocated a weighted running sum method to reduce
the variance in time series caused by stage-specific counts.
There are many ways to assess population size and trends in abundance beyond
nesting beach surveys (Table 11.3). An alternative for assessment and monitoring is
to combine beach surveys with in-water surveys and absolute abundance estimates
with analyses of changes in survival rates (mark–recapture) and size or age distributions (e.g., Chaloupka and Limpus, 2001). All capture and census methodologies
include biases, and many are highly variable (Table 11.3). However, using a variety
of methods to assess changes in populations should improve our ability to detect
problems that may lead to population declines. Likewise, we may be able to eliminate
alternative hypotheses for our observations. For example, loggerhead turtles in North
1123 book.book Page 293 Tuesday, November 12, 2002 7:43 AM