Use of a scent-detection dog for sea turtle nest monitoring of three sea turtle species in Florida

Abstract

Sea turtles are threatened with extinction around the world and rely on sandy beaches for laying their eggs. To protect eggs and locate them for calculation of reproductive success, beach surveyors must find the exact placement of each clutch. Eggs may be buried up to one meter deep under a nest mound several square meters in area. To locate sea turtle eggs, beach surveyors might spend hours searching for these eggs hidden in the sand, especially for difficult-to-locate leatherback (Dermochelys coriacea) and green turtle (Chelonia mydas) eggs. Scent-detection dogs (Canis lupus familiaris) are a novel tool that could provide a means to more accurately identify nests and efficiently locate eggs that need assessment, protection, or relocation. We assessed the effectiveness and feasibility of using a detection dog to locate sea turtle eggs buried in beach sand as compared to the traditional method using human beach surveyors. The detection dog was significantly more accurate in detecting loggerhead sea turtle (Caretta caretta) eggs and more efficient (less time spent and fewer holes dug) in assisting with locating the eggs. This case study presents results on the performance of one detection dog only, and additional research is needed with multiple detection dogs and handlers.

Fig 1. Detection dog final response behavior at green turtle nesting site.

The detection dog presenting a final response (sit) at a green turtle nesting site. The nesting site encompassed the disturbed area, with clear signs of sand being thrown into a mound between the up-crawl and the down-crawl. The eggs were located approximately 55 cm under sand where the dog’s nose is pictured. Printed under a CC BY license, with permission from Disney Worldwide Services, Inc., original copyright 2017.

Fig 2. Accuracy of the detection dog as compared to human surveyors.

Mean accuracy (distance to clutch) by sampling years and species (loggerhead and green turtle) comparing the detection dog with the human surveyors. Statistically significant differences (α = 0.05) between the detection dog’s accuracy and human surveyors’ accuracy are represented by the star symbol (*). All nests combined 2017 (Z₃₀₉ = -2.070, p = 0.038), 2018 (Z₄₂₈ = -8.218, p<0.001), both years (Z₇₃₈ = -6.450, p<0.001). Loggerhead 2017 (Z₁₈₉ = -1.105, p = 0.269), 2018 (Z₄₁₆ = -8.672, p<0.001), both years (Z₆₀₆ = -7.266, p<0.001). Green turtle 2017 (Z₁₁₁ = -1.130, p = 0.258), both years (Z₁₂₁ = -1.038, p = 0.299).

Fig 3. Search time of the detection dog as compared to human surveyors.

Mean search times by sampling years and species (loggerhead and green turtle) comparing the detection dog with the human surveyors. Statistically significant differences (α = 0.05) between the detection dog’s search times and human surveyors’ search times are represented by the star symbol (*). All nests sampled 2017 (Z₂₅₅ = -2.031, p = 0.042), 2018 (Z₄₂₈ = -6.973, p<0.001), both years (Z₆₈₄ = -3.783, p<0.001). Loggerhead 2017 (Z₁₃₅ = -3.869, p<0.001), 2018 (Z₄₁₆ = -7.324, p<0.001), both years (Z₅₅₂ = -4.662, p<0.001). Green turtle 2017 (Z₁₁₃ = -0.023, p = 0.982), both years (Z₁₂₂ = -0.021, p = 0.983).