Ecohydrology & Hydrobiology 19 (2019) 54–65

Contents lists available at ScienceDirect

Ecohydrology & Hydrobiology

journal homepage: www.elsevier.com/locate/ecohyd

Original Research Article

Kinetics and simulation of nitrification at various pH values

of a polluted river in the tropics
Thi Thu Huyen Le a,[147_TD$IF] *, Joachim Fettig b, Günter Meon a
a
Abteilung Hydrologie, Wasserwirtschaft und Gewässerschutz, Leichtweiß-Institut für Wasserbau, Technische Universität
Braunschweig, Germany
b
Fachgebiet Wassertechnologie, Hochschule Ostwestfalen-Lippe, Germany

A R T I C L E I N F O A B S T R A C T

Article history: Nitrification is simulated in existing water quality models by multiplying the rate with the
Received 2 October 2017 attenuation factors for water temperature and dissolved oxygen. The effect of pH,
Accepted 28 June 2018 substrates and bacteria is mostly neglected. Moreover, the two-step nitrification is
Available online 10 July 2018
simplified to a one-step process. The aim of this study is to investigate the kinetics of two-
step nitrification and to simulate this process. Consequently, the importance of considering
Keywords:
the effect of pH, substrate and bacteria in nitrification modelling of river basins is
Nitrification rate
emphasized. For this, an experiment was performed on water taken from the Tay Ninh
Bacterial growth
Nitrification simulation River in Vietnam. The results show a low actual ammonium oxidation rate in the water
Water quality modelling column of the river. The ammonium oxidation is completely inhibited at pH 5 while the
Water management nitrite oxidation is strongly inhibited at pH 8.5. In comparison to the ammonium oxidizers
pH (AOB), the nitrite oxidizers (NOB) are rarely inhibited at low pH. Numerical simulation of
the experiment delivered ammonium oxidation rates kA and nitrite oxidation rates kN for
different pH values. The kA for pH values lower than 6 is very low. It increases rapidly as the
pH increases from 6 to 6.5. We also found out that pH values of 7.5 and 7 are optimal for the
activity of AOB and NOB, respectively. Our results emphasize the importance of considering
the effect of pH, substrate and bacteria on the nitrification research in river basins. In doing
so, nitrification can be simulated more realistically and sustainable water management can
be carried out more successfully.
ß 2018 The Authors. Published by Elsevier B.V. on behalf of European Regional Centre for
Ecohydrology of the Polish Academy of Sciences. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction rivers take advantage of the increasing amounts of

ammonium to generate energy for growth and mainte-
Ammonium pollution of inland water due to waste- nance by means of the nitrification process (Laanbroek and
water discharge is rapidly increasing in densely populated Bollmann, 2011), whereby more oxidized forms of
watersheds (Brion and Billen, 2000; Laanbroek and inorganic nitrogen, such as nitrite and nitrate, are
Bollmann, 2011). The nitrifying bacteria in lakes and produced simultaneously. Nitrification is a two-step
process of the biological oxidation carried out by two
different chemolithotrophic bacteria: the ammonium
oxidizers (AOB) and the nitrite oxidizers (NOB). These
* Corresponding author at: Beethovenstr. 51a, D-38106 Braunschweig,
bacteria are obligate autotrophic and use energy from
Germany.
E-mail addresses: [email protected], [email protected] nitrification to fix CO2. The ammonium oxidation is
(T.T. Huyen Le). comprised of two stages, in which two different enzymes

https://doi.org/10.1016/j.ecohyd.2018.06.006
1642-3593/ß 2018 The Authors. Published by Elsevier B.V. on behalf of European Regional Centre for Ecohydrology of the Polish Academy of Sciences. This
is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 55

