Mapping & Ecology.
There has long been an important link between geography and ecology. From habitat modeling, to understanding individuals home ranges, to visualizing migration routes and important habitat corridors, learning where populations and individuals exist and spend their time allows us to improve management strategies for these species (Yamada et al. 2003; Burt, 1943; Rabinowitz & Zeller, 2010). When analyzing current and historic nesting data on Cumberland Island National Seashore (CUIS), it becomes apparent that there are certain stretches of beach that receive significantly more nests than others, but no analysis has been done on why this is the case.
In the past several decades with the popularization and continued improvement of Geographic Information Systems (GIS) and Remote Sensing Systems, the potential applications of spatial ecology have continued to grow. GIS and Remote Sensing make a multitude of new data available to ecologists, allowing for more in-depth models at higher resolutions (Kerr & Ostrovsky, 2003; Osborne, 2001).
Cumberland Island, Loggerheads, and Seaturtle.org.
Cumberland Island is the largest and southernmost barrier island in Georgia. Its 27.5 kilometers of beach is undeveloped, providing prime nesting habitat for sea turtles. Cumberland Island loggerheads are part of the Northern Recovery Unit of this species, believed to nest exclusively from North Carolina to the northern Atlantic Coast of Florida. On average, Cumberland accounts for 25% of the nests in the state annually. Four (4) species have been known to nest on Cumberland, including Loggerhead, Green, Leatherback, and Kemp’s Ridley sea turtles. The majority of nests (;95%) deposited on the Georgia coast are from Loggerhead sea turtles (Caretta caretta) and those will be the focus of this project.
Female Loggerheads reach reproductive age at 30-35 years and will typically nest every two to three years once reproductively active. Nesting females emerge onto the beach at night, dig a chamber where they lay an average of 100 eggs, and return to the ocean. Females will nest between four and six times during the years they are nesting, typically spaced fourteen days apart.
Nesting Site Selection theory’s & past research.
There have been a number of theories to what factors determine a good nesting site for females. Past studies have hypothesized that any number of environmental factors could affect nesting preference. The relationship between abundance of nesting females and site factors like beach temperature, sand moisture, beach size, and abundance of vegetation has been studied in depth (Bustard & Greenham, 1968; Mortimer, 1990; Stancyk & Ross, 1978). Later studies review geographic characteristics like beach slope and offshore Bathymetric features (Horrocks & Scott, 1991; Provancha & Ehrahart, 1987; Mortimer, 1982). Though never quantified, Mortimer (1982) suggests these features may play a more important role in nest site selection than other beach characteristics. It is notable that the importance of different beach characteristics has been known to differ between species, and site preference might even change to some extent between individuals (Whitmore and Dutton, 1985). However, this study aims to analyze the overarching trends seen within C. caretta populations nesting on Cumberland Island. It is consistently found that onshore beach slope is highly correlated with nest selection in several species including Green and Loggerhead sea turtles(Mortimer, 1982; Provancha & Ehrahart, 1987). Onshore beach slope is also often highly correlated with several offshore factors including offshore Bathymetry and wave energy (Bascom, 1964; Komar, 1976). Provancha and Ehrahart (1987) suggest that because of this, females likely determine a possible nest site prior to emergence, either based on offshore variables correlated to steep beaches, or perhaps they use these offshore parameters as an indication of steeply sloping beach.
While some studies have attempted to analyze the effect of these geographic beach features on turtle nest density, none that I was able to find do so on this scale. Provancha and Ehrahart (1987) looked at nesting densities of Loggerheads on Canaveral National Seashore in a similar manner, however due to logistical problems in the early year of the project many nesting data sets are approximate. This project was also done without the use of GIS software and overall taking another look at these claims with updated technologies and methods would be beneficial. Another study in 2012 analyzes this relationship across more than twenty beaches in southern Florida, where the total number of turtles nesting on each beach was compared to that beach’s attributes including onshore and offshore slope, elevation, and rugosity (Yamamoto et al. 2012). However, with no data as to where on each beach these nesting densities occurred, this study’s scale is much different than what I propose for Cumberland. Yamamoto et al. (2012) supports the idea that geographic features may play a greater role in nest site selection than sand characteristics, but additionally suggests that onshore topography might be more influential than offshore features. I have found no recent studies using the currently available tools to analyze the relationship between morphological characteristics and nest density across the same beach at these scales. The longevity, consistency, and abundance of data collected on Cumberland Island make it well suited for this study.
Research Goals Note For Mark: This is the primary area I am interested in your opinions on, namely, does the research described in the methods sound like it would be useful in answering or adding to a management question.
The goal of this research is to assess the link between offshore morphology and nest site selection for Loggerhead Sea Turtles. With a better understanding of this data, managers might be able to improve critical habitat models for this species and predict where higher concentrations of nests are likely to occur. A better understanding where Loggerheads are likely to nest on beaches has implications in both wildlife management and ecotourism.
Cumberland Turtle Survey Methods.
Cumberland Island monitors its beach for nesting turtles every day between the months of May and October. Surveys cover 27.5 km of beach on the east side of the island, starting at the south jetty and running to North Point. Surveys begin before sunrise to minimize the chance that tracks will be blown over and disappear. Technicians patrol the beach in a vehicle and look for turtle crawls (or tracks). When approaching a crawl, technicians assess if it is a false crawl, when a female did not nest, or a possible nest. At possible nests, technicians probe the area with a wooden dowel rod looking for density differences in sand indicative of the egg chamber. When the egg chamber is found, it is dug into to confirm the presence of a nest. One egg is taken for genetics research and preserved, the rest are recovered. A coarse mesh screen is staked down covering the nest as a method of nest predator mitigation and a numbered picket is secured behind the screen to label the nest. The latitude and longitude of each nest are taken at the centroid of the nest chamber using a Garman GPSMAP 64s. Locations are recorded to the fifth decimal place and are considered accurate within nine feet.
