A Study On The Use Of Macroinvertebrates And Benthic Diatoms In Water Quality Bioassessment

Use of benthic diatom and macroinvertebrates in water quality biomonitoring is an effective tool to assess water quality, stream ecosystem health and protect stream biodiversity. However, challenges to establish a sound bio-metric are still exist, such as:1) Are there any differences in the response of different organism to various disturbants, such as farmland and urbanization 2) How to build a reliable and robust biological index independ of the impact of natural environmental gradient? 3) The relative role of environmental variables at different spatial scales relating to the response of bio-metrics. We conducted the study and try to answer these questions in this thesis.We collected benthic diatom, benthic macroinvertebrates and environmenat vairables at 15 reference sites,27 farmland-disturbed sites and 12 urbanization-disturbed sites, in Xi’an segement of Wei Basin. We used nometric multidimensional scaling (NMDS) and analysis of similarity (ANOSIM), t-test and canonical correspondence analysis (CCA) to measure the response of diatom and macroinvertebrate to three types of disturbances. Our results showed that most of the chemical variables, habitat variables and percentage of land use types were of significant difference among disturbant types, as well as the compostion of diatom and macroinvertebrates. In addition, the six metrics of macroinvertebrates such as No. oftaxa、No. of EPT taxa、Shannon diversity index、1/Simpson index and Pielou index also showed significant difference among distrubant types. However, not all the diatom metrics such as PTI、%Eutraphentic, No. of taxa、Shannon diversity index、 1/Simpson index、Pielou index showed significant difference among disturbed types, especially between farmland and urbanization disturbed sites. CCA results showed that the critical variabes of NO2-, total phosphorus,%forest, average water depth and eltivaiton could explain 17.5% diatom assemblage variation, while 26.2% variation of macroinvertebrate assemblages could be explained by turbdity, temperature, pH, chemical oxygen demand, NO3-,%farmland and NO2-.We developed natural gradients variation adjusted multimtrics index (MMI.RF) using random forest method and un-adjusted multimtrics index (MMI.Ori) for Xi’an segement of Wei Basin based on diatom data. The performance evaluated by precision, bias, responsiveness and sensitivity of MMI.RF were better than that of MMI.Ori. The regression analysis showed that%urban,%farmland,NH4+ and TN were the key stressors impacting MMI variation. However, the explained variation of %farmland to MMI.RF (5.13%) was much less than to MMI.Ori (17.64%). This high difference of natural gradients among reference sites and farmland disturbed sites might contribute this variation.We also developed MMI.Ori and MMI.RF for diatom and mactoinvertebrates date at 78 sites in Qiantangjiang Basin. Results showed that both diatom and macroinvertebrate’s MMI.RF had improved performance in precision, responsiveness, and sensitivity, than MMLOri. The diatom’s MMI.RF (D.MMI.RF) had more improvement than macroinvertebrate’s MMI.RF (M.MMI.RF) in bias and responsiveness, but less improvement in precision and sensitivity. The explained variation by natural predictors to D.MMI.RF is 0, but it was still more than 10% variation was explained by natural predictors to M.MMI.RF. A total of 9 key environment variables could explain 82.83% M.MMI.RF variation, of which variables at reach scale explained 73.46% and 78.43% at watershed scale,the importance of watershed scale environment variables was a little more than reach scale. The 74.45% variation of D.MMI.RF was explained by 8 key environmental variables, and the most important spatial scale is reach scale (71.66%), followed by riparian scale (52.05%) and watershed scale (39.44%).