are involved. In the first stage, ammonium is converted to the separate oxidation of ammonium and nitrite need to be
hydroxylamine (Hagopian and Riley, 1998; Prosser, 1990; taken into account. In this study, the pH effect on the
Wrage et al., 2001). The next stage in ammonium ammonium and nitrite oxidation of inland water is
oxidation is from NH2OH to NO2. Hydrogen ions are examined. The objectives of our work are to investigate
released when ammonium ions are oxidized to nitrite. the kinetics of the two-step nitrification, to simulate this
This release may lower the pH of lightly buffered waters. process and to point out the importance of considering the
The NOB oxidize nitrite produced by the oxidation of effect of pH and bacteria population in the nitrification
ammonium directly to nitrate without intermediate modelling of river basins. Our findings challenge the
products (Wrage et al., 2001). Due to the oxygen commonly accepted approaches used to simulate the
consumption by nitrification, a river system can suffer nitrification in river basin management and restoration.
from oxygen depletion, which can affect the aquatic To do this, nitrification kinetics of a tropical river, the
organisms negatively. Furthermore, increasing ammoni- Tay Ninh river, were investigated under different pH values
um pollution causes increasing concentrations of un- through observations of in situ changes in concentrations
ionized ammonia NH3. NH3 and NO2, which can exist of ammonium, nitrite and nitrate over time by means of a
under unfavourable conditions during the nitrification microcosm experiment. This method was also applied by
process, are toxic to fish at low concentrations (Hagopian Cavari (1977), Cirello et al. (1979), Cooper (1986),
and Riley, 1998). Thus, controlling ammonium inputs into McCutcheon (1987), Pauer and Auer (2000), Sebilo et al.
a river system and ensuring a functional nitrification (2006), Xia et al. (2004) and Yongming (1988). In those
process are important to guarantee a good environment studies, in which the nitrification microcosm experiments
for the aquatic organisms as well as for the human were applied to fresh water systems, the effect of pH on the
community. nitrification process was almost neglected and it is unclear
Nitrification is affected by several parameters such as whether the pH was kept constant during the experiment
pH, dissolved oxygen (DO), water temperature, substrate or not. Moreover, in many cases the water temperature
concentration and nitrifying bacteria population (Chen was kept at 20 8C, at which the nitrification does not reach
et al., 2006; Gujer, 2010; Rheinheimer et al., 1988). Among its maximum rate. Furthermore, the existing studies
these parameters, pH has a strong effect on the nitrification examined the influence of pH, water temperature and
process because it not only affects the bacterial growth DO on the nitrification process simultaneously. In doing so,
rates but also modifies the acid base equilibriums NO2/ the sole effect of pH on nitrification is overlaid by the effect
HNO2 and NH4+/NH3 and hence affects the substrate of other factors. Some authors examined the sole influence
availability for AOB and NOB (Jiménez et al., 2011). This of pH on nitrification, but their studies focused only on
effect was studied by many authors such as Grunditz and sludge, soils, wastewater and pure cultures (Bae et al.,
Dalhammar (2001), van Hulle et al. (2007), Bae et al. (2001) 2001; Groeneweg et al., 1994; Jiménez et al., 2011).
and Jiménez et al. (2011). The results of these authors show Concerning the separate observation of ammonium
the optimum pH for AOB and NOB in neutral and slightly oxidation and nitrite oxidation for fresh water using a
alkaline conditions. According to Grunditz and Dalhammar microcosm experiment, only Cavari (1977), Xia et al.
(2001), the optimal pH for AOB is 8.1 and for NOB 7.9. In the (2004) and Yongming (1988) investigated this aspect in
studies of Bae et al. (2001) and Jiménez et al. (2011), the their research. In this study, the ammonium oxidation and
optimal pH for AOB as well as NOB is about 8.6. Grunditz nitrite oxidation at various pH values was performed under
and Dalhammar (2001) and Bae et al. (2001) also found out optimum DO and at optimum water temperature.
that a temperature between 30 8C and 40 8C is optimal for
the growth of nitrifying bacteria. The results are in 2. Materials and methods
agreement with the results of Hellinga et al. (1999) and
Groeneweg et al. (1994). Although many researchers 2.1. Study area and sampling site
reported that lower DO inhibits the growth of AOB and
NOB, the critical values of DO recorded in the literature The study site for this work is the Tay Ninh River located
were different (Peng and Zhu, 2006). According to Bansal in the northwest of Ho Chi Minh City, Vietnam (Fig. 1). The
(1976), the nitrifying bacteria cannot grow if the oxygen Tay Ninh River has been highly polluted for many years.
content falls below 0.5 mg/L in stream waters. Wheaton Nitrogen contamination has become a serious problem,
et al. (1994) suggested to maintain 2 mg/L as the minimum which is mainly due to high concentrations of ammonium
oxygen level in aquaculture nitrification biofilters. (Le et al., 2017). The ammonium concentration has
Nitrification modelling as a part of water quality extreme temporal fluctuations resulting from the inter-
modelling is increasingly applied for sustainable water mittent discharge of untreated industrial wastewater.
quality management at the catchment scale. Most of river Within the catchment, there are 14 industrial companies
basin water quality models simulate nitrification only in producing tapioca starch and 3 rubber companies (Fig. 1).
dependence of DO and water temperature (Brunner, 2008; On average, a tapioca company in the catchment produces
Chapra, 1997; Chapra et al., 2012; Cole and Wells, 2011; 100 t of tapioca starch per day and releases about 1500 m3
DHI, 2003; Müller, 2001; Neitsch et al., 2011). The effect of of wastewater highly loaded with organic compounds
pH in the modelling of freshwater systems is mostly (Fettig et al., 2013; Fettig and Pick, 2013; Le et al., 2017).
neglected. Moreover, these models consider nitrification as The concentrations of BOD5, NH4-N, and PO4-P in tapioca
a one-step process. To simulate the nitrification process in wastewater have an average values of 6700 mg/L, 30 mg
detail and follow its natural behaviour, the pH effect and N/L, and 58 mg P/L, respectively (Fettig and Pick 2013).
[(Fig._1)TD$IG]
56 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65

