Striped dolphin abundance in the Gulf of Corinth, Greece: dataset simulations help improve sampling design

Estimating wildlife population abundance is essential to evaluate the extinction risk of local populations and support conservation management. Photographic Capture mark-recapture (CMR) is a widely-used technique to assess cetacean abundance. However, reliability of estimates is strongly dependent on sample size. CMR software such as MARK can be used to simulate datasets sampled from virtual populations of known size. Population size can be estimated with CMR models from such datasets and, being the real abundance (Ntrue) known, is then possible to assess the models’ performance under different sample size scenarios. We used photo-identification data from June-September 2012 to estimate population size (N) as well as capture and survival probability (p and phi, respectively) of striped dolphins Stenella coeruleoalba in the semi-enclosed Gulf of Corinth. We then used these estimates to simulate datasets obtained with different levels of sampling effort. Finally, we estimated N from the simulated datasets and 1) compared the accuracy and precision of estimates obtained with closed and open population CMR models with samples of different size and 2) identified an 'optimal' sampling strategy. Reliable estimates (100% included in a ± 50 interval around Ntrue) with both models were obtained with 7+ capture occasions and a minimum capture probability of 0.5 for closed and 0.7 for open models. These results provide a way forward for improved sampling design.