Wastewater treatment using MBBR in cold climates

Proceedings of Mine Water Solutions in Extreme Environments, 2015 April 12-15, 2015, Vancouver, Canada

Published by InfoMine © 2014 InfoMine, ISBN: 978-0-9917905-7-9

1

Wastewater treatment using MBBR in cold climates

Caroline Dale, Veolia Water Technologies Inc, Cary, USA

Marc Laliberte, Veolia Water Technologies Canada Inc, Montreal, Canada

David Oliphant, Veolia Water Technologies Canada Inc , Mississauga, Canada

Maria Ekenberg, Veolia Water Technologies, Lund, Sweden

Abstract

Biological wastewater treatment in cold climates can be a challenge due to the low reaction rates;

however the alternative of heating large flows of wastewater results in significant operating costs.

Using a fixed biomass instead of a suspended biomass allows the process to be operated at much

lower temperature as sludge age is no longer a designing parameter. The MBBR (Moving Bed Biofilm

Reactor) is a preferred example of such a fixed film (or “attached growth”) process. Biomass develops on

the inner surface of a carrier which is in continuous movement in the reactor.

Extensive laboratory trials have been undertaken using the MBBR process to demonstrate that long

term nitrogen removal can be sustained down to 4°C and selenium removal can be sustained down to 7°C.

Results from 2 separate studies treating mine effluents are presented.

The first case study refers to a 2 stage MBBR system treating synthetic water to simulate a mine

effluent containing 8 mg/l NH4-N and 18 mg/l N-NO3 where complete nitrogen removal was sustained at

4°C for a period of 3 months. The second case study refers to a trial using a mine effluent containing 25

mg/l N-NO3 and 45µg/l Total Se. A 2-stage MBBR was used to remove both nitrate and selenate at an

operating temperature of 7°C.

Introduction

The storage of tailings from mineral processing operations is achieved in large impoundment areas

(variously known as tailings ponds, tailings storage facilities or tailing impoundment areas). While

specific regulations vary between sites, mines are not allowed to discharge water acutely toxic to specific

species of fish such the rainbow trout (Oncorhynchus mykiss) or invertebrates such as Daphnia magna,

and discharge limits are being increasingly tightened. Treatment for the removal of metals, nitrogen

MINE WATER SOLUTIONS IN EXTREME ENVIRONMENTS, 2014 ● VANCOUVER, CANADA

2

(ammonia and/or nitrate and nitrite, principally contributed by ANFO explosive used in blasting) as well

as chloride, sulphate and/or total dissolved solids (TDS) is increasingly required prior to discharge to

meet regulatory standards.

Biological treatment has been applied for many years to remove nitrogen from municipal effluent,

even at cold temperatures. With a growing need for nitrogen removal from tailings ponds, laboratory

studies have been undertaken to determine whether nitrification/ denitrification could be sustained at very

low temperatures (<5°C) to reflect the conditions of Canadian mines.

In recent years, selenium present in mine effluents as selenate and selenite has also become a

concern. Selenium is highly toxic to aquatic life (Chapman et al, 2009) and the discharge limitations for

total selenium are becoming increasingly stringent. Some discharge criteria for release of mine effluent

into fresh water systems have been set to < 4.7 µg/l Total Se. Selenate is challenging and costly to remove

with physico-chemical methods, but biological treatment has presented itself as a viable alternative over

the past years. The process of biological reduction of selenate and selenite is very similar to

denitrification. Experiments have shown that a denitrifying sludge can be acclimatized to reduce Se

compounds (Takada et al 2008). During biological selenium reduction, microorganisms utilize selenate

and selenite as electron acceptors. Selenate and selenite are reduced to particulate elemental selenium.

Particulate elemental selenium can either be found as nanospheres inside the cell or external to the cell

(Oremland et al, 2004) and thus can be separated from the wastewater by traditional liquid-solid

separation methods. As with denitrification, biological treatment of selenate and selenite requires anoxic

conditions and the presence of an electron donor, usually an organic carbon compound.

Laboratory studies have been undertaken to use the experience gained in operating MBBR for

denitrification to develop a reliable process for selenate and selenite reduction. Operation at low

temperature has also been studied.