Fig. 1. Study area and location of the investigated station (CG).

This causes a huge amount of highly contaminated organic pH, high concentrations of ammonium and organic
wastewater in the catchment. This wastewater is also substances (Le, 2014). The soils in the catchment are
characterized by low pH (around 4.5). The raw wastewater dominated by ferralic acrisols (84%). Ferralic acrisols are
is discharged into a pond system without any technical strongly weathered acidic soils with low base saturation.
measures to prevent infiltration into the groundwater and The pH values of the river range between 4.4 and 6.9. The
is further discharged irregularly into the Tay Ninh River mean pH of the river is 6.0, at which the nitrification does
when the storage capacity of the ponds is exceeded. The not reach its maximum rate. The mean DO concentrations
uncontrolled release of untreated wastewater, mainly range from 3.4 to 6.1 mg/L, whereby concentrations under
from the tapioca production, causes a severe deterioration 2.0 mg/L were found frequently. Monitored ammonium
of water quality and nitrogen pollution in the Tay Ninh concentrations range from 0.1 to 21.0 mg N/L and the
River system (Le et al., 2017). average is 1.7 mg N/L. Compared to the high ammonium
Fig. 2 presents the spatial variation of water quality concentration, the nitrite and nitrate concentrations are
parameters affecting nitrification (water temperature, pH, rather low with an average of 0.04 mg N/L and 0.36 mg N/
DO) and nutrient species measured in 2009, 2010 and L, respectively. This indicates an inhibition of the
2013 at 10 stations in the catchment. The locations of the nitrification due to low pH and low DO.
stations are shown in Fig. 1. The average water tempera- In this study, the sampling site for the investigation of
ture in the catchment is about 29 8C and has only slight the nitrification is located at the station Cau Gio (CG) in the
spatial and temporal fluctuation which results from the middle of the river (Fig. 1). A water column sample was
catchment’s location in the tropics. The river suffers from collected as a surface grab sample in the central point of
low pH and low buffer capacity due to the geological this station on July 22nd, 2013. This station has a mean
conditions and untreated wastewater discharge with low discharge of 20 m3[148_TD$IF]/s in the flood season (from September
[(Fig._2)TD$IG] T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 57

Fig. 2. Spatial variation of water quality parameters affecting the nitrification and of nutrient species measured in 2009, 2010 and 2013 along the Tay Ninh
River and its catchment.

Table 1
Analytical methods applied for each parameter.

Parameter Description Unit Method

NH4+-N Ammonium mgN/L Indophenol blue

NO2-N Nitrite mgN/L Griess reaction
NO3-N Nitrate mgN/L 2,6-dimethylphenol
TN Total nitrogen mgN/L Peroxodisulfate oxidation, 2,6-dimethylphenol
COD Chemical oxygen demand mg/L Chromosulfuric acid oxidation, chromate determination

