Recently, the marine mammal research team of the Institute of Deep Ocean Science and Engineering of the Chinese Academy of Sciences published a report entitled Modeling intraspecific variation in habitat utilization of the Indo-Pacific humpback dolphin using in Ecological Indicators Research paper on self-organizing map. This study uses a self-organizing map (SOM) to analyze the habitat utilization of three adjacent Chinese white dolphin populations in Hainan, Sanniangwan and Zhanjiang, and finds that there are obvious differences in distribution patterns and habitat utilization between different populations, and puts forward the hypothesis that "Chinese white dolphins may have two ecological types".
Ecotype is a taxonomic subspecies taxonomy, which refers to different populations of the same organism that have diverged and differentiated in a specific environment over a long period of time, and have different ecological, morphological and physiological characteristics formed by natural selection. The study of ecotype is conducive to exploring the process and causes of intraspecific differentiation and stereotyping, which is an important part of the study of the formation process of new species. The concept of ecotype is widely used in the geographical population differentiation of cetaceans. For example, bottlenose dolphins can be divided into coastal ecotype and offshore ecotype. The offshore type inhabits shallow offshore waters, while the offshore type can reach a depth of several hundred meters and a distance of more than 200 kilometers from the shore. The division of killer whale ecotypes is more complex, according to different food and distribution patterns, it can be divided into three ecological types: resident ecotype, transient ecotype and offshore ecotype. According to the different body size, external morphology and distribution area, killer whales in the southern hemisphere can be divided into four ecological types: type A, type B, type C and type D, of which type B can be further divided into two ecological types: big type B and small type B.
Chinese white dolphin is a small cetacean living in the shallow waters of the western Pacific Ocean and the eastern Indian Ocean coast, generally inhabiting shallow waters with a distance of less than 20 km from the shore and a water depth of less than 20 m, and was included in the first batch of national first-class key protected wildlife catalogue in the 80s of the 20th century. On the mainland, Chinese white dolphins mainly inhabit waters such as Ningde and Xiamen in Fujian, western Taiwan, Pearl River estuary in Guangdong, Zhanjiang and Shantou, Sanniang Bay in Guangxi and southwest Hainan. Based on the pattern of spots, Taiwanese populations have been classified as subspecies and assessed separately as "critically endangered", while little is known about the degree of differentiation between other geographic populations. Exploring the degree of differentiation between Chinese white dolphin populations is of great significance for assessing its endangered status, exploring the impact of human activities and formulating targeted conservation measures.
In recent years, with the vigorous development of machine learning and artificial intelligence technology, artificial neural network models have been widely used in the fields of animal ecology and conservation biology. As one of the most popular artificial neural network models, SOM can adjust its weights through competitive learning to achieve the purpose of training, and finally map the high-dimensional complex and difficult dataset to a simple and easy-to-understand low-dimensional map, and retain the topology of the original dataset. Using the same field survey methods and environmental factor collection methods, the research team introduced SOM into cetacean ecology research for the first time to analyze the population distribution and habitat use variation of Chinese white dolphins in Zhanjiang, Sanniang Bay and Hainan. Studies have shown that the population of Sanniang Bay has the smallest distribution area and only one core distribution area. Hainan has the largest and very scattered population distribution, with at least 4 independent hotspot distribution areas; The Zhanjiang population is somewhere in between, with two hotspot distribution areas, one concentrated at the entrance of Zhanjiang Port and the other scattered in Leizhou Bay (Figure 1). The SOM model further found that there were great differences in habitat utilization: Sanniang Bay and Hainan had low degree of intra-population segmentation but large differences between populations, the former inhabiting shallow water, low salt and low conductivity seas, while the latter habitat was characterized by relatively deep water, high salt and high conductivity. There is obvious internal segmentation in Zhanjiang population, that is, the sites are clustered into two distinct taxa, one group is distributed in turbid coastal waters, and the other taxa is superimposed with Hainan group, with deep water and high salinity environmental characteristics.
The results show that the distribution patterns and habitat use of Chinese white dolphins in the three regions have been significantly differentiated. At the same time, combined with the research results of the differences in cluster size, vocal behavior, morphological body color and genetics of the three geographical populations of Chinese white dolphins discovered by the scientific research team in the early stage, the study proposed for the first time in the world that Chinese white dolphins can be divided into estuarine and non-estuarine ecological hypotheses: estuarine type (Sanniang Bay population) is concentrated in the estuaries of large rivers, feeding on the abundant fish resources in the estuaries, the distribution area is small and concentrated, usually the cluster size is small, and the habitat is mainly characterized by low salinity and high turbidity; Non-estuarine (Hainan populations) inhabit seas that are not located in the estuaries of large rivers, have a wide and scattered distribution, usually have large clusters (improve predation success rates to accommodate relatively low levels of food resources), and habitats have high salinity due to lack of fresh water injection. In the setting of protected areas, the estuarine group is suitable for establishing a whole large area of protected area, while the non-estuarine type seems to be more suitable for integrating multiple small protected areas with connecting corridors. The above ecological hypothesis is only based on the currently available data, and the accuracy and universality need to be further verified.
The research work was supported by the National Natural Science Foundation of China and the Ocean Park Conservation Fund of Hong Kong.
Figure 1.Distribution map of Chinese white dolphins in southwest Hainan, Sanniang Bay and Zhanjiang waters using minimum convex polygon method, kernel density estimation method, grid counting method and hot spot analysis method.
Figure 2.The above figure shows the clustering of 290 Chinese white dolphin discovery sites in the self-organizing map at three study sites. The figure below shows the distribution of six environmental parameters (offshore distance, water depth, conductivity, pH, turbidity, salinity) in the self-organizing map, and the environmental factor characteristics of different taxa can be seen compared with the above figure.
Source: Institute of Deep Ocean Science and Engineering, Chinese Academy of Sciences
This account manuscript opens WeChat "Quick Reprint" by default
Please indicate the source of reproduction
For other channels to reprint, please contact [email protected]