The MBBR Process

The Moving Bed Biofilm Reactor (MBBR) was developed in Norway specifically to achieve

nitrogen removal at cold temperatures. Since most Norwegian wastewater treatment plants are inside

buildings or underground caverns, a compact alternative to activated sludge was desirable when the

requirement for nitrogen removal was implemented for larger wastewater treatment facilities in the

1990’s (Ødegaard et al 1999).

The MBBR is a biofilm process which utilises a high density polyethylene carrier as biomass

support. The carriers provide a high protected surface area for biofilm development. The original

AnoxKaldnes K1 carriers provide 500m2/m3. Today, the preferred carrier K5, shown in Figure 1, provides

800m2/m3.


3

Figure 1 : K5 media, 800m2/m3

The carriers, which have a density of 0.95 to 0.98 kg/dm3, are maintained in continuous movement

using air in aerobic systems and mechanical mixers in anoxic systems. The process has been described in

past literature (Ødegaard et al 1994, Ødegaard et al 1999). Since the introduction of MBBR technology

in the 90’s, there have been over 100 plants built or retrofitted using this process in Scandinavia and over

750 installations worldwide. A simplified process schematic of the MBBR is shown in Figure 2.

Figure 2 : Simplified schematic of aerobic MBBR (left) and anoxic MBBR (right)

Extensive pilot studies were under taken on municipal wastewater to demonstrate prolonged

operation at low temperature (4 – 5°C for 3 months) prior to full scale installations (Rusten et al 2000).

During these trials, it was demonstrated that the nitrification rate could be controlled by regulating the

concentration of dissolved oxygen, such that increasing the DO concentration led to higher nitrification

rates in a well established system, allowing process performance to be maintained under a higher load or

at a lower temperature.

Tailings Characteristics and Impact on Wastewater Treatment Plant Design

As stated above, the chemical composition of tailings pond effluent will vary depending on the mining

operations and the ore being processed. A full effluent characterisation will therefore be required prior to


4

the design of the wastewater treatment plant. Ammonium and nitrate will usually be found in mine water

when explosives such as ammonium nitrate fuel oil (ANFO) are used for blasting. Further, the minimum

operating temperature will be used in the design of the biological process, impacting the removal rates

that can be achieved. Depending on their location, tailings ponds can experience wide temperature

variation throughout the year. Figure 3 shows a typical temperature profile for a tailings pond in Ontario,

Canada, where it can be seen that the water temperature falls and stays below 5°C for approximately 3

months of the year. In some cases, there will be sufficient holding capacity to allow the flow (and,

therefore, the nitrogen daily load) to treatment to be decreased during the winter months, while still

maintaining biological activity throughout the cold season to ensure that discharge guidelines are met.

Figure 3: Temperature profile of water pumped from a tailings impoundment area

The nitrogen removal rate in a biofilm system, typically expressed as g Nremoved per m2 of protected media

surface area per day (gN/m2.d), has been shown to be highly dependent on operating temperature (Rusten

et al 1995, Welander et al 2003). A study using fixed film biofilters to treat a mine effluent demonstrated

that a nitrification rate of 0.33 g NH4-N/m2.d could be maintained at 5°C (Zaitsev et al 2008). Typically,

the design of the MBBR will be done on the lowest operating temperature to ensure that there is sufficient

capacity to achieve treatment under the worst conditions unless heating can be provided to maintain the

design temperature. In Canadian mine operations, the minimum temperature at which nitrogen removal

will need to be maintained will frequently be below 5°C. Although a long term study of MBBR operation

downstream of municipal lagoons has shown that nitrification could be sustained at temperature as low as

1°C (Hoang et al 2014), little information was available on the removal rate achievable for mine effluents

at temperatures below 5°C until the study described below.


5

Methods

Nitrogen Removal

As previously described, complete nitrogen removal will require both nitrification (occurring under

aerobic conditions) and denitrification (occurring under anoxic conditions) to convert ammonium and

nitrate to nitrogen gas. The study was performed in a laboratory model of an MBBR-process consisting of

2 stages – first a nitrification stage then a denitrification stage. The nitrification reactor had a volume of

1070 mL and contained AnoxKaldnes media K1 to a fill ratio of 55 %. The denitrification reactor had a

volume of 1400 mL and contained AnoxKaldnes media K1 to a fill ratio of 50 %. The K1 media was

taken from a municipal wastewater treatment plant and thus had an active biofilm at start-up. After

several weeks of operation, the denitrification reactor was divided into two reactors to optimize the

denitrification rate; each with a volume of 700 mL and a filling degree of K1 of 50% (the same media that

was used when the denitrification reactor was one-stage), see Figure 4.