to November), of 5 m3/s in the dry season (from December Table 1. Water temperature, pH and DO were measured
to August) and an average depth of 1.6 m. Before this using a WTW MULTI 340i instrument with a SENTIX 41 pH-
experiment took place, a preliminary nitrification experi- electrode and a CELLOX-3 oxygen sensor. The original
ment with a water sample taken from the same site was ammonium, nitrite and nitrate concentrations of the
performed on January 26th, 2013. This preliminary sampled water are 1.0, 0.06 and 0.94 mg N/L, respectively.
experiment was recorded carefully in a protocol for a In order to unambiguously determine the changes of
systematic investigation of the nitrification rate of the river ammonium, nitrite and nitrate concentrations during the
later on. experiment, the initial ammonium concentration was
increased to 8.1 mg N/L by adding 420 mL from a stock
2.2. Sample collection and experimental approach solution of 1.0 g/L NH4Cl to the river water sample. Such an
ammonium concentration is realistic for the Tay Ninh River
Rates of ammonium and nitrite oxidation were and stays in the range of the monitored data. In the
measured by observing changes in concentrations of nitrification experiments of Pauer and Auer (2000), Cirello
ammonium, nitrite and nitrate over time during the et al. (1979) and Hsiao et al. (2014), ammonium chloride
experiment at different pH values. In order to do this, was also added to the water samples to achieve the desired
15 L of river water was collected from the CG station in the ammonium concentration. Additionally, because of the
middle of the river as mentioned in Section 2.1. The water low buffer capacity of the river water as mentioned in
sample was kept at 4 8C, transported to the laboratory and Section 2.1, NaHCO3 was added to the river water to raise
analysed immediately. NH4+-N, NO3-N, NO2-N, total its buffer capacity. Thus, the alkalinity of the sample was
nitrogen (TN) and COD (chemical oxygen demand) were increased from 0.59 to 2.65 mmol/L. This higher buffer
quantified spectrometrically (NOVA 60, Merck) during the capacity helps to prevent significant changes of the pH
experiment. Analytical techniques are summarized in value during the experiment. The water sample was
58 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65

