Relationships between stream macroinvertebrates and environmental variables at multiple spatial scales

Summary 1. Aquatic communities are structured by multiple forces, and identifying the driving factors over multispatial scales is an important research issue. The East Asian monsoon region is globally one of the richest environments in terms of biodiversity, and is undergoing rapid human development, yet the river ecosystems in this region have not been well studied. We applied a hierarchical framework to incorporate regional and local environmental effects on stream macroinvertebrate communities in this region. The knowledge gained is expected to improve the understanding of the importance of spatial scale on regional and local diversity in the East Asian monsoon region. 2. A national data set of benthic macroinvertebrates and environmental variables (geographical, land-use, hydrological, substratum and physicochemical elements) in Korean rivers was used to determine the habitat preferences of macroinvertebrates. 3. Latitude, proportion of forest coverage, riffle habitat, silt substratum and temperature were the most important determinants for the ordinations of macroinvertebrate communities in each category evaluated by canonical correspondence analysis (CCA). The optimal habitats for stream macroinvertebrates are not the same for all species, and overall community metrics and abundance of sensitive species tended to be lower in open agricultural and urban streams than in forested streams. The sensitivity of mayflies and stoneflies to anthropogenic disturbances implicated them as good indicators to assess the effects of urban and agricultural activities. 4. A partial CCA was used to evaluate the relative importance of macrohabitat and microhabitat variables on community composition at three spatial scales (whole country, the large Han River basin and two small sub-basins in the lowlands and highlands). The majority of community variation (17–22% for each environmental element) was explained by macrohabitat variables at the regional spatial scale. In contrast, large proportions (15–18%) were explained by microhabitat variables at the local spatial scale. 5. Our findings indicate that the relative importance of habitat scales should be determined by geographical size and that comprehensive understanding of multispatial scale patterns can be important for implementing sound biodiversity conservation programmes.