No abstract

Summary1. Marine protected areas (MPAs) have been advocated for the protection of threatened marine mammals, but there is no empirical evidence that they are effective. In 1988, the Banks Peninsula Marine Mammal Sanctuary was established to reduce gillnet mortalities of Hector's dolphin Cephalorhynchus hectori, an endangered dolphin species endemic to New Zealand. This study assesses the effectiveness of the MPA in improving the survival rate of Hector's dolphin at Banks Peninsula. 2. Over 21 years, we undertook photo-identification surveys of Hector's dolphins along standardized transects from small outboard-powered boats. From 1986 to 2006, we photographically captured 462 reliably marked individuals. We estimated mean annual survival during the pre-sanctuary and post-sanctuary periods by applying a Bayesian random effects capture-recapture model to the data. Population growth was estimated from population simulations using a stage-structured matrix model. 3. We estimate a 90% probability that survival has improved between the pre-sanctuary and postsanctuary periods, with estimates of mean survival probability increasing by 5AE4% (from 0AE863 to 0AE917). This improvement in survival corresponds to a 6% increase in mean annual population growth (from 0AE939 to 0AE995). 4. Synthesis and applications. Our study demonstrates improvement in a demographic parameter of an endangered marine mammal species following conservation action. Our results provide evidence that area-based protection measures can be effective for marine mammals. We note that estimating demographic parameters in marine mammals requires many years of data to achieve sufficient precision to detect biologically meaningful change. MPAs should be established with a commitment to long-term monitoring.

Knowledge about home ranges is essential for understanding the resources required by a species, identifying critical habitats, and revealing the overlap with anthropogenic impacts. Ranging behavior of Hector's dolphins (Cephalorhynchus hectori) was studied via coastal photo-ID surveys in the Banks Peninsula Marine Mammal Sanctuary (BPMMS) between 1985 and 2006. Univariate kernel density estimates of alongshore home range were calculated for 20 individuals with 15 sightings or more. For each individual, sighting locations were transformed into a univariate data set by projecting sightings onto a line drawn 1 km from the coast and measuring the distance along this line relative to an origin. Sightings were weighted by survey effort. Ninety-five percent (K 95 ) of the density estimate was used as a measure of alongshore home range, and 50% of the estimate (K 50 ) was used to reveal core portions of coastline where dolphins concentrated their activity. The mean estimates of K 95 and K 50 were 49.69 km (SE = 5.29) and 17.13 km (SE = 1.89), respectively. Four distinct hubs were apparent where the core areas of different individuals coincided. Three of the dolphins' alongshore ranges extended beyond the current northern boundary of the BPMMS, raising fresh concerns that the sanctuary is not large enough. Proposed changes to gill netting regulations, if enacted, will result in the alongshore ranges of all the dolphins in our study being protected.

No abstract

Populations of the once common bottlenose dolphin (Tursiops truncatus) in Argentina have precipitously declined throughout the country in the past decades. Unfortunately, local declines of common species are easily overlooked when establishing priorities for conservation. In this study, demographics of what may well be the last remaining resident population in the country were assessed using mark—recapture analysis (Pollock’s Robust Design) of a photo-identification dataset collected during 2006–2011 in Bahía San Antonio (Patagonia, Argentina). Total abundance, corrected for unmarked individuals, ranged from 40 (95%CI: 16.1–98.8) to 83 (95%CI = 45.8–151.8) individuals and showed a decrease over the years. Adult survival rates varied between 0.97 (± 0.037 SE) and 0.99 (± 0.010 SE). Average calving interval equalled 3.5 ± 1.03 years, with 3.5 births/year in the entire population and a minimum annual birth rate of 4.2%. However, data suggest that calves may have been born and lost before being documented, underestimating birth rate, calf mortality, and possibly the number of reproductive females. Either way, the recruitment rate of calves appears to be insufficient to support the size of the population. This population is relatively small and declining. Considering the disappearance of populations north and south of the study area, an incessant decline will have severe consequences for the continuous existence of this species in Argentina, indicating an urgent need for serious conservation efforts. This study provides insight into how the failure to recognize local population declines can threaten the national (and eventually the international) status of a common species like the bottlenose dolphin.

A temporal and geographical analysis of echolocation activity can give insights into the behaviour of free-ranging harbour porpoises Phocoena phocoena. Seasonal and diel patterns in the presence and foraging activity of harbour porpoises were investigated based on a year-long passive acoustic monitoring data set recorded at 5 sites in the western Baltic Sea. Diel patterns in detection rates were found at 4 sites. A year-round rhythm in presence, however, was found at only 1 station, whereas the other 3 stations showed diel rhythms for 2 to 3 seasons. Three of the sites showed diel patterns in foraging sequences on a seasonal level, but no station showed such patterns for the complete year of investigation. Both diurnal and nocturnal patterns in harbour porpoise detections were observed, indicating that diel rhythmic behaviour is more complex than previously reported. In contrast, foraging behaviour showed only nocturnal rhythms. Owing to the limitations in passive acoustic monitoring, all categorized foraging sequences are a minimum estimate. Therefore, classified foraging sequences are most likely pelagic foraging, while bottom grubbing could have been missed. Differences in the occurrence of foraging sequences between station, season and time of day lead to the assumption that the long-term echolocation diel patterns of porpoises strongly depend on the temporal changes in food availability and composition within a certain habitat. Echolocation behaviour of foraging porpoises is strongly influenced by seasonally available prey resources, which require adaptive foraging strategies. Therefore, owing to seasonal variations, analyses of diel patterns need to be conducted over sufficiently long time periods and large geographic scales to allow generalized interpretation of the findings. Consequently, no general conclusion regarding diel rhythms in harbour porpoise echolocation was found. We hypothesize that porpoises in the study area alternate between foraging on benthic prey in shallow waters at daytime and in the pelagic in deeper waters at night.

No abstract

Capture‐recapture techniques have been extensively used to estimate survival rates of Hector's dolphins at Banks Peninsula, but not abundance. We analyzed nine seasons of photo‐identification data using a model‐fitting approach in the computer program MARK, and then used MARK's estimates of capture probabilities to calculate the abundance of distinctive individuals. We extrapolated these estimates to include unmarked individuals using five seasons of data on the proportion of identifiable individuals in this population, obtained from “random photography.” This capture‐recapture approach suggests a 1996 population of about 1,100 (CV = 0.21). This is very similar to the 1997 line‐transect estimate of about 900 (CV = 0.28), especially considering that the two techniques do not necessarily measure the same thing. An important advantage of the capture‐recapture approach stems from the inherent versatility of photo‐ID data. If the sampling design is appropriate, an unbiased abundance estimate can be achieved as a spin‐off from work directed at other questions. However, in our view, line‐transect estimates are easier to interpret because the sampling design is explicit.