aerated to increase the dissolved oxygen (DO) to about ammonium concentration was neglected. Moreover, dur-
7 mg/L before starting the experiment. ing the first 46 h, there was no visible increase in the
The whole water sample was then divided into 9 glass ammonium concentration in beaker 1 to beaker 7. Beaker
beakers. The beakers were sterilized before the experi- 8 and beaker 9 were not considered in this evaluation due
ment. Each glass beaker received about 1.5 L of prepared to ammonia volatilisation into the air. Denitrification as a
river water. To prevent nutrient assimilation by algae and sink for nitrogen can be excluded in this study, since the
inhibition of nitrification by light (Rheinheimer, 1991; oxygen contents in the beakers were always kept above
Schön and Engel, 1962), the 9 glass beakers were incubated 3 mg/L. It is well known that denitrification, as an
in the dark. The water was then adjusted to different pH anaerobic respiratory process, does not occur in the
values: pH 5.0, pH 5.3, pH 5.6, pH 6.0, pH 6.5, pH 7.0, pH 7.5, presence of oxygen at concentrations higher than about
pH 8.0, pH 8.5 with either NaOH (0.1n) or H2SO4 (0.1n). 0.2 mg/L (Sebilo et al., 2006).
During the period in which the experiment was performed,
water temperature and DO were controlled regularly. The 3. Results and discussion
pH values of the glass beakers were monitored continu-
ously and maintained at the defined pH by adding NaOH 3.1. Effect of pH on the nitrification process in the beakers
(0.1n) or H2SO4 (0.1n) as needed during the experiment.
According to Admiraal and Botermans (1989), oxygen Temporal changes in NH4+-N, NO2-N and NO3-N at
concentrations above approx. 3 mg/L are generally different pH values over the experimental period of 190 h
thought to saturate microbial nitrification. Stenstrom are presented in Fig. 3. The initial ammonium, nitrite and
and Poduska (1980) gave 0.3 mg/L as the lowest DO nitrate concentration of all examined beakers were 8.1,
concentration at which nitrification can occur. Bansal 0.06 and 0.94 mg N/L, respectively. Measuring the total
(1976) reported that nitrifying bacteria cannot grow if DO nitrogen (TN) concentrations helps to quantify the
in stream water falls below 0.5 mg/L. Therefore, DO was ammonia loss due to ammonia volatilisation during the
kept above 3 mg/L during the experiment by aeration. experiment. In beaker with pH 8.5, ammonium and TN
According to Wetzel (2001), the nitrifying bacteria are concentrations decreased significantly due to ammonia
mesophilic with a wide temperature tolerance ranging volatilisation into the air. Ammonium in fresh water acts as
from 1 8C to 37 8C. Grunditz and Dalhammar (2001) and a weak base in a dynamic equilibrium between ammonia
Bae et al. (2001) proposed that nitrifying bacteria grow (NH3) and ammonium (NH4+). The acid base equilibrium
optimally in the temperature range between 30 8C and NH4+/NH3 is dependent on pH and water temperature
40 8C. During the experiment, the temperature was kept (Hagopian and Riley, 1998; Pambrun et al., 2006). At pH
between 30 8C and 32 8C to ensure the optimal tempera- 8.5 and a water temperature of 30 8C, about 20% of the total
ture for nitrification. Since throughout the experiment, DO ammonium concentration is present in NH3 form.
was always above 3 mg/L and the water temperature was In the beakers with pH 8.0 and 8.5, the first apparent
between 30 8C and 32 8C, the effect of these two increase of nitrite concentration was observed after 46 h of
parameters on nitrification were negligible. Therefore, incubation. This indicates that the ammonium oxidation
only the effects of pH, the nitrifying bacteria population started after 46 h. In contrast to the rapid increase of nitrite
and the substrate concentration had an influence here. concentration after the start of the ammonium oxidation,
The glass beakers were incubated at optimal water the nitrate concentration at pH 8.5 did not change
temperature, optimal DO level and given pH values for significantly. It can be concluded that the NOB, in contrast
about 8 days (190 h) in the dark. Aliquots were removed to the AOB, are strongly inhibited at pH 8.5. At pH 8.0, the
from the microcosms and analysed for ammonium, nitrite, nitrite oxidation is less inhibited. At the end of the
nitrate, total nitrogen (TN) and COD. Based on the experiment, the nitrate concentration at pH 8.0 had
experience of the preliminary experiment (Section 2.1), increased from 0.64 mg N/L to 5.1 mg N/L.
NH4+-N, NO3-N, NO2-N, total nitrogen (TN) and COD of No change in the concentrations of ammonium, nitrite
each beaker were measured in the first 4 days only 1 time and nitrate at pH 5.0 over the investigated period of 190 h
per day at 8 AM, as during these 4 days a low nitrification is noticeable (Fig. 3). This indicates a strong inhibition of
rate was expected. From the fifth day on, the measure- the ammonium oxidation at pH 5.0. From the results of the
ments took place twice per day, around 8 AM and 7 PM. experiment, it could not be determined whether the nitrite
During the experiment, COD was measured to estimate oxidation at pH 5.0 is inhibited due to the sole effect of pH.
the effect of organic substances. BOD (biochemical oxygen The reason for this is the substrate inadequacy for the
demand) was not measured due to lack of equipment, so activity of NOB due to the inhibition of AOB at pH 5.0. An
that the concentration of biodegradable organic substance inhibition of NOB due to insufficient substrate availability
could not be quantified. The COD concentration of the could be possible. At pH 5.3–7.5, the transformation of
original water sample collected at the station CG at the ammonium to nitrite and finally to nitrate did not take
beginning of the experiment was 11 mg/L. Theoretically, place in the first 23 h after starting the experiment. After
degradation of organic matter can be a source of 46 h, only a slight decrease of ammonium concentrations
ammonium as a natural result of ammonification. Howev- at pH 5.6 to pH 7.5 was detected. At pH 5.3, the ammonium
er, since nitrogen is only about 4–10% of the dry weight of oxidation started after 70 h and remained at a very low
organic substrate (Boyd, 2000) and the COD concentrations level. After 70 h, the ammonium concentrations at pH
of all beakers did not change significantly over the 5.6 to pH 7.5 decreased rapidly and the nitrite concentra-
experiment, the influence of organic substrate on the tions started to increase significantly. At pH 6.5 to pH 8.5,
[(Fig._3)TD$IG] T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 59

Fig. 3. Changes in the concentrations of NH4+-N, NO2-N and NO3-N at different constant pH values during the experiment.

