1  Introduction

1.1 Statement and Relevance of Problem

There is a small population of loggerhead sea turtles that choose to nest on Pensacola Beach. There are many studies that have looked into the nesting patterns of the loggerhead population else in Florida, but few have examined specifically within the Florida Panhandle. This research is part of an effort to fill in the gaps for the Florida Panhandle area.

The Florida Fish and Wildlife Conservation Commission (FWC) perform “turtle patrol” during nesting season and record observed turtle nests, however, it is difficult to determine where they choose not to nest. In order to observe an “absence,” researchers can use observed false crawls or computer-generated pseudo-absence points. A false crawl is when a turtle comes ashore but does not nest. Pseudo-absence points are randomly generated locations on the beach where no nest is observed. Both methods have been used in previous studies, but there is a lack of research comparing the two as the “no nest” observation.

In this study, we are looking to determine how false crawl and pseudo absence points compare as the “no nest” observation. We are also comparing multiple binary logistic regressions to a single multinomial logistic regression to determine, between the two options for modeling methods, if one stands out as a “better” model.

1.2 Literature Review

The Computational Geomorphology and Modeling Lab at UWF has been investigating nest sites of loggerhead sea turtles on Pensacola Beach for several years. In 2022 Jordan Iserman conducted the study “Beach Morphology Characteristics of Sea Turtle Nesting Sites: A Statistical Analysis” in which they compared the beach characteristics within 100 meters of a nesting site (1). For this study they only evaluated the beach characteristics of observed nests. The results of their Wilcoxon signed rank showed a difference in foreshore slope and beach slope (1).

Aiden Jensen conducted the study “An Exploration of Presence and Pseudo-Absence Data in the Analysis of Loggerhead Sea Turtle Nesting Behavior in the Florida Panhandle” in 2022 (2). The study was a pseudo simulation in which the dataset for the 10:1 ratio was resampled to generate different ratios of pseudo absence points to presence points. They used logistic regression to evaluate bias across ratios and found that as the number of psuedo absence points increased, the bias also increased (2).

In 2024, Cheyenne Long did “Modeling Loggerhead Nesting Patterns: How Many Pseudo-Absence Points are Necessary?” by doing a full Monte Carlo simulation study (3). They looked at 1:1, 2:1, 5:1, and 10:1 ratios and were able to specify the true relationships, then examined bias and mean square error of the resulting models. Like in Jensen’s results, their results showed an increased bias for \(\beta_i\) as the pseudo-absence point ratio increased (3).

Most recently, Angelina Scamardo worked on “From the Sea to Statistics: Using Machine Learning to Predict Sea Turtle Nesting Patterns” in which they compared machine learning methods to examine accuracy in predicting nesting status between presence and pseudo-absence points (4). The results of the study found that the random forest approach had higher accuracy, sensitivity, and specificity than the support vector machine approach (4).

Outside of Pensacola, there are several researchers who have conducted similar studies on the loggerhead population. Afford did the “Using multivariate analysis to determine characteristics of sea turtle nest selection along the Florida Panhandle” study in 2016 (5). They examined the nest selection of loggerhead sea turtles in Okaloosa county using pseudo absence points as the “no nest” sites. Using Bayesian logistic regression, classification trees, and random forest analyses they looked to identify environmental predictors of nesting (5).

Most recently, Hernandez conducted “Sea Turtle Nesting on Nourished Beaches With Different Construction Designs: A Case Study in Southeast Florida, USA” in which they examined nest selection of loggerhead and green sea turtles in South Florida (6). They used false crawl points as their “no nest” sites and employed descriptive statistics and zone-based proportions to compare, rather than inferential statistics (6).

Byrd (2022) examined environmental and anthropogenic drivers of loggerhead sea turtle false crawls on Jekyll Island, Georgia, using historical nesting records (2008–2019) and binary logistic regression analyses on contemporary field data (2020–2021) (7). They evaluated the effects of revetment proximity, lighting, sand temperatures, human presence, and beach obstructions on nesting status using false crawls as their “no nest” sites. Byrd found that false crawl rates were consistently and significantly higher near the large shoreline revetment, with additional context-specific effects from subsurface sand temperature and human presence, indicating that major anthropogenic structures are the primary contributors to elevated false crawling on this beach (7).

Manestar conducted a study examining the influences on the nest site selection of loggerhead sea turtles using unmanned aerial vehicle surveys and traditional methods in Boca Raton, Florida (8). With false crawl points as their “no nest” sites as well, they used logistic regression to identify predictors of nesting (8).

1.

Iserman J. Beach morphology characteristics of sea turtle nesting sites: A statistical analysis [Undergraduate Proseminar]. University of West Florida; 2022.

2.

Jensen A. An exploration of presence and pseudo-absence data in the analysis of loggerhead sea turtle nesting behavior in the florida panhandle [Undergraduate Proseminar]. University of West Florida; 2024.

3.

Long C. Modeling loggerhead nesting patterns: How many pseudo-absence points are necessary? [Undergraduate Proseminar]. University of West Florida; 2024.

4.

Scamardo A. From the sea to statistics: Using machine learning to predict sea turtle nesting patterns [Undergraduate Proseminar]. University of West Florida; 2025.

5.

Afford H. Using multivariate analysis to determine characteristics of sea turtle nest selection along the florida panhandle [MS Thesis]. Duke University; 2016.

6.

Hernandez L. Sea turtle nesting on nourished beaches with different construction designs: A case study in southeast florida, USA [MS Thesis]. University of South Florida; 2025.

7.

Byrd B. Effects of environmental and anthropogenic factors on loggerhead sea turtle (caretta caretta) false crawl rates on jekyll island, GA [MS Thesis]. University of North Florida; 2022.

8.

Manestar S. Using unmanned aerial vehicle (UAV) surveys and traditional methods to examine influences on loggerhead sea turtle (caretta caretta) nest site selection [MS Thesis]. Florida Atlantic University; 2023.