Google Tag Manager icon


Are biofilms secretly lurking in your cooling water systems?

Water has excellent heat transfer efficiency, which is the driving force behind medium to large commercial and manufacturing facilities’ preference for water-cooled systems over their air-cooled counterparts. Ensuring optimal ambient air temperatures for guest comfort and precise process temperature control, water-cooled systems play a crucial role in various industries. 

Microbiological fouling is an ongoing concern in open recirculating cooling water systems. Biofouling can lead to corrosion damage, losses in heat transfer efficiency, and Legionella development. By detecting, eradicating, and monitoring biofilms using modern DNA tools, effective biocontrol methods will achieve superior results at lower cost.

Credit: H.Wang (1)

Key Underlying Principles

Legionella is a parasitic organism that requires the presence of certain amoeba and protozoa host organisms in a cooling water system to survive and thrive. Legionella infects and multiplies in host organisms until such point that they lyse, exploding out of the host often in thousands of new Legionella organisms, which then search for new hosts and repeat the process.

Legionella hosts require the presence of biofilms (slime layers) which can form in cooling water systems and adhere to metal surfaces. Biofilms have a diverse ecology, providing food for protozoa and amoeba. Biofilms facilitate Microbial Influenced Corrosion (MIC) and are more thermally insulating than calcium carbonate scale. When biofilms form on heat exchange surfaces they can cause significant increases in energy costs. Biofilms provide a layer of protection from biocides and disinfectants, enabling more resistant and resilient microbial communities.

Credit: B. Hayes/NIST

Case Study: DNA Insights into Cooling Water Biofilms

Microbe Detectives analyzed 3 cooling water samples using next gen 16S and 18S DNA sequencing and qPCR for an experienced industrial water treatment company and specialist in treating cooling water systems. Each sample was collected from a different cooling water system operating during a warm summer day and treated to prevent scaling, biofouling, and MIC.

Microbial control was described as poor for sample #1, excellent for sample #2, and good for sample #3. Summary bio-characteristics of these samples are provided in the below table.

Estimated Unique Identities

209 unique genera were observed in sample #1, 216 in sample #2, and 150 in sample #3, measured as Operational Taxonomic Units (OTUs).

Estimated Quantities

The largest estimated quantity of microbes observed was in sample #1 with ~39 million 16S and 18S gene copies/mL. Approximately 220 thousand were observed in sample #2, and 514 thousand in sample #3. These values correlate with the descriptions provided by the customer about the degree of microbial control in each sample.

Shannon Diversity Index (SDI)

SDI biodiversity observed was similar across the samples in the mid-to-high 2’s, on a scale of 0 – 5, with 5 representing the highest possible biodiversity value. This is the first clue that there may be more to the story. I would expect to see a greater difference in biodiversity between excellent microbial control of a cooling water system and poor control. Specifically, excellent microbial control should normally correlate to a lower SDI biodiversity value compared to a poorly controlled system. There are always site specific considerations to account for, however, as a good microbe detective, this provides reason to dig further.

% Relative Abundance of Kingdoms

  • Bacteria (59% – 71%) were the primary kingdom observed in all three samples, with a % rel. abundance of 71%, 59%, and 69% for samples #1, #2, and #3 respectively. The ecology of bulk recirculating cooling water with good to excellent microbial control should mainly consist of bacteria. Higher life forms are indicative of biofilms.
  • Protista (13% – 22%) were the second most abundant kingdom observed in all three samples, ranging from 13% to 22% rel. abundance. Protista is the kingdom of known Legionella hosts Acanthamoeba, Naegleria, and Vermamoeba (2). This discovery provided more evidence that biofilms may be lurking undetected in all three cooling water systems. In fact, the % rel. abundance of Protista in sample #2 (22%) was the highest observed. This is conflicted with the description of excellent microbial control for this sample. The % rel. abundance of Protista observed in sample #3 (13%) was about the same as in sample #1 (14%). Sample #3 was described as having good microbial control.
  • Fungi (9% – 11%) were the third most abundant kingdom observed in all three samples, with similar rel. abundance ranging from 9% to 11%.
  • Animalia and Plantae (2% – 6%) were the fourth and fifth most abundant kingdoms observed in all three samples, ranging from 2% to 6%.

The presence of Legionella hosts and other Eukarya are an indication of biofilm(s) lurking in the cooling water system.