the maxima of nitrite concentrations were reached when lag period for ammonium oxidation was longer in the
the ammonium concentrations dropped nearly to zero research of Cavari (1977) with 7 and 29 days for water
(Fig. 3). The nitrate concentrations at pH 5.3 to 6.5 and at taken during the turnover period and the stratification
pH 7.0 to 8.5 started to increase slowly after 95 h and 46 h, period of Lake Kinneret in Israel, respectively. Xia et al.
respectively. A rapid increase of nitrate was noticed after (2004) found different lag periods in their experiments
nitrite concentrations reached their maxima. investigating the effects of suspended solids on nitrifica-
tion in river water of the Yellow River in China. The lag
3.2. Estimation of the lag period periods for ammonium oxidation were 0 day and 5 days,
respectively, for 2 different water samples without
The growth of bacteria is divided into 3 phases: initial suspended solids.
phase, transition phase and exponential phase, whereby Regarding the second step of the nitrification process,
the initial phase and transition phase are considered Cavari (1977) proposed that the lag periods of the nitrite
together as the lag period (Prats, 2008). The lag period for oxidation of water samples taken during the mixed period
nitrification has been observed in lake and river studies by and the stratification period were 12 days and 0 day,
some authors (Cavari, 1977; Pauer and Auer, 2000; Xia respectively. Xia et al. (2004) reported different lag periods
et al., 2004; Yongming, 1988). However, their studies of the nitrite oxidation in water samples. In one water
focused on water bodies located in the temperate zone sample, the nitrite oxidation took place immediately and in
(Pauer and Auer, 2000) and in the subtropics (Cavari, 1977; the other water sample, the lag period was 6 days.
Xia et al., 2004; Yongming, 1988). Furthermore, these Based on the results of the experiment in this study, the
authors did not consider the pH effect on ammonium lag period of the ammonium oxidation for pH 5.3–8.5 can
oxidation as well as on nitrite oxidation. In the study of be 70 h (2.9 days), whereby the initial phase is 46 h
Yongming (1988), the nitrification of water from the (1.9 days). In the Tay Ninh River, the ammonium oxidation
Toujiang River (China) was examined for 1 week under is completely inhibited at pH 5.0. Concerning the lag period
aeration, whereby the water sample in the experiment was of the nitrite oxidation, it is divided into two lag periods.
covered at the bottom with sediments. The results showed The lag period for pH 5.3–6.5 can be 142 h (5.9 days) and
a rapid ammonium and nitrite oxidation in the first day of for pH 7.0–8.5 it can be 128 h (5.3 days). The initial phases
the experiment (nearly no lag period). are 95 h (4 days) for pH 5.3–6.5 and 46 h (1.9 days) for pH
Regarding the first step of the nitrification process, 7.0–8.5. When assessing the lag period of the nitrite
Pauer and Auer (2000) studied the ammonium oxidation in oxidation, it is important to consider that the activity of
the water column of the Onondaga lake and the adjoining NOB at the beginning of the experiment, in contrast to the
Seneca River, located in New York, using microcosm AOB, was inhibited by a substrate inadequacy. The NOB
experiments. No change of ammonium was observed in received the maximum substrate availability only after the
the water column experiments over the first 3–4 days. The available ammonium was almost completely oxidized. In
60 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65

