A Retrospective Analysis on the Population Viability of the Yangtze River Dolphin or Baiji (Lipotes vexillifer)

The baiji or Yangtze River dolphin (Lipotes vexillifer, Miller), which inhabits only the middle and lower reaches of the Yangtze River in China, was the world’s most endangered cetacean species (Reeves et al., 1994; Reeves et al., 2000; Hu, 2003; Dalton, 2006). A multi-vessel visual and acoustic survey in 2005 of the Yangtze River system did not encounter any baiji, leading the team of international scientists to declare that the baiji was most likely extinct (Turvey et al., 2007). Population models have been successfully applied in species protection (Brook et al., 2000a, 2000b; Lancy. 1993; Chen et al., 1992; Zhang et al., 1994; Zhang et al., 1999) and fish resource assessment (Gayanilo et al., 1996, 2005; Gayanilo et al., 1997; Hacer., 2018; Wu et al., 2018). Research on population dynamics can help people understand the laws of population change and provide decision support and scientific basis for the endangered species protection. Based on this, we analyzed the population viability of the baiji in 1985 using Vortex model 10.3.7.0, to explore the possible causes of the functional extinction of the baiji, to provide guidance for the effective management of another freshwater mammal, the Yangtze finless porpoise (Neophocaena asiaeorientalis asiaeorientalis). This suffers from many of the same impacts that affect the baiji and the population of main stream of the Yangtze River have been continuously decreasing in recent years (Zhang et al., 1993; Wang et al., 1996, 1998; Mei et al., 2012; 2014; Huang et al., 2019). The latest research results suggest that the critically endangered Yangtze finless porpoises should be considered a unique incipient species (Zhou et al., 2018).

The baiji population was about 300 in the mid-1980s and only about 200 in the early 1990s (Chen et al., 1980; 1989; 1993; Lin et al., 1985; Zhou et al., 1980; 1982; Zhang et al., 2003; Turvey et al., 2007). Because of the small population size of the baiji, opportunities to observe them in the wild were very rare. The baiji’s scattered distribution and low density made them difficult to find and count. Population survey results could have large errors, Therefore, using mid-1980s as the starting point for the Yangtze River baiji population viability review, the relative error may be smaller. Lin et al., (1985) found that the baiji was distributed over 17 sections of the Yangtze River, each river segment had 8, 8, 8, 8, 21, 6, 4, 8, 8, 24, 11, 5, 6, 8, 12, 3, 8, respectively.
The total number of baiji was estimated to remain at around 156 individuals. Surveys during 1997-1999 provided a minimum estimate of only 13 animals (Zhang et al., 2003). The last authenticated baiji records were of a stranded pregnant female found in 2001 and a live animal photographed in 2002 (Turvey et al., 2007). The MVP concept emerged in 1981 from Shaffer’s pioneering paper. Operationally, time horizons of 50-100 years and extinction risk of 5% became the most frequently used criteria (Reed et al., 2003; Lacy et al., 2009; Curtis et al., 2011). The estimation method was: to analyze the population viability of initial population size, Initial population size, Ni in 100 years, when the survival rate of the population reached 95%, Ni was MVP.
        
The distribution range of the baiji was estimated to extended over 1,559 km (Lin et al., 1985). The baiji was population fragmentation, as the distances between used habitats and groups were increasing (Zhang et al., 2003), suggesting movement between 17 sections of the Yangtze River was now infrequent. So we set up each of the 17 sections diffused 1% to each other, with an equal probability of female and male and a survival rate of 50%.The baiji’s mating system was polygyny. Lifespan was about 30 years. The highest reproductive age was 20. The mature age of female was 8 and that of male was 6 (Zhang et al., 1994). The baiji had no record of producing twins in the wild, the proportion of newborn males was set at 0.5, 0.6 and 0.4 respectively, which constituted six simulated scenarios, as shown in Table 1 for details. The population growth of this species should be density-dependent (Zhang et al., 1994). Fowler (1981) considered that the parameter B of the change intensity, taking 2 could better simulate the growth of the density-dependent population and the parameter A of the decrease of the female mating rate also taking 2. For the baiji, if N was close to the environmental capacity K, the proportion of breeding females was 25% and when N was close to 0, the proportion of breeding females was 70% (Zhang et al., 1994). The proportions of old, adult and immature baiji individuals were approximately estimated at 57%, 26% and 17% respectively (Zhang et al., 2003). So the annual reproductive rate of adult female baiji could be about 30% (Zhang et al., 1994).


        
Zhang et al., (1994) the baiji set mortality rate of 30%-60% for 0-1 age group, 10%- 20% for 1-2 age group and 5%- 10% for the other age groups. An annual rate of population decrease was roughly estimated as 10% (Zhang et al., 2003). Therefore, we set the the baiji 0-1 age group mortality to be 60% for 0-1 age group, 20% for 1-2 age group and 10% for the other age groups. The baiji was the victim not of active persecution but of incidental mortality resulting from massive-scale human environmental impacts, primarily uncontrolled and unselective fishing (Turvey et al., 2007). So we set up two types of disasters with a frequency of 10% (Zhang et al., 1994), simulating the population dynamics at 95% and 95% survival and reproduction rates, respectively.
        
