Variations in Reproductive Indicators

The drop of the unmarked ratio in 2012 meant the number of the unmarked individuals was declining; in other words, the calf proportion was declining, which may in turn suggest a low birth and calf survival rate. The estimates of the crude birth rate further confirmed this phenomenon, with the decline occurring much earlier in 2010.

It should be noted that bias could occur in the estimations above when calves displayed negative reactions to the survey boat and thus decreased the capture probability and frequency. Although Taiwan’s Chinese white dolphins, like the population in Hong Kong (Jefferson 2000), was not considered to have obvious positive or negative behaviors towards small boats (Wang et al. 2007), the behavior of the calves remain still unknown, which may have biased these estimates, risking underestimates of reproductive indicators. But such bias would have been consistent over the survey years and should not significantly affect the general pattern found in this study.

In addition, the change of study area and survey team members also could bias the results here, the previous one could show different age structure among regions, especially it was known that old dolphins preferred Northern regions rather calves and

juveniles frequently appeared in South regions (Chang 2011), and the latter one could bring observation bias among different observers like the sighting rate of calves may be lower, and also bias the photo quality when too many new cameramen joined the survey, as the calves were generally little and swimming fast for camera shooting. The evidence of this bias came from personal communication of Dr. Chiou-Ju Yao, another Chinese white dolphins’ survey team leader, who encountered at least one neonate in Miaoli in 2012, indicated the bias existed in this study’s dataset. But it should be noted that the crude birth rate was only an indirect and “crude” reproductive indicator here, and the general decline was still worth noting considering even these bias.

To conclude, the birth rate of this population was confidently declining over at least the past three years. This decline combined with the increase in mortality mentioned above may be the main reason for the estimated abundance decline in the past three years.

3.4.4 Spatial and Temporal Variations in Population Estimates

The generally small estimates of abundance for the populations in nearby waters showed the limited distribution of Chinese white dolphins in each location. The abundance status of the Taiwanese population was quite similar to that of Xiamen, where the small population issue has been noted and a Nature Reserve has been established for the Chinese white dolphins (Chen et al. 2009). In Taiwan, a

2010. In Wang et al.’s study in 2012, it was mentioned that the survey efforts increased over time and, as the photo-ID process was begun in 2007, only the final year’s estimate was confident enough. On the other hand, Yu et al. (2010) and this study benefitted well by the solid photo-ID database assembled by Lien-Siang Chou’s team. Specifically the abundance estimates used only the data from 2008 and on, excluding the early data from 2005 through 2007 due to the limited effort amounts as well as the fast increase in the number of marked individuals in the early survey years to avoid the estimation bias. As the identified individuals were known to have already reached the plateau in 2008, , the resulting estimates could therefore be considered accurate, and this was supported by the consistent estimates between the best estimate (75) from Wang et al. (2012) and this study (75) for 2010, despite the huge difference in effort amounts. In this study, moreover, although the annual survey efforts also varied greatly due to different survey projects for different parts of the study area, the estimates remained quite stable even between 2008 and 2011 because of high capture/re-capture rate. Therefore, the results from this study could be considered as a valid reflection of the population trend of the Chinese white dolphin population in Taiwan.

Despite their use of the same dataset, also worth considering were two major differences between the present study and Yu et al.’s. First, the study in 2010 used Bailey’s modified Petersen estimation for simple closed populations, while this study used the robust design. Secondly, the unmarked ratios estimated by Yu et al. were 0.16-0.17 (2008-2009), but none of the estimates in this study were higher than 0.15 (0.027-0.13). This difference may have been caused by the different sampling methods.

Yu et al.’s sampling method was based on manually and deliberately selected original photos of large sighted dolphin groups, whereas this study utilized randomly selected photos cropped to show only the characteristics of the marked animals. The previous

sampling method may result in overestimation since it only focused on large groups which were naturally with much more calves, while the method in this study may cause underestimation if the calves were harder to be captured by camera-shooting due to small body size, mother’s protection behavior, or negative behavior. However, there were no evidence of negative behaviors of avoidance in calves, and in practical survey experiences this kind of trap-shy response was considered to be minor (Wang et al.

2012). Additionally, the low unmarked ratio estimates in this study were similar to Wang et al.’s study in 2012, so the estimates here were considered to be confident enough to correct the total population abundance and indicate relative birth rates in this study.