this experiment, the ammonium oxidation at the begin- the ammonium oxidation; øpH_A: attenuation factor for
ning was dependent on pH and concentrations of AOB. pH on the ammonium oxidation; øN_A: attenuation factor
Dissolved oxygen, water temperature and substrate for substrate ammonium on the ammonium oxidation;
concentration (ammonium) were kept at their optima øB_A: attenuation factor for bacteria AOB on the ammo-
for AOB. In contrast to AOB, the activity of NOB depended nium oxidation; kNmax: maximum nitrite oxidation rate
on the substrate concentration (nitrite), pH and temporal (d1) under optimum water temperature (T), optimum
concentrations of NOB. DO, optimum pH, excess substrate (N) and excess
bacteria (B); øT_N: attenuation factor for water tempera-
ture on the nitrite oxidation; øDO_N: attenuation factor for
3.3. Estimation of nitrification in the river
dissolved oxygen on the nitrite oxidation; øpH_N:
attenuation factor for pH on the nitrite oxidation;
Pauer and Auer (2000) and Cavari (1977) concluded
øN_N: attenuation factor for substrate nitrite on the
that the very low nitrification rate at the beginning of
nitrite oxidation; øB_N: attenuation factor for bacteria
the experiment in the initial phase was related to the
NOB on the nitrite oxidation.
low initial nitrifier density. This initial density depends
The attenuation factors represent the influence of
on the number of nitrifying bacteria present in the river
the mentioned environmental factors on the maximum
water at the time of sampling. Nitrifying bacteria
oxidation rate and range between 0 and 1.0, whereby
concentration will change during the experiment. In
0 means a complete inhibition regarding the effects
this study, the ammonium oxidation rate and the
and 1.0 means no inhibition. In this study, due to
nitrite oxidation rate for all pH values at the beginning
optimal values for T and DO in the experiment,
of the experiment appeared to be zero. The low
Eqs. (1)–(3) can be rewritten when applied on each
ammonium oxidation rate is probably due to the low
pH as follows:
initial AOB density. The low nitrite oxidation rate is
perhaps due to the low initial NOB population or due to
the substrate inadequacy at the beginning of the dNH4
¼  kAmax ;N A ;B A NH4 ¼ kA NH4 (4)
experiment. These rates represent the actual ammoni- dt
um oxidation rate and nitrite oxidation rate of the Tay
Ninh River. This conclusion is in agreement with the
dNO2
conclusion derived by Pauer and Auer (2000) and ¼ kAmax ;N A ;B A NH4  kNmax ;N N ;B N NO2
dt
Cavari (1977). The difference of this study compared to
¼ kA NH4  kN NO2 (5)
the studies of Pauer and Auer (2000) and Cavari (1977)
is only that the lag period of the ammonium oxidation
is about 3 days for pH 5.6 to 8.5 and thus shorter. In the dNO3
¼ kN NO2 (6)
Tay Ninh River, the ammonium oxidation is completely dt
inhibited at pH 5.0.
where kA ðkA ¼ kAmax ;N A ;B A Þ is the ammonium oxidation
rate, determined for each pH, taking into account the effect
3.4. Simulation of the nitrification at different pH values of substrate ammonium and bacteria. kN ðkN ¼
kNmax ;N N ;B N Þ is the nitrite oxidation rate, determined
Since the nitrification is affected by water temperature, for each pH, taking into account the effect of substrate
dissolved oxygen, pH, substrate and bacteria population, nitrite and bacteria.
the balance for ammonium, nitrite and nitrate can be The oxidation of ammonium and nitrite is related with
simulated using the functions below (Chapra, 1997; Senzia the growth of nitrifying bacteria. The lag period (lag time
et al., 2002; van Hulle et al., 2007; Volcke, 2006; Wett and tL) in the bacterial growth is followed by the logarithmic or
Rauch, 2003): the exponential phase (Baty and Delignette-Muller, 2004;
Zwietering et al., 1990). Eqs. (4)–(6) do not consider the lag
dNH4
¼  kAmax ;T A ;DO A ; pH A ;N A ;B A NH4 (1) time and therefore can describe only the growth of
dt
nitrifying bacteria in the exponential phase. Some authors
dNO2 have dealt with the simulation of the bacterial growth and
¼ kAmax ;T A ;DO A ; pH A ;N A ;B A NH4 have proposed several models to simulate the bacterial
dt
growth (Annuar et al., 2008; Augustin and Carlier, 2000;
 kNmax ;T N ;DO N ; pH N ;N N ;B N NO2 (2)
Buchanan et al., 1997; Mochizuki and Hattori, 1987;
Ocampo-López et al., 2015; Pamment et al., 1978; Rosso
dNO3
¼ kNmax ;T N ;DO N ; pH N ;N N ;B N NO2 (3) et al., 1996; Zwietering et al., 1990). Although many efforts
dt have been made, these models are either too complicated
where NH4, NO2 and NO3: nitrogen concentration (mg N/ (too many parameters need to be estimated) or do not
L) of ammonium, nitrite and nitrate; kAmax: maximum consider the lag phase of the bacterial growth. Further-
ammonium oxidation rate (d1) under optimum water more, most of the empirical models contain mathematical
temperature (T), optimum DO, optimum pH, excess parameters rather than parameters with biological mean-
substrate (N) and excess bacteria (B); øT_A: attenuation ing (k, tL) (Zwietering et al., 1990). Deterministic models
factor for water temperature on the ammonium oxida- with parameters having biological meanings are easier to
tion; øDO_A: attenuation factor for dissolved oxygen on understand and to calibrate.
 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 61