For the baiji which was distributed over 17 sections of the Yangtze River, we set up two kinds of simulation, one kind was each section had 8, 8, 8, 8, 21, 6, 4, 8, 8, 24, 11, 5, 6, 8, 12, 3, 8 heads of baiji respectively (Lin et al., 1985). The other kind was each section had double heads of baiji mentioned above, as shown in Table 1 for details. In addition to the above 6 simulation scenarios, another 3 simulation scenarios were constructed, for a total of 9 simulation scenarios. Scenario 7 based on scenario 4, with a diffusion survival rate of 100%. In scenario 8 and 9, based on scenario 4, were assumed to have 5% heads diffusion, equal female and male probability and 50% and 100% survival rates, respectively. Brook (2000b) believed that the prediction results of the population viability model were credible when the maximum amount ever distributed in history was taken as the environmental capacity. Zhang et al., (1994) set the environmental capacity of the Yangtze River for the baiji at 1,000 heads, taking into account the strong long-distance swimming ability of baiji, in order to conduct more reliable correlation analysis, we set the environmental capacity at 100 for each of the 17 each section. Environmental carrying capacity declined by 5% per year for 15 consecutive years, starting in 1985.
        
Inbreeding decline was one of the most important factors affecting the long-term survival of small populations. Ralls et al., (1988) studied the lethal equivalence coefficient of 40 mammalian populations and concluded that each diploid had 3.14 lethal gene equivalents. So we set 3.14 as the lethality equivalence coefficient of the baiji (Zhang et al., 1994). There was no breakthrough in the artificial breeding of the baiji. Therefore, it was unrealistic to supplement the baiji population in Yangtze river. At the same time, there was no regular fishing of baiji (Zhang et al., 2003). The Chinese government’s conservation plan involves capturing baiji in the Yangtze River and subsequently releasing them into a temporary refuge such as Tian-e-zhou semi-natural ex situ reserve (Wei et al., 2002). A female baiji was captured in the main stream of the Yangtze River on 19 December 1995 and was subsequently introduced into the semi-natural reserve. At the time of capture, she weighed 150 kg and measured 229 cm in body length (Wang et al., 1999), no other fishing reports have been seen since.

The baiji population viability retrospective analysis
 
It turned out that in six scenarios, the average extinction time of the baiji population was 22.5-28.9 years, between 2007 and 2013. Further simulation analysis revealed that, whether the total number of baiji was 156 or 312, the outcome was similar, the average extinction time of the former was 22.5-24.2 years and the corresponding extinction time of the latter was 26.6-28.9 years, as shown in Table 2 for details.


  
The average extinction time of the baiji population was between 2007 and 2013, this was more in line with the official announcement (Turvey et al., 2007). It was suggested that the simulation scenario had a better relationship with the historical habitat and the population characteristics of the baiji. Further simulation analysis revealed that, whether the total number of baiji was 156 or 312, the outcome was similar. Zhang et al., (1994) showed that, under natural conditions, the breeding rate of the adult female baiji was about 30% and the mortality from age 0 to 1 baiji was 30%-60% and from the results of the retrospective analysis, the mortality rate may be closer to 60%, therefore the direct cause of baiji ‘s functional “extinction” may be “one lower and one higher”, that was, the lower rate of the breeding for the adult female baiji and the higher mortality among newborn baiji (Zhang et al., 2003). Habitat destruction and so on, led to the baiji population passive fragmentation distribution, so that the “one low and one high” phenomenon continued to deteriorate (Turvey et al., 2007).
 
The minimum viable population of the Yangtze River baiji
 
On the basis of the above, we first adopt a single species model, it was found that, under existing conditions, single change of environmental capacity, initial number, disaster event and sex ratio could affect the population viability of Yangtze River baiji, but could not reach the MVP standard. The MVP with a breeding rate and the mortality from age 0 to 1 was shown Table 3 and 4 for details. Later, we set the breeding rate of the adult female baiji at 50%, the mortality from age 0 to 1 was 30%, MVP of the baiji with different environmental carrying capacity was obtained, as shown Table 5 for details.






    
In a relatively ideal state, the minimum MVP of the baiji was 61 heads, indicating that the relocation protection was very difficult for the baiji. With the deterioration in the ecological conditions of the Yangtze River because of expanding human activities and the likelihood that this trend may last for many more years, there may be little hope for the long-term survival of baiji in the river (Reeves et al., 2000).

This paper analyzed the population viability of the Yangtze River dolphin or baiji (Lipotes vexillifer) and 1985 was used as the starting point. It turned out that the average extinction time of the baiji population was 22.5-28.9 years, between 2007 and 2013. It was suggested that when the population was faced with the “Vortex” of extinction, focusing on the protection and restoration of high-quality habitats, especially those related to parents-children, was important for the protection of the endangered species. When the population was at risk of extinction, human-induced migration of the baiji and other species to specific natural habitats and strengthen conservation, may be one of the necessary means.