Abstract
An accurate and precise estimate of mark rate (the proportion of individuals with marks suitable for photo-ID) is essential for adjusting estimates of abundance from capture-recapture models to include the unmarked portion of the population. Capture-recapture analyses on marine mammals typically assume that mark rate is constant over time, but this may not be true if management has decreased the frequency of interactions with fishing gear (a known source of marks). Considering that mark rate of Hector’s dolphins may have changed due to expanded set net protections around Banks Peninsula, this research had three primary aims: (1) to estimate an updated mark rate (for 2016), (2) to determine whether mark rate has changed since protection measures have expanded, and (3) to explore the implications of a low mark rate on estimating a key demographic parameter, survival rate. To determine the 2016 mark rate, several different mark rate estimation strategies were trialled and compared. Mark rate was then compared to an earlier period when set net restrictions were still relatively new (1992 – 1996). To investigate whether a decreasing mark rate affects precision of survival rates in Hector’s dolphins, I simulated capture histories by resampling from the population’s original capture history dataset. The comparison of different mark rate estimation strategies suggests that past studies underestimate the variance of mark rate by assuming simple random sampling of individuals, when in reality dolphins are encountered in groups (meaning observations are not truly independent). Whether a change in mark rate was detected between the two periods was method dependent. The frequentist method did not detect a statistically significant difference, but results from the Bayesian model indicated a 98% probability that the mark rate in 2016 (0.069, 95% HDI: 0.049 – 0.090) is lower than the 1992 – 1996 period (0.107, 95% HDI: 0.080 – 0.137). The simulation showed that although there was no evidence a lower mark rate would bias estimates of survival rate, it is likely to result in reduced precision. This may affect the ability to track population trends and determine whether current management is adequate.