Figure 4: Experimental set-up for nitrification followed by two-stage denitrification. Insulation

was removed before the photo was taken.

Throughout the study, the temperature was controlled by circulating water from a cooling bath

containing glycol through the double walls of the reactors. At lower temperatures, the set-up was

insulated. In the nitrification stage, the media was mixed by the aeration only. In the denitrification

stage(s), magnetic stirring was used for mixing.

The process was fed a synthetic wastewater, the characteristics of the feed are given in Table 1. The

feed was kept in a refrigerator below the bench on which the set-up was placed, and pumped with a

peristaltic pump with a continuous flow to the nitrification reactor. The effluent from the nitrification


6

reactor was then pumped to the denitrification reactor with a continuous, slightly lower flow than the feed

flow.

Table 1: Synthetic feed water characteristics

Compound Concentration NaHCO3 (sodium bicarbonate) 0-0.5 g/L

KH2PO4(potassium-dihydrogen phosphate) 4.93 mg/L, corresponding to 1.12 mg/L PO4-P

NH4Cl (ammonium chloride) 30.53 mg/L, corresponding to 8 mg/L NH4-N

NaNO3 (sodium nitrate) 18.8 mg/L, corresponding to 30 mg/L NO3-N

Peptone extract 5 mg/L

Trace metal solution (see table 2) 0.5 mL/L

Table 2: Trace metals stock solution

Parameter Concentration MgSO4*7H2O 4.8 g/L

MnCl2*2H2O 1.6 g/L

CaCl2*2 H2O 5.8 g/L

CoCl2*6 H2O 0.48 g/L

NiCl2*6 H2O 0.24 g/L

ZnCl2 0.26 g/L

CuSO4*5 H2O 0.10 g/L

FeCl2*4 H2O 1.44 g/L

BH3O3 0.0005 g/L

Na2MoO4*2 H2O 0.0022 g/L

Na2SeO3*5 H2O 0.00114 g/L

Na3WO3*2 H2O 0.0014 g/L

Carbon source was added to the first stage denitrification reactor with a separate peristaltic pump.

Ethanol was used initially (day 1- 16) as it can be assimilated by a wider range of bacteria than methanol;

a faster start-up can thus be expected when using ethanol as carbon source. The carbon source solution

was switched to a mixture of methanol and ethanol from day 16 to 41 to allow the biomass to adapt to

methanol. After this period, methanol was used as sole source of carbon as it is the most commonly used

carbon source on full scale systems.


7

Selenium Removal

The studies were performed in laboratory models of MBBR-processes that were continuously fed

with a coal mine effluent, except from day 51- 68 when a synthetic feed was used due to a shortage of

mine effluent. The characteristics of the mine effluent are given in Table 3. The mine effluent was diluted

by 50% with tap water before being fed to the MBBR to reflect the average selenium concentration

observed at the mine.

Table 3: Mine effluent water characteristics

Parameter Concentration:Batch 1 Concentration:Batch 2

Soluble COD, mg/L 4 <20

Total COD, mg/L 5 <20

Total Se, µg/L 77 77

Soluble Se, (0.2 µm) , µg/L 78 72

NO3-N, mg/L 46 37.6

NO2-N, mg/L 0.04 0.022

NH4-N, mg/L 0.02

PO4-P, mg/L 0.06

Sulphate, mg/L 843 850

A 2-stage MBBR system was set up, each reactor had an operating volume of 600 mL, with a fill

ratio of media of 50 % K5 in each reactor. Virgin media was used and the process was seeded with

biomass from a passive Se-reducing system. The start-up was undertaken at a temperature of 15°C. The

temperature was then gradually decreased to 7°C. The process temperatures were controlled by

circulating water from cooling or heating thermostat baths through the jackets of the reactors. The

reactors were closed with rubber lids to ensure anoxic/anaerobic conditions and the media was mixed

with magnetic stirrers. Denatured ethanol was added as carbon source. The ORP was measured with

probes in both reactors.

Effluent samples were collected and treated batchwise with either filtration through 0.2 µm

membrane filters or with coagulation and flocculation using ferric chloride and polymer.