Legionella was detected by a standard culture test and by 16S DNA sequencing in all three samples. Est. quantities of Legionella were about 7,300 in sample #1, 300 in #2, and 6,600 in #3. Legionella hosts, specifically Vermamoeba, were observed in all 3 samples. Est. quantities were about 500 in sample #1; 70 in #2, and 90 in #3 (18S gene copies/mL). 

Other Biofilm Indicators

As shown in the above table, slime forming and corrosion associated microbes were observed in all three samples. Slime formers help produce biofilms. MIC microbes often thrive underneath biofilms and cause underdeposit corrosion. The best way to interpret this data is by monitoring trends and observing results. In general, the goal is to minimize or eliminate slime forming and MIC microbes from the cooling water system.

Key Performance Indicators (KPIs) of Biofim Removal

Shannon Diversity Index (SDI)

  1. One way to verify biofilm removal from open recirculating cooling water systems is to first measure the SDI of the bulk recirculating water prior to shutdown for scheduled maintanance.
  2. During this scheduled maintenance, thoroughly review the system for potential biofilm locations. Look for biofilms in deadlegs, any strainers or filters, and any low or no-flow areas.
  3. Locate and completely remove the biofilm(s) from the system.
  4. Return the cooling system to operation and apply a strong biocontrol treatment to provide further assurance that the system has been properly restored.
  5. Once returned to normal treatment operation, re-measure the SDI of bulk recirculating water. You should find SDI values to be lower. Those values can then be used as an indicator or benchmark for excellent microbial control of the cooling water system when biofilms are known to have been eradicated.

Direct Measurement of Legionella Hosts

Generally speaking, if biofilms do not develop, Legionella hosts are not likely to develop because they need the diverse ecology of microbes and nutrients provided by biofilms to survive and thrive. If Legionella hosts do not develop in the system, Legionella are not likely to develop in the system.

Therefore, by detecting and tracking Legionella hosts, you are one step closer to a more predictive indicator of biofilm and Legionella development.

The Take Away

This data demonstrates that Legionella hosts can secretly develop in a cooling water system, even when microbial control is believed to be excellent. As demonstrated, if Legionella hosts are present, Legionella bacteria are likely to be present. Making things worse, underdeposit corrosion beneath biofilms, and loss of heat transfer efficiency are also threats lurking in the system.

By detecting, eradicating, and monitoring Legionella hosts and biofilms using modern DNA tools, effective biocontrol methods will achieve superior results at lower cost.

By applying modern tools such as DNA, cooling tower owners, operators, and service professionals, are in a much stronger position to detect and eradicate biofilms and prevent them from wreaking havoc.

Contact us today to learn more.


  1. ACS Appl. Bio Mater. 2023, 6, 8, 3213–3220, Publication Date: July 10, 2023, Copyright © 2023, American Chemical Society
  2. Boamah DK, Zhou G, Ensminger AW, O’Connor TJ. From Many Hosts, One Accidental Pathogen: The Diverse Protozoan Hosts of Legionella. Front Cell Infect Microbiol. 2017 Nov 30;7:477. doi: 10.3389/fcimb.2017.00477. PMID: 29250488; PMCID: PMC5714891.
  3. Chadee, Amanda and Skovhus, Torben Lund. “Linking Microbiologically Influenced Corrosion to Microbiological Activity UsingMolecular Microbiological Methods.” Materials Performance, May 2020.
  4. “Mapping the microbiome of… ” The Forefront, University of Chicago, November 2017.
  5. Thompson, , Sanders, J., McDonald, D. et al. “A communal catalogue reveals Earth’s multiscale microbial diversity.” Nature, November 2017.
  6. Keele, “Using eDNA to test for pathogens in reused water.” U.S. Department of the Interior, Bureau of Reclamation, September 2016.
  7. Ghylin, “DNA based microbial analysis detects and locates potential contamination in distribution systems.” Journal AWWA, March 2014.
Fecal contamination observed near wastewater plant along Florida’s coast was from non-human sources

Fecal contamination observed near wastewater plant along Florida’s coast was from non-human sources

Fecal contamination observed near wastewater plant along Florida’s coast was from non-human sources

A municipal wastewater treatment plant (City WWTP) fielded complaints regarding fecal matter on local beaches. In review of the potential sources of fecal contamination with Microbe Detectives (MD), and MD technical advisor Alison Ling, PhD, three had been contemplated including animal feces, human feces from the City’s homeless population, and the City WWTP.  The objectives of this sampling and analysis program were to:

  • Evaluate the presence/absence of overall fecal contamination and human fecal contamination in samples collected from stormwater outfalls.
  • Evaluate the relative contribution of human fecal contamination in samples collected.
  • Evaluate how presence/absence and relative contribution of human fecal contamination change with time.

