Changes in Community Structures, Seriation, Cyclicity, and Diversity Patterns of Aquatic Insects under Environmental Impacts in the Upper Keelung River, Northern Taiwan

Aquatic insects were sampled at five sites along the upper Keelung River to examine the community structures, seriation, cyclicity, and diversity patterns. The upstream sites 1 to 3 significantly differed from the downstream sites 4 and 5 in taxa richness, density, Shannon diversity index, and the proportion of dominant taxon in the univariate analyses. Differences in aquatic insect assemblages at the upstream three sites were found using the two ordination techniques, canonical correspondence analysis (CCA) and non-metric multidimensional scaling (MDS). In the CCA, conductivity, biochemical oxygen demand (BOD), water temperature, and pH were the best water variables that explained about 42.9% of the variance in aquatic insect assemblages on the first two axes. In the MDS, chemical oxygen demand (COD) and BOD showed the highest correlation with aquatic insect assemblages in the upper Keelung River.
An engineering practice was proceeded at site 3 from August to September 1997. Differences in population- and community-based parameters between sites 2 and 3 were used in the Before-After Control-Impact with Paired-sampling (BACIP) design. The abundances of Stenelmis formosana in population-based parameters and the total abundances of aquatic insects in community-based parameters showed good abilities of impact detection. Result of the clustering analyses implied that it took two months for aquatic insects to recover from the construction.
Spatial seriation and temporal cyclicity trends were examined in the aquatic insect assemblages in the upper Keelung River from August 1999 to July 2000. The MDS plot exhibited significantly different distribution patterns between sampling sites based on both taxonomic and functional data. The index of multivariate seriation (IMS) in the seriation analysis showed that the longitudinal patterns based on the taxonomic structure at the five sites broke down under the disturbances, mainly elevated discharges. The longitudinal patterns based on functional compositions also showed more sensitive to elevated discharge. The cyclic patterns based on taxonomic data were more resistant to the elevation of discharge, which was detrimental to the patterns based on functional data. The engineering practice at site 5 appeared to be more influential on cyclic patterns based on taxonomic data.
Diversity patterns of aquatic insects were examined with different measures of alpha diversity, taxonomic diversity, beta diversity, and partitioning of diversity. The alpha diversity of aquatic insects decreased from upstream to downstream. The taxonomic distinctness (Δ*) accentuated the diversity at sites 2 and 3, and the average taxonomic distinctness (Δ+) provided a statistical test and indicated the impacted condition at sites 4 and 5. Similar spatial patterns of changes were recorded between βW and βT, and βI and βE over the sampling period. Partitioning of diversity divided taxa richness (TR), Shannon diversity (H’), and Simpson diversity (D) into components at each hierarchical level. For each sampling sites, mean within-point TR accounted for less than 30% of total TR at the five sites, while within-sample H’ and D accounted for more than 70% and 80% of total H’ and D. The results of partitioning of diversity suggested that the uses of Shannon and Simpson diversity for single sample at each site could give quite clear illustration. Replicated samples over a long period at each site were necessary for understanding the taxa composition. Partitioning of species diversity would be helpful in understanding of the factors controlling the spatial and temporal distribution of biodiversity, and of potential applications in conservation biology.