To tag ornot to tag: comparative performance of tagging and photo-identification in a long-term mark-recapture of Juvenile Green Turtles (Chelonia mydas)

Abstract Capture-mark-recapture (CMR) methods are widely used to estimate population parameters and to collect data on animal demography, migration, and life history. Sea turtle research programs generally use artificial tags, an invasive method. Photo-identification (PID) methods have become an important tool for animal identification. Herein, we assessed the effectiveness of a PID method for marking green turtles ( Chelonia mydas ) compared to traditional methods (artificial tags). As a part of a long-term CMR study, green turtles have been tagged and photographed since 2001. We analyzed 1917 captures with left and right side photographs of tagged turtles using Wild-ID software, these results were compared with tag-recapture data to assess error rates (false positives and negatives), and different effectiveness metrics. A combination of PID and tags (a match from either method was considered a recapture) was the most error-free and efficient criterion for identification of recaptures; however, it was the most time consuming and invasive criterion as well. We also assessed the effect of image quality indicators on the error rates of PID. We found that turtle cleanliness increases the similarity of images (indirectly related to false negatives), but we found no effect of sharpness, angle, light condition, or width and height in pixels of images on error rates. We could conclude that if image quality is improved, tags could be substituted by PID. However, we strongly recommend researchers to consider local situations (occurrence of by-catch or stranded dead turtles, for which tags are still necessary) before deciding to apply only PID.