溪頭地區北勢溪水棲昆蟲群聚結構及功能組成

Abstract

為取得防洪工程與生態保育的平衡達到河川治理的目的，溪頭營林區已採取生態工法整治園區內土石流危害之溪流。水棲昆蟲為溪流中主要無脊椎動物成員，其分布最主要受棲地環境因子的影響，故可以藉由水棲昆蟲群聚結構之改變評估棲地環境狀況。自 2004 年 3 月至 2005 年 8 月，在溪頭北勢溪設置七個採樣站，其中第 4 至 7 樣站以蘇伯氏定面積水網 (Surber net) 對淺瀨型之棲地 (riffle habitat) 進行定量採樣，而第 1 至 3 樣站由於缺乏淺瀨型之棲地，故以鑷子或手網進行30分鐘之定性採樣；唯第3站在2004年9月後由於流量增加，改以定面積水網採樣。各樣站計算水棲昆蟲之分類群豐度 (taxa richness)、密度 (density)、歧異度 (Shannon diversity index) 、均勻度 (Pielou’s evenness) 及優勢種比例 (dominant taxon ratio) 等群聚參數，以瞭解群聚之變動；並應用多維空間尺度 (multi-dimensional scaling, MDS) 及典型對應分析 (canonical correspondence analysis, CCA)之多變質排序技術 (ordination techniques)，以分析群聚結構在時間及空間上的變異，及與環境因子間之關係。同時利用理化指標之河川污染程度指標 (river pollution index, RPI) 及生物指標之 Hilsenhoff 科級生物指標 (family-level biotic index, FBI)、EPT index 評估水質之變化。各樣站並同時調查環境因子的變動 (水溫、溶氧量、電導度、酸鹼值、濁度、氨氮、硬度、總鹼度、溪流量、遮蔽度、懸浮固體值、硝酸鹽、磷酸鹽、生化需氧量及底質粒徑)。 定量樣站部分，由單向變方分析 (one-way ANOVA) 之結果顯示，水棲昆蟲各單變質群聚參數在各樣站間僅以密度及優勢種比例呈顯著差異；就整個研究區域而言，以四節蜉科 (Baetidae) 之 Baetis spp. 為優勢分類群 (佔27.74 %)，在科級亦是以四節蝣科為優勢之分類群 (佔43.25 %)；各項環境因子中僅有酸鹼度、流量及遮蔽度在樣站間呈顯著差異；水棲昆蟲之功能攝食群 (functional feeding groups) 組成比例在各樣站皆以聚集採食者 (gathering collectors) 為優勢群，碎食者 (shredders) 群豐度最低，樣站間以刮食者 (scrapers) 有顯著差異，在月份間各攝食群則皆呈現顯著差異；FBI及 EPT index在樣站間無顯著差異，FBI 水質等級 (water quality) 介於 「Excellent」與「Good」之間；群聚參數與環境因子相關性中，種類數及歧異度之變化皆與硝酸鹽呈顯著之負相關，且分類群豐度與水溫亦呈顯著之負相關。以 MDS 分析群聚結構之相似性，顯示因蚋科 (Simuliidae) 之Simulium sp. 及 Prosimulium sp.、短尾石蠅科 (Nemouridae) 之 Nemoura sp.、囊翅石蛾科 (Hydrobiosidae) 之Apsilochorema excisum 等種類多集中園區內第 3 站，因而將第 3 站與其他四樣站分離為園區內外兩群組的分布型態；CCA 排序圖亦因指石蛾科 (Philopotamidae) 之 Dolophilodes sp.、舌石蛾科 (Glossosomatidae) 之 Agapetus sp.、鱗石蛾科 (Lepidostomatidae) 之 Lepidostoma sp.、牙蟲科 (Hydrophilidae) 之 Pelthydrus sp.、糠蚊科 (Ceratopogonidae) 之 Atrichopogon sp. 及 細蚊科 (Dixidae) 等種類僅出現於第 3 站而反應相同趨勢；而單變值分析則無法顯示園區內外樣站間群聚的差異性。進一步以相似性矩陣的變方分析 (analysis of similarity, ANOSIM) 檢視各樣站間之群聚結構，由 ANOSIM 的分析結果為各樣站間的群聚結構呈顯著差異，此結果在單變質參數中無法得知；而配對檢測季節群聚結構差異，顯示春季與夏秋冬三季間呈顯著差異。以多變質方法分析影響樣站間水棲昆蟲群聚分布的主要環境因子，MDS 顯示以硝酸鹽及海拔高度二環境參數組合之排序圖與水棲昆蟲群聚排序圖具最高之相似性 (ρw = 0.450)，且硝酸鹽為各樣站共同具影響力之環境因子。而以 CCA 分析環境變數對物種變異值的影響，顯示第一軸萃取之物種的變異量為最大值，物種主要分布於第一軸，並指出影響物種分布之主要環境因子於第一軸為海拔高度、底質粒徑、水溫、遮蔽度為，第二軸為總鹼度、硝酸鹽，累積解釋的變異量約為 37.6 %；季節的變化在園區外較明顯，尤其夏季更為明顯。 半定量樣站部分，分類群豐度在樣站間無顯著差異；而以 MDS 分析有/無類型資料 (presence / absence data, P / A data)，顯示第 1、2 及 3 站在空間排序圖上之位置有明顯分群，以第 1、2 站間群聚相似度最為接近 (72.09 %)；以水溫及電導度與水棲昆蟲種類有無構成之排序圖具最高之相似性 (ρw = 0.202)。 全樣區部份，種類數在樣站間呈顯著差異；並以 MDS 分析有/無類型資料得知園區內第 1 及 2 站為一群，第 3 站半定量樣本混雜於第 1 站類群中，園區外第 4、5、6 及 7 站樣本及第 3 站定量樣本為另一群，顯示為暫時水域與永久水域之物種組成差異，或為定量與半定量採獲物種之差異；海拔高度之排序圖與水棲昆蟲有/無類型資料之群聚排序圖具最高相似性 (ρw = 0.679)。 關鍵字：水棲昆蟲、群聚結構、功能組成、生物指標、河川污染指標、多維空間尺度、典型對應分析、生態工法

