台灣水鹿泥浴打滾行為與環境利用偏好：以林業及自然保育署嘉義分署轄區為例

Abstract

本研究從林業保育署嘉義分署轄區的台灣水鹿點位資料，分析水鹿環境利用的偏好，從自動相機監測發現水池是水鹿活動的熱點，但在大尺度分布模型中因為水池圖資的缺乏而未能顯現其重要性。因此，本研究透過3處水池樣點水鹿泥浴打滾利用的情形進行分析，探討水鹿泥浴打滾行為與水池環境受降雨變化的影響。 研究方法採用Ecological Niche Factor Analysis (ENFA)分析1*1公里網格中水鹿分布環境因子的邊際值與專一值，並藉由嘉義分署自動相機長期監測的資料作印證。水鹿泥浴打滾的資料則透過自動相機拍攝的個體進行辨識與計算活動模式，搭配鹿林山地區降雨指標與水坑比例的分析，並進行廣義線性模型的計算。 研究結果水鹿分布環境因子以海拔與年均溫的邊際值最大，但最具限制性的因子為網格內的道路長度，嘉義分署自動相機監測的水鹿熱點則在鹿林山與楠溪林道。鹿林山地區水鹿泥浴打滾的活動模式相近，中海拔水池較大且較無人為干擾的水池則有更高比例的水鹿日間活動與泥浴打滾利用。水鹿泥浴打滾的個體以鹿角為2叉3尖型態的成熟雄鹿為主，但也有亞成體雄鹿泥浴打滾的事件。統計鹿林山兩樣點的水鹿皆較偏好使用水坑(78%)而非泥坑(22%)，其中2023年是鹿林山3年度降雨最少與最長連續不降雨日的年份，反映在水鹿泥浴打滾的利用也減少。而每月的水鹿泥浴打滾的事件做GLM模型預測，CY15以雄鹿隻數做預測為最佳模型，CY17則以OI_3值、水坑比例與月均溫3個變數做預測較佳。 本研究證實水池是水鹿泥浴打滾利用的熱點，建議後續研究或經營管理上可針對更多樣棲地特徵或不同海拔區段的水池進行長期的自動相機監測與分析，以彌補大尺度空間分析中水池圖資缺乏的限制。

This study analyzes the habitat preferences of sambar deer within the jurisdiction of the Chiayi Branch, Forestry and Nature Conservation Agency. Data collected from camera traps indicate that pools are hotspots of sambar deer activity. However, due to the lack of spatial data of pools, their significance is not reflected in large-scale species distribution models.Therefore, this study focuses on three pool sites to examine the wallowing behavior of sambar deer and to explore how this behavior and pools influenced by rainfall variability. The methodology employs Ecological Niche Factor Analysis (ENFA) to investigate the marginality and specialization of environmental factors associated with sambar deer distribution within 1×1 km grids. The analysis is validated using long-term camera trap monitoring data from Chiayi Branch. Sambar deer wallowing activities are obtained by camera trap images through individual identifying and activity pattern calculation. The analysis is combined with rainfall indices and the proportion of puddle-type pools in the Lulin mountain area, and further evaluated using Generalized Linear Model. The results show that elevation and annual mean temperature had the highest marginality, while the most limiting factor is the total road length within each grid. Camara trap monitoring data showed that the sambar deer hotspots in Chiayi branch are in the Lulin mountain area and Nanxi Forest Road. In Lulin mountain area, sambar deer exhibited similar wallowing activity patterns, while the activity pattern of sambar deer in mid-altitude camera trap had higher proportion during daytime and more frequent wallowing use, due to the larger pool size and less human activities. Sambar deer that wallow are mostly mature males with two prongs and three tined antlers, but there are also juvenile males wallowing. Sambar deer in Lulin Mountain area prefer to use puddle (78%) rather than mud pit (22%). In the year of 2023, it was observed the least rainfall and the longest consecutive dry days in Lulin Mountain in the three-year observation, and this leads to the sambar deer wallowing events decreased. Generalized Linear Model (GLM) predictions of monthly wallowing events showed that at site CY15, the best predictor was the number of male deer. At site CY17, a combination of OI_3, the monthly puddle-type ratio, and mean monthly temperature provided the best predictive performance. This study confirms that pools are hotspots for wallowing behavior in sambar deer. It is recommended that further research or management can conduct long-term camera trap monitoring at pools with diverse habitat or altitude, in order to compensate the limitations posed by the lack of pool spatial data in large-scale distribution models.