오염하천 생태계 관리를 위한 생태정보학 기반 통합수질지수 연구

Abstract

A significant amount of various wastewaters including domestic and industrial wastewaters was increasingly discharged into the environmental since 1980s in South Korea which experienced a rapid population growth and a fast industrialization and urbanization, causing a serious ecological perturbation and destruction in aquatic ecosystems. Physico-chemical parameters have been conventionally used to monitor water quality, and then biological indicators were introduced to monitor water qualities in the late 1970s for streams, lakes and other water bodies (Wang, 2001). One of the major limitations in management of water quality in the field is that the management is performed based on maximum allowable concentration levels for control parameters, and a number of biological indicators which reflect a limited information of biological influences caused by the pollution.

Physico-chemical monitoring of aquatic environments is effective in providing a rapid and specific information for water quality management but it has a significant limitation in demonstrating an overall status of an aquatic ecosystem affected by the exposure of biological communities to the various pollutants. Therefore, it will be necessary to develop an intergrative water quality index which makes possible a comprehensive analysis of water quality based on an understanding of ecosystem function on spatial and temporal scales.

The purpose of this study was to develop an integrative water quality index for the management of stream ecosystems. In order to effectively evaluate the water quality for the ecosystems, it will be necessary to combine physicochemical parameters and biological parameters complementarily, which may comprehensively represent the quality of the ecosystem. This study, therefore, focused on development of an integrative water quality index that took into account the community structures of producers (algae), consumers (macro-invertebrates) and degraders (microorganisms) and their environmental factors as the ecosystem components.

Data for the biological and environmental parameters were periodically collected from 21 sites in the five different pollution sources (domestic sewage, farming, livestock, industry and restoration sites). The biological parameters included population densities of algae, macroinvertebrates and microorganisms, and the environmental parameters were atmospheric temperature, water temperature, pH, conductivity, DO, BOD, turbidity, water depth, flow rate, COD, TOC, T-P, T-N and NO3--N. The microbial community analysis was performed by PCR-DGGE (denaturing gradient gel electrophoresis) of amplified 16S rDNA fragments. Total data collected were 3,289 (13 items × 253 sites) for the environmental parameters, 21,505 (85 family × 253 sites) for macro-invertebrates, 29,095 (115 species × 253 sites) for microorganisms and 15,180 (60 species × 253 sites) for algae.

All the environmental and biological parameters were first processed through SOM analysis to pattern the complex ecological data sets. SOM patterning of sampling sites using macroinvertebrate densities as an input showed generally a good correlation between the site pollution status and their relevant species while there was little correlation in case of microorganisms and algae. In fact, the macroinvertebrate community was clustered according to the pollution gradient while microbial and algal communities appeared to cluster on a seasonal basis.

The SOM analysis was also performed for the representative indices calculated from these taxa data. This procedure was necessary to obtain an essential data useful for development of the integrative water quality index. The correlations between macroinvertebrate and algal communities were significant to a certain level while the microbial community showed little correlations with the other taxa. The correlation analyses between environmental parameters and the three taxa indicated that each taxon reflected the water quality independently. Microbial index has been barely developed and the microbial populations showed little correlations with the environmental parameters in this study, which made difficult the development of the index.

To circumvent this problem, an attempt has been made to develop a microbial water quality index that could represent the environmental status. For development of the index, microbial populations which could well represent the environmental status were first selected and characterized based upon multi-variate analysis (factor analysis) and the probability model of the microbial communities. This trial, however, was not successful because of non-linear responses of the microbial populations to the environmental parameters.

The factor analysis made it possible to extract a common factor out of the high-dimensional environment variables (parameters) that could well represent a pollution status of the environment. The data for the common factor were essentially classified into clean site () and relatively polluted site (), leading to a better elucidation of relationships between the common factor and the environmental parameters. Emergence patterns for the microbial populations were analyzed based on the probabilistic model derived from the microbial densities for each site. The microbial index for the water quality was developed by utilizing characters and emergence pattern of each microbial population. A water quality index based on macroinvertebrate community data was also developed in the same way. Each index developed from microbial and macroinvertebrate community data was evaluated against the measured environmental data to see if they could reflect the environmental parameters well. The evaluation results for the microbial and macroinvertebrate indices generally showed a good predictability for the water quality. However, their prediction capability becomes lower in the highly polluted area (e.g., industrial pollution sites) and the mid-level pollution (BOD, 14.65 ～ 35.08 mg/L ) sites. By the way, the microbial index was able to better predict the water quality in the extremely polluted (BOD, 44.66 ～ 76 mg/L) environment than the macroinvertebrate index. This is mainly because most of the microbial population data were collected from the clean sites and the highly polluted sites, and hence the data were less representative of mid-level pollution sites, and had more missing values than the data of environmental parameters and the other taxa. To avoid this kind of bias, it will be necessary to collect data from more diverse sites of different levels of pollution and to secure more diversified microbial populations that could represent the environmental pollution status more extensively. Furthermore, the macroinvertebrates cannot be a good indicator for the extremely polluted sites and oligotrophic sites since they could not survive at these sites.

The developed macroinvertebrate index was also compared with the indices of BMWP for macroinvertebrates to test and evaluate the efficacy of the index. It was shown that the site evaluation results based upon the indices developed in this study were similar to those based upon BMWP. It was concluded that the integrative index developed in this study was able to predict the water quality in the polluted streams when compared with the conventional BMWP index of macroinvertebrates used to monitor the water quality of the stream ecosystems.