Fecal Coliform Measurements

Fecal coliform and similar culture-based microbiology testing methods count how many bacteria grow on a specific type of food under specific laboratory conditions. As such, the term “fecal coliform” is misleading in that it describes bacteria that grow in a certain way rather than bacteria of a specific origin. Fecal coliform tests, including Standard Method 9222D, account for bacteria that can be present in animal feces as well as naturally occurring environmental bacteria. These methods were developed to indicate whether fecal contamination of any original exists in drinking water sources (typically very low in bacterial counts).  They routinely detect bacteria that are not fecal in origin, including bacteria common in plant materials and industrial effluents not associated with sewage (Reference 1).

Bacteria that proliferate during a controlled fecal coliform test can be found naturally in numerous environmental settings, including soils, sediments, algae, and lake water columns.  These “indigenous” fecal coliforms can be present in natural waters, even in cold climates.  When those waters are sampled, the fecal coliform laboratory method can sometimes indicate that fecal coliforms are present, in effect presenting a false positive (Reference 2).

The EPA states that “fecal coliform” test results do not meet the World Health Organization criteria for effective fecal indicators.  The EPA recommended in 1986 that states use either E. coli or enterococci tests instead of fecal coliforms to set water quality criteria, as fecal coliforms can present false positives with no association to fecal pollution (References 2, 3).

Molecular methods using qPCR Bacteroides spp. Markers

Due to these well documented method deficiencies, there has been an emphasis in recent years on the development of molecular biology methods to measure specific types of bacteria associated with fecal contamination rather than relying on culture-dependent methods with high false positive rates.  In addition to being more precise (only about 1% of bacteria can be grown in the lab), these molecular biology methods are less subject to false positives.  Specific methods used for evaluating fecal contamination are designed to measure specific types of bacteria that are prevalent in animal guts, but they are not good at growing in the environment, including Prevotella and Bacteroides genera.  Bacteriodes, specifically, include several types that evolve with the host animal and thus present host-specific targets for analysis.  These targets have been used to track the source of fecal contamination to different sources, including bovine, waterfowl, and human wastes (Reference 5).  Bacteriodes also have the benefit of comprising a significant portion of the bacteria present in human guts, and are 1,000 times more abundant in feces than fecal coliforms (References 2, 6).  Numerous studies have shown that Bacteriodes concentrations were better predictors of human pathogens than fecal coliforms (References 2, 7).

Bacteriodes spp. bacteria can be detected and quantified using quantitative polymerase chain reaction (qPCR) of the 16S ribosomal RNA gene (rRNA).  The 16S rRNA gene is the industry standard for identifying what types of microorganisms are present in a given sample, because it falls in the sweet spot of being conserved enough to be present in all organisms but variant enough to detect differences between specific species.  The qPCR method uses specific “primers” segments of DNA that bookend a targeted portion of the gene.  The primers are designed to be specific to the targeted organism (in this case Bacteroides spp.).  When combined with extracted DNA from the environmental sample and DNA replication enzymes and supplies, the qPCR process replicates DNA from only that target and back-calculates the initial concentration of the targeted DNA fragment.

Specific qPCR primers for general Bacteriodes (indicating fecal contamination) and host-specific Bacteriodes species and strains (Bacteriodes spp.) have been developed to assist with fecal source tracking efforts in environmental settings (References 5, 7). The HF183 marker targets a subset of Bacteriodes bacteria that evolved specifically with humans and is thus only present in human guts. HF183 was developed in 2005 (Reference 8) and an updated version of the assay is generally considered the most accurate method for detecting human fecal contamination (References 5, 9).

Samples collected during the first and fifth sampling events occurred after a rain event. All other samples were collected in the absence of rain before or during sampling. Fecal coliforms were analyzed by the City WWTP according to their methods. Host specific analysis for fecal bacteria were analyzed using molecular biology methods that target DNA of specific host-associated gut bacteria, specifically Bacteroides spp. (general fecal origin) and HF183 (human fecal origin).  The method involves collection of water in DNA-free containers, overnight shipment to the lab, filtration, DNA extraction, and quantitative polymerase chain reaction (qPCR) using primers developed for the specific targets.

Sampling Steps

Microbiology methods are especially sensitive to contamination from other materials. Samples were collected by authorized personnel at the City of Client WWTP, avoiding disturbance to the ground surface near or under the water close to the sampling location. The following steps were followed during sample collection.

