微生态分析方法的研究是近年来国内外热点研究课题。本文基于酵母微好氧（溶解氧<0.5mg/L）处理技术可行性研究探讨了细菌-酵母二元体系的微生态分析方法，重点探讨了荧光原位杂交——流式细胞术（FISH-FCM）和Biolog方法等微生态分析手段在复杂水处理系统微生态检测中的应用，并基于微生态研究阐明了进水浓度和HRT变化时酵母微好氧反应器中微生态的协同变化规律以及关于废水处理效果变化的微生态机制。（1）微好氧和无氧条件下批量培养试验表明：微好氧条件下微生物有很高的活性，能有效去除COD（去除率为91%）；无氧条件则导致生物量下降，COD去除率仅为12%。（2）微好氧连续小试结果表明：当进水TN和TP保持在1000和230 mg/L，进水COD浓度分别为11000、 15000和30000 mg/L时，COD去除率均稳定在70%-80%，氮磷去除率分别是18% - 68%和23% - 68%。随着进水COD浓度的上升，污泥浓度由2.0 mg/L上升到7.3 mg/L，COD比去除速度由2.3 kg/（kg•d）下降到1.7 kg/（kg•d）；当水力停留时间由42 h缩短到25 h时，COD去除率降至60%，污泥浓度从7.3 mg/L下降至6.0 mg/L，COD比去除速度从1.7 kg/（kg•d）上升至2.8 kg/（kg•d）。对连续小试出水进行批量再处理可使出水降至1000mg/L，总去除率可达90%以上（COD从30000 mg/L降至3000mg/L以下），降低出水MLSS浓度可以明显提高COD去除速度，单纯升高出水pH值对对去除效率无显著影响。（3）染色-FCM检测表明：同时适用于混合酵母样品（酵母假菌丝）和混合细菌样品（活性污泥絮体）的最佳超声分散条件为100 W 、60-90 s，过强超声条件（120 s）对酵母Candida tropicalis纯培养物（或细菌Escherichia coli纯培养物）的超声粉碎效果完全不同于混合酵母样品（或混合细菌样品）。（4）本文以C. tropicalis和E. coli分别作为酵母和细菌的模式微生物，采用双探针杂交的FISH-FCM技术检测了背景微生物（C. tropicalis或E. coli）浓度为107个/mL时二元体系中目标微生物（102~107个/mL）的数量。流式细胞仪能够明显区分噪音和二元体系中的酵母和细菌，因而一次进样时能够同时分析两种微生物数量，并且目标微生物浓度可以精确到104个/mL，此时微生物含量仅为0.1%，其中细菌浓度甚至可以精确到103个/mL（相应含量仅为0.01%）。然而，当二元体系中目标微生物浓度过低时(<104 个酵母/mL或<104 个细菌/mL)，背景微生物的存在会严重影响FISH-FCM技术的测量准确性。 （5）5种Biolog微孔板（GN、GP、ECO、YT、FF）所反映的代谢相似性聚类分析规律完全不同，其中ECO板所反映的代谢相似性聚类分析规律与PCR–DGGE提供的种群结构聚类分析规律更为接近，FF板所反映的代谢相似性聚类分析规律与PCR–DGGE提供的种群结构聚类分析规律一致；在检测细菌群落时，除Biolog GP多样性指数（P = 0.001）和Biolog GN多样性指数（P = 0.046）外，超低温冻存预处理对Biolog GN、GP、ECO代谢活性、丰富度指数和Biolog ECO多样性指数（H’）分析结果均无显著性影响(P > 0.101)，可以作为上述指标分析的微生物样品保存手段。在检测真菌群落时，超低温冻存处理影响显著影响Biolog YT代谢活性（P = 0.023）和Biolog FF多样性指数(H’)（P = 0.041），但对两种微孔板所反映的其它指数如代谢活性、丰富度指数(S)、多样性指数(H’)分析结果均无显著性影响（P > 0.132）。（6）FISH-FCM分析结果表明，在连续小试全部4个处理阶段中曝气柱污泥中酵母含量始终保持在＞99.9％的水平。随着进水COD浓度的上升，曝气柱污泥真菌群落结构多样性指数由2.05上升至2.19，代谢多样性指数由4.42上升至4.45；而随着HRT下降，真菌群落结构多样性指数和代谢多样性指数则分别从2.19和4.45降至0.79和4.36。作为提高进水有机负荷的主要措施，提高进水COD浓度和缩短HRT对于废水处理效果和曝气柱微生态存在截然相反的影响。

In recent years, many yeast species have been isolated from nature and used for biological treatment of those special HOEs under aerobic and open condition. However, high operational cost can be a drawback for the process due to high aeration requirements. In this dissertation, we demonstrated the feasibility of treatment of high-strength organic effluents at low DO level(0 - 0.5 mg/L). Beside, we hope to study Micro-ecology Mechanism on Microaerobic Treatment Performance at Various Influent COD Concentrations（11000、15000、30000 mg/L）or HRTs(42、25 h) by using fluorescent in situ hybridization (FISH) in combination with flow cytometry method (FCM) to enumerate the yeast and bacterial cells and using Biolog plates to investigate metabolic activities and communities similarity. However, we must solve some questions before applying these two methods to the wastewater treatment field, such as the pretreatment of the samples, the accuracy of enumeration and the choice of the Biolog plates’ types. Therefore, we start relative research in FISH-FCM and Biolog field.(1) Batch experiments show oxygen supply had an important effect on yeast growth and COD removal, DO levels of 0.1 - 0.4 mg/L, the microorganism grew well while COD removals were 91%, however, without oxygen supply, COD removal was only 12% following with the decrease of MLSS level. (2) In the continuous running at low DO levels, average COD, TN, TP removals of 67% - 88%, 18% - 68%, 23% - 68% occurred at influent COD, NH4+-N and TP levels of 11000 – 30000, 1000 and 230 mg/L, respectively. The influent COD: TN: TP levels played an important role in TN/TP removals. Microscopic observations showed no bacterial or mold contaminants in reaction column were observed over the experimental periods and the microscopic morphology of dominant yeasts varied. Increasing influent COD concentration brought on a rising MLSS (2.0~7.3mg/L) and a reduced specific COD removal rate [2.3~1.7 kg/(kg•d)]. Shortening HRT brought on a reduced MLSS (7.3~6.0 mg/L) and a rising specific COD removal rate [1.7~2.8 kg/(kg•d)].An