Process optimisation was undertaken by gradually reducing the HRT whilst maintaining the outlet

Se concentration below 5 µg/l.


8

Analytical Methods

The progress of nitrification and denitrification was followed by measuring ammonium (NH4-N),

nitrate (NO3-N) and nitrite (NO2-N), while the removal of organic material was followed by measuring

chemical oxygen demand (COD) and biochemical oxygen demand (BOD5). Effluent total suspended

solids (TSS) were also measured. To check that phosphorus was added in sufficient amounts, phosphate

(PO4-P) was measured on the effluent. The alkalinity was also measured regularly, given the importance

of this parameter on the nitrogen removal processes.

Samples for measurement of nitrogen compounds, phosphate and COD were taken as grab-samples

and analyzed directly after sampling. Samples for effluent BOD and TSS were collected over a period of

1-3 hours. TSS were analyzed directly after sampling. COD and BOD5 were also measured after filtration

through a filter with a pore size of 10 µm (Munktell V150).

COD was analyzed using Hach-Langes test LCK 114. BOD5 was analyzed according to SS-EN

1899-1:1998. ((SS= Swedish Standard). NH4-N, NO3-N, NO2-N and PO4-P were analyzed with

HachLanges methods LCK 303/304 and LCK 349 respectively. Sulphate was measured with HachLanges

method LCK 153. Alkalinity was measured as mg/L CaCO3 using HachLanges methods LCK 362. TSS

were measured according to SS-EN 872:2005. All these analyses were made in-house.

All analyses of NH4-N, NO3-N, NO2-N and PO4-P were performed after filtration through glass-fiber

filters with pore size 1.6 µm (Munktell, MGA).

Total selenium was analysed using ICP-MS/AES. Samples for selenium speciation were analysed

with HPLC-ICP/MS.

The following samples were prepared for analyses of total selenium:

• Feed: not filtered (taken from the feed vessel)

• Effluent from MBBR1 outlet. This sample was collected by opening the connecting tube

between the reactors and then collecting the effluent until sufficient sample volume was obtained.

The sample was then filtered first through a glass fiber filter (1.6 µm (Munktell MGA) and then

through membrane filters (0.20 µm Sartorius, cellulose-acetate).

• Effluent from MBBR2 outlet (two samples). The effluent was collected over a couple of hours

to obtain sufficient volume. A 50 mL sample was then filtered first through glass-fiber filters and

then through membrane filters as described above- this is referred to as filtered effluent. A 400

mL sample was treated with coagulation and precipitation. The sample was left to settle. A

sample of supernatant was taken for analysis – this is referred to chemically treated effluent.


9

The pH, temperature and dissolved oxygen concentration (DO) were checked with hand-held probes

(“HachLange HQ40d” for DO and temperature and “Eutech pH 5” for pH). ORP was measured with a

HachLange HQ40d instrument.

The flow out from the process was checked by measuring the volume of collected effluent from the

last stage reactor over 1-3 hours.

Results

Nitrogen Removal: Nitrification

The reactors were operated at an initial temperature of 10°C. Almost complete nitrification was

obtained within the first week of start up. Figure 5 shows the NH4-N profile with time across the MBBR

process. The temperature was gradually decreased to 4°C.

Figure 5: Temperature and NH4-N profile across MBBR system

The reactors were operated at this temperature for approx. 80 days (days 60 to 140). During this

period, the alkalinity was also reduced in order to simulate the low alkalinity of tailing effluents.

However, the nitrification deteriorated when the alkalinity was below 50 mg CaCO3/L (shown by

increased levels of NH4-N in the nitrification stage, day 60 -70). When the alkalinity in the feed was

increased again through chemical addition, nitrification recovered within approximately one week.


10

Concentrations of NH4-N in the nitrification reactor remained low for the rest of the study. The

temperature was decreased to 2°C between day 145 and 160. During this period, NH4-N was not

completely oxidized to NO3-N as previously observed. Figure 6 shows the NOx- N profile across the

nitrification reactor. It can be seen that until day 130, the NO2-N concentration was very low, indicating

complete nitrification. After day 130, an increase in NO2-N was observed, and towards the end of the

study, when the process was operated at 2°C, the concentration of NO2-N was approximately 5 mg/L.