  • Each sample container was submerged in the water at the outfall to fill the container. A one liter of water per sample was collected using sterile bottles provided by Microbe Detectives.  The inside of the bottle cap or rim of the sample bottle was not touched with anything.  If touched, the sample bottle was discarded and a new one was used.
  • The required volume of salt water was collected for each sample for fecal coliform analysis.
  • Gloves were used while sampling and were replaced often.  If gloves touched dirt or skin or clothing, they were replaced with new, clean gloves straight from the box.
  • Each sample bottle was clearly labelled with the sample name, Sample ID, date and time collected.
  • The labels on the bottles were covered with clear tape to protect them from getting wet and rubbing off from the ice and shipping.  “Bacteriodes spp.” and “HF183” were recorded on the COC for the host-specific fecal markers.


The below table summarizes test results from the first five rounds of sampling.

CT value reflects the threshold cycle number for a qPCR assay, which is the cycle number where the measured fluorescence exceeds a set threshold.  Smaller CT values correspond to larger concentrations of the target gene (Reference 11).

The below table compares results of the fecal coliform test versus the qPCR Bacteriodes spp test on the basis of whether fecal associated microbes were detected or not detected. A comparison of results is summarized by sampling event, and by sample location. The results confirm the stated testing inaccuracies of the fecal coliform test that is summarized above, as compared to the more accurate qPCR Bacteriodes spp test.


Results from the first five rounds of molecular testing suggested the following main takeaways:

  • Human-specific fecal contamination (Human Specific HF183) was not detected at any of the four sites during the five sampling events.
  • Samples from all four sites tested positive for fecal contamination using the molecular method (for Bacteroides spp.)
  • Positive fecal coliform results did not always correspond to positive fecal marker detection, suggesting that some fecal coliform tests may be subject to false positives.  Specifically, false positives were observed in the first sampling event (1 of 4) and the fifth sampling event (4 of 4). 
  • Positive fecal coliform and negative molecular results suggested a potential false positive fecal coliform test.  False positives on fecal coliform tests are a common concern and can be caused by naturally occurring bacteria from soil or other sources that can also grow on fecal coliform media.
  • In the third sampling event, molecular fecal markers were detected despite no detection of fecal coliforms in 3 of 4 sample locations.  This likely reflects the higher sensitivity of the molecular method over the culture-based fecal coliform method, and therefore is a false negative result.


  1. Doyle, MP and Erickson, MC. Closing the door on the fecal coliform assay. Microbe. 2006. Vol. 1, 4.
  2. US Environmental Protection Agency. Assessment of Fecal Indicators in Ambient Waters. 2015.
  3. World Health Organization. Guidelines for drinking-water quality. Recommendations – First addendum to the third edition. Geneva, Switzerland: World Health Organization.
  4. Zhang, Y. & Liu, WT (2019). The application of molecular tools to study the drinking water microbiome – Current understanding and future needs, Critical Reviews in Environmental Science and Technology, 49:13,  1188-1235,  DOI: 10.1080 /10643389.2019.1571351
  5. Ahmed, W, Hughes, B and Harwood, V. Current status of marker genes of Bacteroides and related taka for identifying sewage pollution in environmental waters. Water. s.l. : 8, 2016. Vol. 6.
  6. King, CH, et al. Baseline human gut microbiota profile in healthy people and standard reporting template. PLoS ONE. 2019.
  7. Savichtcheva, O, Okayama, N and Okabe, S. Relationship between Bacteroides 16S rRNA genetic markers and presence of bacterial enteric pathogens and fecal indicators. Water Research. 2007. Vol. 41.
  8. Seurinck, S, et al. Detection and quantification of the human-specific HF183 Bacteroides 16S rRNA genetic marker with real-time PCR for assessment of human fecal pollution in freshwater. Environmental Microbiology. 2005. Vol. 7, 2.
  9. Green, HC, et al. Improved HF183 quantitative real-time PCR assay for characterization of human fecal pollution in ambient surface water samples. Applied and Environmental Microbiology. 2014. Vol. 80, 10.
  10. Gronewold, AD and Wolpert , RL. Modeling the relationship between most probable number (MPN) and colony-forming unit (CFU) estimates of fecal coliform concentration. Water Research. 2008. Vol. 42, 13.

