中華白海豚在台灣之分布預測與活動模式

Distribution Prediction and Ranging Pattern of Indo-Pacific Humpback Dolphins (Sousa chinensis) in Taiwan

葉志慧(Chih-Huey Yeh)

指導教授 ： 周蓮香

共同指導教授 ： 李培芬(Pei-Fen Lee)

國立臺灣大學/生命科學院/生態學與演化生物學研究所/碩士(2011年)

https://doi.org/10.6342/NTU.2011.01405

摘要

中華白海豚多出沒於25米水深以內，由於與人類活動區域高度重疊，白海豚遭受許多人為活動的干擾等威脅。台灣西海岸的族群目前僅剩不到一百隻，已於2008年由IUCN紅皮書列為「嚴重瀕危」的保育等級。本研究的目的為透過分析2006-2010年海上調查所得資料，了解其空間分布，棲地特徵及活動模式，分為兩大部分進行。
第一部份是由族群的層級來探討影響台灣中華白海豚分布的因子，並進一步預測其分布模式。在ArcGIS 9的操作平台上建立調查區域環境因子的網格資料及白海豚目擊的點位資料，分別以三個預測模式運算：廣義累加模式Generalized Additive Models (GAMs)，最大熵物種分布模式(Maximum Entropy Model, MaxEnt)及遺傳演算法(Genetic Algorithm for Rule-set Prediction, GARP)，篩選出能顯著區分白海豚出沒與否的因子組合及合適的推估模式。結果顯示網格內水深的最大值及標準差、鹽度標準差及離岸距離為影響海豚分布的關鍵因子。三種模式疊合分析後，預測在調查區域內中華白海豚的適合棲地僅侷限於沿岸水域，而在此棲地內，以苗栗縣南至大肚溪口(北)、雲林縣至外傘頂洲沿岸(南)之出現機率較高，另目擊率資料顯示為台灣中華白海豚兩個密度熱區。
第二部分由個體的層級比較於各年齡層(少、青、成、老年)及育幼與否於活動模式上的差異，包括範圍大小及棲地忠誠性兩個方面的探討。由已辨識出的71隻中華白海豚中篩選目擊十次以上之個體57隻，以最小多邊形法(Minimum Convex Polygon, MCP)計算海豚的活動範圍大小，平均MCP範圍為192.6 km2，另範圍平均南北長為69.1km。於各年齡層及個體育幼與否皆無顯著差異，然老年個體使用的MCP範圍偏小。棲地忠誠性方面，使用群集分析區分不同地域的使用者，發現分別有只侷限在兩個熱區的居留者，以及往返南北的遷徙者。計算個體於兩個熱區內的月目擊率，發現居留者與遷徙者於目擊率上的分隔不明顯；又以南區遷徙者的目擊率與居留者的差異不顯著─多數的個體傾向為密集使用此區但偶爾離開此區活動的遷徙者。相較之下，利用北區的個體傾向為當地的居留者或是偶爾拜訪的遷徙者。此外，曾育幼之個體有較高的區域目擊率。
結合空間分布預測及個體的棲地忠誠性分析得知中華白海豚在台灣西海岸有南北兩個主要分布的區域，為海豚所密集利用，確為台灣中華白海豚之重要棲地；另外曾育幼的個體對於這兩個區域的忠誠性也較高。未來應加入棲地利用的詳細研究，以對海豚於棲地內各區的偏好及其功能有更清楚的了解。

關鍵字

中華白海豚 ； 空間分布 ； 廣義累加模式 ； 最大熵物種分布模式 ； 遺傳演算法 ； 活動模式 ； 棲地忠誠性

並列摘要

Indo-Pacific humpback dolphins (hereafter Chinese White dolphins) inhabit in shallow coastal waters, almost within 25 m depth. Due to the proximity to areas of human activity, Chinese white dolphins have encountered tremendous threats caused by many anthropogenic activities. In Taiwan, the population size is no more than 100 individuals and the population has been listed as “Critically Endangered” (CR) by IUCN since 2008. With a view to providing practical information to conservation management, the purpose of this research is to identify critical regions within habitat, including finding key factors that influenced the distribution pattern in population level and investigating differences in ranging patterns among age classes and breeding statuses in individual level as described in the following two parts.
In the first part, habitat modeling techniques were applied to build prediction models. The environmental raster data within 2006-2010 survey coverage and occurrence of dolphins including sighting and tracking positions (n=2289) were input to ArcGIS 9. Three models were applied, including Generalized Additive Models (GAMs), MaxEnt (Maximum Entropy Modeling) and GARP (Genetic Algorithm for Rule-set Prediction). The appropriate models were selected, containing key factors including maximum water depth, standard deviation water depth, standard deviation salinity and distance to shore. An ensemble models was generated by overlaying these three models to obtain map of suitable areas for Chinese white dolphins in Taiwan. Predicted suitable areas were limited inshore, with the probability of occurrence was higher in coastal area of southern Miaoli County to Dadu River and coastal area of Yunlin County to Waisanding Sandbar.
In the second part, individual ranging pattern was investigated and comparisons were made between age classes and breeding statuses. For the range estimate, 57 individuals with more than ten sightings were selected from 71 identified dolphins. Minimum Convex Polygon (MCP) method and latitudinal length were the two estimators for the range size. The average MCP size was 192.6 km2, and the average latitudinal length was 69.1km. Old individuals tended to have smaller ranges despite non-significant result.
As for site fidelity, individuals were divided into residents and transients to north and south region respectively based on cluster analysis in range use, but the monthly sighting rate didn’t show clear stratification. Sighting rates of transients were close to those of residents in southern region, where most individuals tended to utilize as intensively as residents, but visited other areas on occasion. On the other hand, many dolphins tended to be either residents or transients that visited other areas frequently. In addition, females who have been observed calving have significantly higher sighting rates.
To sum up, the result of distribution prediction, ranging patterns and site fidelity revealed two suitable areas that were occupied by residents. Moreover, dolphins who used to be mothers used these regions more frequently. Specific research in habitat use should be carried out in future to obtain detailed information about habitat preference and possible functions of regions within habitat.

並列關鍵字

Chinese white dolphins ； distribution pattern ； GAMs ； MaxEnt ； GARP ； ranging pattern ； site fidelity

參考文獻

Akamatsu, T., K. Nakamura, R. Kawabe, S. Furukawa, H. Murata, A. Kawakubo, and M. Komaba. 2010. Seasonal and diurnal presence of finless porpoises at a corridor to the ocean from their habitat. Marine Biology 157:1879-1887.

Allen, M. C., A. J. Read, J. Gaudet, and L. S. Sayigh. 2001. Fine-scale habitat selection of foraging bottlenose dolphins Tursiops truncatus near Clearwater, Florida. Marine Ecology Progress Series 222:253–264.

Anderson, R. P., M. Gomez-Laverde, and A. T. Peterson. 2002. Geographical distributions of spiny pocket mice in South America: insights from predictive models. Global Ecology and Biogeography 11:131–141.

Anderson, R. P., D. Lew, and A. T. Peterson. 2003. Evaluating predictive models of species’ distributions: criteria for selecting optimal models. Ecological Modelling 162:211–232.

Andrews, K. R., L. Karczmarski, W. W. L. Au, S. H. Richards, C. Vanderlip, B. W. Bowen, E. G. Grau, and R. J. Toonen. 2010. Rolling stones and stable homes: social structure, habitat diversity and population genetics of the Hawaiian spinner dolphin (Stenella longirostris). Molecular Ecology 19:732-748.