attempt was made to further lower the COD level of the effluent from primary microaerobic treatment. MLSS adjustment make COD levels decline very fast along with the sharp increase of MLSS levels during 2 d. pH adjustment did not make significant differences in further COD removal in these batch cultures. These results demonstrate that further COD removal be efficiently achieved only after excess yeast cells were removed. (3) Mixed yeast samples with developed pseudomycelia and mixed bacteria samples, i.e. activated sludge flocs, were sonicated and investigated using the FCM following the staining with fluorescent dyes. The optimal sonication condition for both biological samples was 60-90 s at 100W. Excessive sonication treatment, i.e. for 120 s, had different effects on pure cultures of bacteria (or yeasts) and mixed samples of bacteria (or yeasts). (4) The simultaneous enumeration of yeasts and bacteria in binary samples was investigated by FISH-FCM using two probes (PF2、EUB338). The binary samples were prepared using Candida tropicalis and Escherichia coli as model yeast and bacterium microorganisms, respectively, and contained 107 cells/mL of background microorganisms and varied concentrations of the target microorganisms ranging from 102–107 cells/mL. Flow cytometry could clearly distinguish yeasts, bacteria and background microorganisms, and one injection provided enumeration results for both target microorganisms simultaneously. FISH-FCM effectively measured concentrations of the target microorganisms as low as 104 cells/mL for yeasts and 103 cells/mL for bacteria, corresponding to 1% and 0.1%, respectively. However, at concentrations < 104 cells/mL for yeasts and < 103 cells/mL for bacteria, the target microorganisms could not be accurately enumerated due to the presence of background microorganisms in the binary sample. (5) For bacterial community, the 3 Biolog plates (GN、GP、ECO) offered significantly distinct metabolic similarity cluster analysis results, and the result from ECO plate was close to the PCR-DGGE structure similarity analysis; For fungal community, FF plate offered a metabolic similarity analysis result absolutely different from that by YT plate. The Biolog FF metabolic similarity analysis result nearly corresponded to the PCR-DGGE structure similarity analysis result. With the exception of Biolog GN and GP diversity index (H’), the cryopreservation pretreatment had no significant influence on bacterial metabolic activities and richness index(S) expressed by 3 kinds of microplates, and Biolog ECO diversity index (H’). The cryopreservation had no significant influence on Biolog fungal metabolic activities, richness index(S) and diversity index(H’) (P > 0.132) except Biolog YT metabolic activity（P = 0.023）and Biolog FF diversity index (H’)（P = 0.041）. (6) Based on above study, the molecular microbiological technologies, including Fluorescent In-situ Hybridization - Flow Cytometry (FISH-FCM), Polymerase Chain Reaction - Denaturing Gradient Gel Electrophoresis (PCR-DGGE), and Biolog - FF assay method, were used to detect the variation of microbial community in the aerated column during the four pseudo-steady-state periods. The yeast contents remained >99.9% throughout the overall experimental periods according to FISH-FCM. Increasing influent COD concentration brought on structural (PCR-DGGE) / metabolic (Biolog FF) diversity index values of fungus community in the aerated column had an increase of (2.05~2.19) / (4.42~4.45). Shortening HRT brought on structural (PCR-DGGE) / metabolic (Biolog FF) diversity index values of fungus community had a decrease of (2.19~0.79) / (4.45~4.36). Increasing influent COD concentration or shortening HRT has an absolutely adverse effect on the microaerobic treatment performance and micro-ecology in the aeration column although either of them can increase influent COD loading up to a higher level.