This showed that the second step in nitrification, where NO2-N is oxidized to NO3-N, was more affected

by the low temperature then the first step (oxidation of NH4-N to NO2-N).

Throughout the study, the effluent NH4-N concentration was maintained below 0.2 mg/L NH4-N,

with the exception of day 121 due to the conversion of the denitrification step from one-stage to two-

stage. The low NO3-N concentration observed on day 135 is most likely an experimental error as no

increase in residual NH4-N or NO2-N concentration was observed on that day.

Figure 6: Temperature and NOx-N profile in MBBR1effluent (nitrification reactor)

Nitrogen Removal: Denitrification

The denitrification reactor was started on day 4 of the study at an operating temperature of 10°C.

The addition of carbon source was started on day 7. Initially, ethanol was added as carbon source in

addition to methanol until day 41. Figure 7 shows that complete denitrification was obtained within 4

weeks of start-up however, upon the removal of ethanol from the carbon source mix, an increase in

effluent NOx-N to 5 mg/l was observed.


11

The operating temperature was gradually decreased to 4°C, operation at this temperature was

maintained for approx 80 days (day 60 – 140). During this period, carbon source was added in large

excess, however a residual NOx-N concentration of 6 mg/L was observed such that on day 121, the

denitrification step was divided into 2 stages to improve the denitrification process. The residual NOx-N

concentration decreased to less than 3 mg/l within a short time. Finally, when the process was operated at

2°C, effluent NOx-N went up to approximately 5 mg/L, 0.78 mg/l of which was NO2-N. The residual

nitrite concentration exceeds the limit set by British Columbia’s Ministry of Environment, in the Water

Quality Criteria for Nitrogen for Protection of Freshwater Aquatic Life (30 day average < 0.2 mg/l NO2-

N, max value < 0.6 mg/l NO2-N for chloride concentration > 10 mg/l) (Environment Management Act,

1981).

Figure 7: Temperature and NOx-N profile across the MBBR system

Selenium Removal

The process was started at a temperature of 15°C and then decreased gradually to 7°C during the

trial. Denitrification was obtained within 2 weeks of start up. Residual N-NOx concentration was

maintained below 0.5 mg/l throughout the study. Since some species of selenium reducers will

preferentially utilize nitrate as electron acceptors when nitrate is available, it is essential to completely

reduce nitrate to achieve complete selenium reduction.

Residual available carbon is also another important parameter to follow to ensure that the reaction

rate is not substrate limited. A profile of the COD concentration across the MBBR system is shown in


12

Figure 8. Assuming that the influent COD in not readily available, it can be said that the system is

substrate limited when MBBR2 effluent is equal or less than the influent COD concentration – such as

can be observed on day 55. It should also be said that excessive COD dosing will result in sulphide

generation which is not desirable due to the potential for toxic H2S gas release.

Figure 8: Soluble COD profile across MBBR system over time

The temperature and selenium profile across the system is shown in Figure 9. Selenium removal to

less than 5 µg/L was achieved within two weeks after start-up. The effluent total Se-concentration

remained below 3 µg/L during the rest of the trial, except during the period when the reactors were fed

with synthetic wastewater due to a shortage of mine water. During a period of approximately 7 days (days

50-57), the influent Se concentration was increased to 52 µg/L due to an error in preparation of the

synthetic water, resulting in an increase effluent Se to 7 µg/L which is likely to be due to a shortage of

carbon source as well as the sudden increase in loading.


13

Figure 9: Total selenium profile across MBBR system over time

Example of Full Scale Plant

A full scale MBBR system treating a mine effluent was recently commissioned in Sweden. The

MBBR is designed for complete nitrogen removal from the effluent of an underground mine extracting

zinc, silver, copper, lead and gold. Effluent is pumped to the surface and is sent directly to the MBBR

hence the influent temperature is fairly constant at 10-15°C. Treated effluent needs to comply with a

monthly average soluble total nitrogen concentration of < 2 mg/l. The influent design parameters are

listed in Table 4.

Table 4: Mine effluent water characteristics

Parameter Value

Total N, mg/L

NO3-N, mg/L

40.6

35

NH4-N, mg/L 5.55

Flowrate , m3/hr 120

A schematic of the process is shown in Figure 10. In this case, a pre-denitrification step was used to

benefit from the alkalinity released during denitrification and thus save on chemical dosing requirements.


