| Summary: | 碩士 === 國立臺灣大學 === 昆蟲學研究所 === 94 === 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
|
|---|