Technicians will relocate a nest if there is a high probability of inundation by the tide. A new egg chamber is dug higher in the dunes, attempting to closely match the dimensions of the original. Latitude and longitude are recorded for both the original nest site and the relocated nest. For the purposes of this project, only the locations of original nesting sites will be used during nesting seasons from 2009 to 2017.
Bathymetry models were created using archived editions of NOAA navigation chart 11489, which covers St Simons River to Tolamato River. These were georeferenced into ArcMap using a two-minute Lat/Long grid for true scale. Each depth measurement on the charts was manually entered in Arc as a point with its depth as a value, intertidal zones were considered “0”. Using a hand drawn polygon that traces the shoreline of the island as a mask, the “Topo To Raster” spatial anylist function was then used to derive a bathymetry map from those point clouds. This process was followed for editions 12, 37, 39, 40 of the NOAA 11489 Navigation chart, creating relative depth chart raster files for 1974, 2007, 2011, and 2015. Raster calculator was used to subtract more recent raster depth files from past, to assess the changes in offshore bathymetry over time.
Study Area & Analysis
This study only considers the 27.5km stretch of beach that is regularly surveyed. The study area was divided into 108 non-overlapping, .25x.25 km squares that follow the shoreline latitudinally and where the longitudinal centers overlap with the highest aggregation of nests within the study area (Fig. 1, Fig. 2). This gives a high resolution look at where nest concentrations occur. Nest locations were retrieved from Seaturtle.org databases for Cumberland Island. Using a spatial join function, the number of nests in each data block were counted. Originally, this data would have been broken up into groups by year, which would have corresponded to which Bathymetry map was most accurate for the given dates. However, in our study area no changes in bathymetry greater than one meter are shown between versions 37 (2007) and 40 (2015), which spans the entirety of the project’s focus. Though it is likely that this is not accurate and just a product of low resolution data, the charts still show the general trend in offshore morphology for the purpose of this project. Due to the lack of observed bathymetry changes, data was evaluated using a single map over the time frame analyzed.
To analyze offshore depth, each data block will have a 30-degree circular segment that represents the area of Bathymetry associated with a turtle’s approach for each block of nests. These polygons will originate from the center of each data block and extend east (Fig. 3). Three sets of analysis will be performed, one concerning up to 250m offshore, one concerning up to 500m offshore, and one concerning up to 1000m offshore (Fig. 4). I was thinking I might try this for some different angles as well, to try to account for the fact that we don’t know exactly where she’s approaching that spot of beach from. Maybe 30, 60 and 90-degree angles? This would give a total of 9 nesting approach models. What do you think of this idea.
For each area of approach, the following will be calculated using the tool Extract Data to Polygon: average depth, maximum depth, standard deviation (depth), average offshore slope, minimum offshore slope, maximum offshore slope, standard deviation (offshore slope), average offshore aspect, and standard deviation (offshore aspect), average change since 1970. Each of these features will be created based off the Bathymetry raster and using tools in ArcMap. The area of the intertidal zone will also be calculated for each block of nests. With the data now synthesized in the offshore approach polygons, their attribute table will be joined to the nest data blocks and then exported to Excel.
A Stepwise multiple regression function will be run in R to determine which, if any, of these variables are significantly affecting nesting site placement. Nest abundance will be the response variable, all other variables considered that have the potential to impact nest site preference will be predictors. Additional analysis will be performed as needed based on the results.
Figuer 1. Visulization of nesting “Hot Spots” on CUIS’s shoreline where darker red represents higher nesting densities and lighter red represents lower nesting densities, data used from 2009-2017.
Figure 2. The shoreline is broken up into 108 Quarter-Mile data blocks that will be the basis of analysis.
Figure 3. Close up of data blocks Figure 4. Example of the polygones that will be used to conduct
Including nests (red) and false crawls analasis of the offshore Bathymetry for 250m, 500m, and 1000m.
(blue) for refernce. Yellow stars represent the center of each data block.
Though some studies have looked at seafloor rugosity, roughness, and Bathymetric position indices as descriptors for offshore bathymetry that might affect a turtle’s approach, these measurements are much more suited to high resolution data sets such as LIDAR, which I did not have access to.
Though Bathymetric LIDAR data from 2017 exists for the study area, it is not yet available to access. Once this data becomes available, re-running the models used in this study with more accurate depth data would be beneficial and help validate the results of this paper. This would also allow for the inclusion of elements like those mentioned previously.
Though there have been major storm events to effect coastal Georgia since 2007, the NOAA depth charts used suggest no significant changes (more than 1 meter) in offshore Bathymetry to the area of interest off Cumberland Island between 2007 and 2015. Though it is likely that this is not accurate and just a product of low resolution data, the charts still show the general trend in offshore morphology for the purpose of this project.
Its important to note that this study does not consider other factors that may affect a turtle’s approach. For example, during shrimping season there are often large aggregations of shrimp trawlers off the south end of the island, an area that sees significantly less nesting. It is possible that the presences of trawlers impact turtles approach. It is also important to note that there are many factors that likely effect sea turtle nesting site preference that had not been taken into account in this study, including presence or absence of humans or predators on a given night.