14

Figure 10: Process flow schematic

An overview of the plant together with the engineering layout of the MBBR is shown in Figure 11.

A circular tank configuration divided into the required nitrification and denitrification cells was used to

save on construction costs, optimizing footprint and pipework.

Figure 11: Site layout, confidential client (Sweden)

The influent and effluent nitrogen profiles are shown in Figures 12 and 13 respectively. The influent

nitrate concentrations during the monitoring period were slightly lower than the design values whilst other

parameters were in line with the expected values.


15

Figure 12: Nitrogen profile in the mine effluent

Figure 13: Nitrogen profile in the MBBR effluent

During the monitoring period shown (2014-03-22 to 2014-04-26), the influent flow to the MBBR

exceeded the design flow on several occasions however, it can be seen from Figure 13 that the monthly

average total nitrogen concentration of < 2 mg/l was achieved.


16

Conclusion

Laboratory scale pilot tests have demonstrated that the MBBR process can be successfully used to

remove nitrogen and selenium from mine effluents at low temperatures. Complete nitrification and

denitrification was maintained down to 4°C without production of nitrite. Although nitrification was

observed at 2°C, nitrite production may result in non-compliance with the BC Fresh Water Quality

regulations. In full scale applications, careful monitoring of the plant effluent would be required during

periods when effluent temperature decreases below 4°C to avoid overloading the reactors.

The process has been demonstrated at full scale on a Swedish mine, consistently achieving very low

total nitrogen concentration.

It has also been shown that the MBBR process can successfully reduce selenate and selenite to

elemental selenium. Total effluent selenium concentrations of less than 5 µg/l can be consistently

achieved down to 7°C. Careful control of carbon source dosing is required to maintain the low selenium

concentrations without excessive sulphide generation.

References

Chapman P., Adams W., Brooks M., Delos C., :uomaq S., Maher W., Ohlendorf H., Presser T. and Shaw P. ( 2009).

Ecological assessment of selenium in aquatic Environment. SETAC workshop. Feb 22-28, 2009.

Hoang V., Delatolla R., Abujamel T.,Mottawea W., Gadbois A., Laflamme E. and Stintzi A.( 2014) Nitrifying

moving bed biofilm reactor (MBB) biofilm and biomass response to long term exposure to 1°C. Water Research 49.

pp 215-224

Metcalf and Eddy (1991) Wastewater Engineering. Third Edition pp429-433

Ødegaard H., Rusten B. and Westrum T. (1994) A new moving bed bofilm reactor – Applications and results. Wat.

Sci. Tech., Vol. 29, No 10-11, pp 157-165.

Ødegaard H., Rusten B. and Siljudalen J. (1999) The development of the moving bed biofilm process – From idea to

commercial product. European Water Management, Vol. 2, No. 2.

Oremland R., Herbel M., Switzer Blum J., Langley S., Beveridge T., Ajayan P., Sutto T., Ellis A.V and Curran

S.(2004). Structural and Spectral Features of Selenium Nanospheres Produced by Se-Respiring Bacteria. Applied

and Environmental Microbiology, Jan. 2004, p. 52–60 Vol. 70, No. 1

Rusten B., Hellström B. G., Hellström F., Sehested O., Skjelfoss E. and Svendsen B. (2000) Pilot testing and

preliminary design of moving bed biofilm reactors for nitrogen removal at the FREVAR wastewater treatment plant.

Wat. Sci. Tech., Vol. 41, No. 4-5, pp 13-20

Rusten R., Hem LJ. and Ødegaard H. (1995). Nitrogen removal from dilute wastewater in cold climate using

moving bed biofilm reactors. Water Environ Resources 67 (1) pp 65-74


17

Takata T., Hirata M., Kokutu S., Toorisaka E., Ozaki M. andHano T.( 2008). Kinetic study on biological reduction

of selenium compounds. Process Biochemistry Vol 43 pp 1904-1307.

Welander U. and Mattiasson B. (2003) Denitrification at low temperatures using a suspended carrier biofilm

process. Water Research 37 pp 2394–2398

Zaitsev G., Mettamen T. and Langwaldt J. (2008) Removal of ammonium and nitrate from cold inorganic mine

water by fixed-bed biofilm reactors. Mineral Engineering 21, pp 10-15

Documents

Wastewater treatment using MBBR in cold climates