In order to strike a balance between flood control engineering and ecological conservation, ecological engineering method has been applied to manage the debris flow streams in the recreational area. Besides, aquatic insects are one of the major groups of streams invertebrates. Their distribution pattern is affected by the environmental factors in habitat, and, therefore, we can assess the changes in habitat conditions by analyzing their community structures. Seven sites were selected in the Pei-Shih Stream in Shi-Tou, and aquatic insects were sampled monthly between March 2004 and October 2005. Suber net was used in quantitative sampling in riffle habitats at sites 4-7, while dipnets and forceps with fixed-time (30 min) were used in semi-quantitative sampling at sites 1-3 where were dominated by non-lotic habitats. At site 3, the sampling method was changed from semi-quatitative to quantitative approach in September 2004 when the discharge increased and the riffle habitats were available. The community parameters, such as taxa richness, density, Shannon diversity index, Pielou’s evenness and dominant taxon ratio, were calculated to study changes in community of aquatic insects at each site. Two ordination techniques, multi-dimensional scaling (MDS) and canonical correspondence analysis (CCA), were conducted to analyze the temporal and spatial variations of the community structures. River pollution index (RPI), Hilsenhoff’s family-level biotic index (FBI) and EPT index were also used to evaluate the variation of water quality. At quantitative sampling sites, the results of one-way ANOVA showed that significant differences were only found in density and dominant taxon ratio among sites. Baetis spp. were the dominant taxon (27.74 %), and the dominant family was Baetidae (43.25 %). Among the environmental factors, only pH, discharge and canopy showed significant difference among sites. Gathering collectors were the dominant group in the functional composition, and shredders were the least dominant. There were significant differences in the abundances of scrapers among sites. Functional groups showed significant difference among months. There was no significant difference in FBI or EPT index, and the water quality was between “excellent” and “good” based on the FBI index. Taxa richness and the diversity of aquatic insects showed significantly negative correlations with ammonia. The negative correlation was also found between the taxa richness and water temperature. The MDS plot indicated that site 3 was separated from the other 4 sites, because Simulium sp., Prosimulium sp., Nemoura sp. and Apsilochorema excisum were observed to be dominant at this site. The CCA ordination diagram revealed the same trend that Dolophilodes sp., Agapetus sp., Lepidostoma sp., Pelthydrus sp., Atrichopogon sp. and Dixidae sp. only appeared at site 3. However, univariate analysis could not show the difference in aquatic insect assemblages between sampling sites inside and outside the recreational area. Analysis of similarity (ANOSIM procedure) was further conducted to compare the community structures among sampling sites. The results revealed that the community structures among sites were significantly different. Pair-wise comparisons indicated that the community structures were significantly different between each pairs of spring and the other 3 seasons. Multivariate methods were used to analyze the main factors that shaped community structures of aquatic insects at each site. The MDS diagram based on the combination of ammonia and altitude showed the highest correlation with the ordination diagram of community structures of aquatic insects (ρw = 0.450). Besides, ammonia was the common influential factor at all sites. In the CCA diagram, the first axis constrained the maximum value on the variation of taxa and the data indicated that the major environmental factors affecting the distribution of taxa were altitude, particle size, water temperature and canopy. The second axis represented the sum of alkalinity and ammonium, and explained about 37.6 % of the cumulative variation. Seasonal changes were much apparent outside the recreational area, especially in summer. At semi-quantitative sampling sites, there was no significant difference in taxa richness among sites. The presence/absence data (P / A data) revealed that sites 1, 2 and 3 could be separated on the ordination diagram. The similarity of community structures was the highest between sites 1 and 2 (72.09 %). Water temperature and conductivity explained most of the P/A similarity (ρw = 0.202). All sampling sites were included in analyses using p / a data. Taxa richness was significantly different among sites. In cluster analysis, samples at sites 1 and 2, which located inside the recreational area, formed a cluster, while samples collected with dipnet at site 3 also merged into the cluster. Samples collected with Suber net at sites outside the recreational area could be clustered into another group. The difference in the assemblages could be attributed to the hydrology (permanent and temporary waters) or sampling methods. The MDS diagram based on altitude showed the highest similarity with the diagram of aquatic insect assemblages based on p/a data (ρw = 0.679). Key words：aquatic insects, community structure, functional composition, biotic index, river pollution index, multidimensional scaling, canonical correspondence analysis, ecological engineering