The Facultative Membrane BioReactor (FMBR)

The Facultative Membrane BioReactor (FMBR)

The Facultative Membrane BioReactor (FMBR)

The FMBR is a one-step, single tank solution for on-site wastewater collection, treatment and reclaimed water, suitable for commercial and residential areas, watershed pollution protection, and wastewater treatment plant upgrading. Compared to traditional sequencing batch reactors (SBR) and MBR systems, it operates with low energy and a small footprint. These design characteristics make for an efficient Decentralized Wastewater Treatment System (DEWATS). The FMBR is also designed for large scale centralized WWT systems and upgrades and has many proven examples.

FMBR was invented by Jiangxi JDL Environmental Protection Co., Ltd. (JDL) of Nanchang, China in 2008. JDL claims 47 invention patents across the USA, UK, France, Japan, China, and other countries, and over 3,000 systems installed and commissioned across 19 countries.

The ecosystem cultivated by an FMBR is designed to have the characteristics of a high concentration of microbes, long Solids Retention Time (SRT), and low Dissolved Oxygen (DO), which enables simultaneously removal of Carbon (C), Nitrogen (N), Phosphorus (P), and organic residual sludge in one reactor.

In the below 9 minute video, I present a summary of the FMBR on behalf of JDL for the New England Water Environment Association (NEWEA) in January of 2022.

FMBR Pilot Demonstration Project

The first FMBR Pilot Demonstration Project installed in the USA was in November 2019 at the Plymouth, Massachusetts Municipal Airport. This was made possible by winning a global competition hosted by the Massachusetts Clean Energy Center (MASSCEC) for wastewater treatment innovations that minimize energy consumption. It was granted the highest available funding of $150,000. The final report summarizing the results of the pilot can be found here: (FMBR Pilot: Final Report).

The FMBR Pilot was selected to replace an SBR wastewater treatment process due to high energy costs. The FMBR was sized to process 5,000 gallons per day (GPD) of wastewater generated by the airport and surrounding restaurants. Effluent discharge permit requirements were: Biological Oxygen Demand (BOD) < 30 mg/L, Total Nitrogen (TN) < 10 mgN/L, and Total Suspended Solids (TSS) < 30 mg/L.

Daily testing of influent and effluent for TN, P, BOD, and TSS showed strong performance of C, N, and P removal. Over one year of operation, the FMBR Pilot observed average daily P removal of: 10.0 mgP/L to <1.0 mgP/L; TN removal: 62.7 mgN/L to 4.1 mgN/L; BOD removal: 371 mg/L to non-detect; and TSS removal: 79 mg/L to non-detect.

As compared to the legacy SBR system (25K GPD capacity), the FMBR Pilot energy savings was estimated @ 70%, although on an “apples to apples” comparison to the SBR, the estimated energy savings was ~30%. Annual volume of biosolids requiring offsite disposal was reduced by an estimated 65%. Footprint was reduced by 75%. Installation and commissioning of the FMBR Pilot system was completed in 30 days.

DNA Study of FMBR Microbiome

Microbe Detectives analyzed 13 samples of the FMBR Pilot’s biology collected between May 2020 and May 2021 to help JDL better identify and quantify the FMBR Pilot’s nutrient removal biology. 16S DNA sequencing methods specialized for wastewater BNR analysis were applied. Results on the first 3 samples revealed Tetrasphaera at a high abundance. Tetrasphaera is a Denitrifying Phosphorus Accumulating Organism (DPAO), that can ferment, produce Volatile Fatty Acids (VFAs), and has shown to be a very robust PAO. Bacteria capable of Simultaneous Nitrification/Denitrification (SND) including Dechloromonas, Pseudomonas, and Thauera, were observed coexisting in the same facultative environment.

In a 2nd phase project, statistical summaries of the presence and abundance of BNR microbes observed in FMBR Pilot samples, were compared to 18 municipal wastewater BNR processes. These statistical summaries are referred to as “BNR benchmarks.” MD 16S DNA sequencing data of FMBR Pilot samples was aggregated with MD 16S sequencing database of 675 samples from 18 municipal wastewater BNR processes, that are dispersed across New England, Midwest, Southwest, Rocky Mountains, and West Coast geographies in the USA. All data was anonymized.

Key FMBR Microbes

Five BNR microbes were detected in all 13 FMBR samples. They include Tetrasphaera, Dechloromonas, Rhodocyclus, Nitrospira, and Pseudomonas. Each have known capabilities to remove C, N, and/or P.

Evidence of Simultaneous C, N, and P Removal

C – Removal

Fermenting bacteria remove C by decomposing organic waste and producing VFAs which is food for PAOs. The % abundance of Fermenting bacteria observed in FMBR Pilot samples averaged 4.2% vs 2.4% in BNR benchmarks. Tetrasphaera represented ~ 95% of fermenters and was identified in every FMBR sample.

