Relative Roles Of Spatial Factors And Environmental Variables At Different Scales In Shaping Meta-Community Structure Of Stream Macroinvertebrate

Spatial distribution pattern and the driving mechanism of local communities is the major concern in community ecology. Spatial factors and environmental variables have been proved to play important role in shaping regional communities structure. Stream ecosystem is a typical hierarchical system. Therefore, stream community composition and structure are affected by environmental factors at different spatial scales. However, the relative roles of spatial and environmental factors, as well as different spatial scales in structuring community structure aredepend on communities, regions, and types of disturbance. In this thesis, we aimed to 1) explore the relative roles of spatial factors, environmental factors 2) explore the key environmental factors and spatial scales in shaping the benthic macroinvertebrate community and EPT (Ephemeroptera, Plecotera and Trichoptera) community structure at the Wei basin (tributary stream in xi’an region) and the Qiantang Basin.We calculated two spatial factors of Principal Coordinates of Neighbor Matrices (PCNMs), and Asymmetric Eigenvector Map (AEMs). Environmental variables consisted of natural predictors characterising natural gradient, and stressors resulted from anthropogenic disturbance. All the environmental variables were classified into three spatial scales:catchment, riparian and reach scale. We used forward selection, redundancy analysis (RDA) and canonical correspondence analysis (CCA) to get the significant spatial factors and environmental variables at different scales related with macroinvertebrate and EPT community. Then, the relative role of spatial factors and environmental variables at different scales on macroinvertebrate and EPT were decompsoed by variation partition or partial CCA (pCCA).A total of 67 EPT taxon and 194 macroinvertebrate taxon (hereinafter referred to as "species") were identified at 25 mountain sites and 23 plain sites in Xi’an segement, Wei Basin. Of which, mountain sites had 159 macroinvertebrate species and 64 EPT species, while plain sites had 96 macroinvertebrate species and 22 EPT species. There were significant difference in 54 natural predictors and 54 stressors between mountain sites and plain sites. The disturbance intensity was significant higher in plain sites than in mountain sites. Results of variation partition and pCCA showed that, pure environmental factors explained 17.9% variation of macroinvertebrate in mountain sites, much higher than pure PCNM(0.5%) and pure AEM(0.5%); pure reach scale environmental factors covered 16.8%, and yet pure riparian scale and catchment scale were both 0. Habitat score, Mg2+ concentration, average water width, bedrock percentage, average velocity and seasonal precipitation variance at site were selected as key environmental variables in reach scale. For EPT in mountain sites, pure AEM was responsible for 9.6% variation, far more than pure PCNM (0) and pure environment(0). Pure environmental factors explained 30.3% variation of macro-invertebrate community in plain sites, much more than pure PCNM(O) and pure AEM(0). Pure AEM covers 25.5% variation of EPT in plain sites, far more than pure PCNM(0) and pure environment(3.7%).A total of 60 EPT species and 258 macroinvertebrate species were identified in Qiantang Basin The results of variation partition and pCCA showed that, pure AEM had a higher explanation (17.1%) on macro-invertebrate than pure environment (8.0%) and pure PCNM(5.2%); and pure environment had much more explanation(21.3%) on EPT than pure AEM(1.1%) and pure PCNM(2.8%). Among three spatial scales, pure reach scale environmental variables covered 6.4% variation of macroinvertebrate, slightly higher than pure riparian scale(3.2%) and pure catchment scale(5.2%); and yet, pure reach scale covered 14.0%variation of EPT, much higher than pure riparian scale(0.8%) and pure catchment scale(3.6%).