N – Removal

An average 17.6% abundance of N – bacteria was observed in FMBR samples vs 6.3% in BNR benchmarks. Bacteria capable of SND represented 94% of N bacteria in FMBR samples.

Dechloromonas (avg. 8.3% in FMBR vs 1.0% in BNR benchmarks), Pseudomonas (avg. 8.1% in FMBR vs 3.1% in BNR benchmarks), and Tetrasphaera (avg. 4.0% in FMBR vs 2.4% in BNR benchmarks), were detected in all 13 FMBR samples, often at a relatively high % abundance. In other studies, Dechloromonas, a known phosphate-accumulating organism (PAO), has been observed in high abundance in alternating anoxic and aerobic conditions, enabling nitrification and denitrification (2, 3).

Pseudomonas has been observed to be capable of performing heterotrophic nitrification and then denitrify their nitrification products under aerobic conditions (4). Tetrasphaera is a Denitrifying Phosphorus Accumulating Organism (DPAO) that can decompose organic waste and produce VFAs, as a fermenting bacteria, and has shown to be a very robust PAO. DPAOs are of particular interest due to their ability to utilize internally stored cellular carbon both for P uptake and denitrification at the anoxic stage, resulting in increased phosphorus and nitrogen removal in BNR systems.

P – Removal

Phosphorus Accumulating Organisms (PAOs) were observed at an average 12.4% in FMBR Pilot samples vs 4.1% in BNR benchmarks. Overall PAO % abundance in FMBR samples was the highest among all benchmarks.

Key Missing Microbes

Key BNR microbes missing in FMBR samples, or at low levels, included Ammonia Oxidizing Bacteria (AOBs) (avg. 0.1%), Nitrite Oxidizing Bacteria (NOBs) (avg. 1.0%), and Nitrate Reducing Bacteria (NRBs) (avg. 0.1%). In a low DO SND condition, these microbes are expected to be missing because the environmental conditions are not suitable. This is additional evidence pointing to simultaneous nitrification/denitrification in a low DO condition as the principal N removal process of the FMBR.

Operating and Biological Data Tell the Same Story

Daily influent and effluent test data of C, TN, BOD, and TSS removal by the FMBR Pilot demonstrate strong nutrient removal capabilities. 16S DNA sequencing conducted in this analysis validates a high abundance of microbes that complete simultaneous nitrification/denitrification in a low DO condition, a high abundance of fermenters, that decompose organic waste and produce VFAs, and a high abundance of PAOs that biologically remove phosphorus. These processes occur simultaneously by facultative microbes that live in the same ecosystem, and compete for the same food. The significant energy savings is expected because of the low DO (<0.5 mg/L) condition. The reduction of biosolids production makes sense because of the greater abundance of fermenters that decompose organic waste and longer SRT. In the end, operational and biological data tell the same story – simultaneous removal of C, N, and P, in a single tank, with a surprisingly small amount of energy, footprint, and biosolids waste.


  1. Kristiansen, R., Nguyen, H., Saunders, A. et al. A metabolic model for members of the genus Tetrasphaera involved in enhanced biological phosphorus removal. ISME J 7, 543–554 (2013).
  2. Xu, D., Liu, S., Chen, Q. et al. Microbial community compositions in different functional zones of Carrousel oxidation ditch system for domestic wastewater treatment. AMB Expr 7, 40 (2017).
  3. Lifang Luo, Junqin Yao*, Weiguo Liu et al. “Comparison of bacterial communities and antibiotic resistance genes in oxidation ditches and membrane bioreactors,” Nature Portfolio, Scientific Reports (2021)
  4. Jin, R., Liu, T., Liu, G. et al. Simultaneous Heterotrophic Nitrification and Aerobic Denitrification by the Marine Origin Bacterium Pseudomonas sp. ADN-42. Appl Biochem Biotechnol 175, 2000–2011 (2015).
  5. Bergey’s Manual of Systematic Bacteriology, Edition Eight, by R.E. Buchanan and N.E. Gibbens, page 517.
  6. Strous M.; et al. “Deciphering the evolution and metabolism of an anammox bacterium from a community genome”. Nature. 440 (7085): 790–794. (2006).
  7. Lv, Xiao-Mei et al. “A comparative study of the bacterial community in denitrifying and traditional enhanced biological phosphorus removal processes.” Microbes and environments vol. 29(3): 261-268. (2014).