In order to simulate the changing of ammonium, nitrite For the optimization of nitrite and nitrate, Eq. (11)
and nitrate for the entire nitrification experiment in this becomes:
study, taking into account of the lag period tL, Eqs. (4)–(6)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uX
are modified as follows: u  
n 2
u fNO2 ðt i ; pH j ; u  NO2i j Þ
m t
X
dNH4   Y ðuÞ ¼ i¼1
¼ kA NH4 1  af;exp½ðt=tL A Þw  (7) n
dt j¼1
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uX n   2
u
u fNO 3 ðt i ; pH j ; u  NO3i j Þ
dNO2   m t
X
¼ kA NH4 1  af;exp½ðt=tL A Þw  þ i¼1
(13)
dt n
  j¼1
 kN NO2 1  af;exp½ðt=tL B Þw  (8)
The calibration (optimization) of nitrite and nitrate was
  carried out after finishing the calibration (optimization) of
dNO3
¼ kN NO2 1  af;exp½ðt=tL B Þw  (9) ammonium. For the optimization, the method Fmin-
dt
searchbnd of MATLAB was chosen. Fminsearchbnd is a
where t is the simulation time, tL_A and tL_B is the lag time derivative-free method. It is used to optimize nonlinear
for AOB and NOB and w is a shape parameter (dimension- functions with inequality constraints. The MATLAB meth-
less). od Fminsearchbnd uses the simplex method of Nelder-
For the simulation of the entire nitrification experi- Mead to solve the optimization problem. The idea of the
ment in this study, the differential Eqs. (7)–(9) were used Nelder-Mead method is to create a sequence of simplexes
and implemented in MATLAB. The MATLAB procedures for that will be ideally smaller and smaller in diameter to
solving differential equations are based on the Runge- reach the desired minimum.
Kutta algorithm. To solve Eqs. (7)–(9), function ode45 was The oxidation rates kA and kN in Eqs. (7)–(9) represent
applied. The MATLAB procedure ode45 is a combination of the specific rates of the ammonium oxidation and the
a 4th[150_TD$IF]49 order Runge-Kutta method with a 5th[15_TD$IF] order method. nitrite oxidation when AOB and NOB grow exponentially.
The calibration (parameter estimation) was carried out The simulated concentrations of ammonium, nitrite and
with an optimization procedure and is done by the nitrate using the proposed Eqs. (7)–(9) are shown in
method of least root of mean square error (RMSE). The root Fig. 4. As the unit of the oxidation rate is d1, the time in the
of mean square error, obtained for each pH, is calculated as x-axis was changed from hours to days. For all pH values, a
follows: good agreement between simulated and measured data
was achieved. This is confirmed by high coefficients of
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uX n determination (r2) and high Nash-Sutcliffe coefficients of
u   2
u
t
f ðt i ; u  g i Þ efficiency (e). The coefficients r2 at pH 5 are 0.00 due to the
i¼1 measured values of ammonium, nitrite and nitrate being
LðuÞ ¼ (10)
n constant over time. Also very low concentrations of nitrite
(<0.5 mg N/L) at pH 5, pH 5.3 and pH 5.6 cause low
where u = (u1, . . ., up): parameters need to be calibrated
coefficients of r2 and e. In contrast to r2, which evaluates
(optimized) (kA, kN, tL_A, tL_B, w); f(ti, u): simulated values of
the linear relationships between simulated and measured
the model at time ti and with the parameter u; n: number of
values, Nash-Sutcliffe coefficient e takes into account the
time steps at which ammonium, nitrite and nitrate were
differences in the measured and simulated values (Krause
measured (n = 13); g: measured values.
et al., 2005; Legates and McCabe, 1999). It is thus more
Since L(u) was calculated for each pH, the error total (Y)
suitable in the model evaluation.
of all L(u) was used for optimization (Eq. (11)). The error
The model parameters in Eqs. (7)–(9) which were
total measures the agreement between the simulated and
optimized during the simulation of the nitrification
the measured values. To determine the Y, all measured
experiment are summarized in Table 2. The ammonium
data and all pH values were taken into account
oxidation rate kA and nitrite oxidation rate kN were
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi estimated for each pH. These values allow a first estimation
uX
u  
n 2
u f ðt i ; pH j ; u  g i j Þ of the influence of the pH value on the oxidation activity of
m t
X i¼1 AOB and NOB. For pH 8.5, the sole ammonium oxidation
Y ðuÞ ¼ (11)
j¼1
n rate could not be correctly estimated due to ammonium
loss as a result of ammonia volatilisation. The kA at pH
where m is the number of the pH values (m = 9). 8.5 represents therefore the sum of ammonium oxidation
For the optimization of ammonium, Eq. (11) becomes: and ammonia volatilisation. Table 2 shows that the highest
activity of the AOB and NOB was found at pH 7.5 and pH
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uX 7.0 for the Tay Ninh River, respectively. For pH values
u  
n 2
u fNH4 ðt i ; pH j ; u  NH4i j Þ lower than 6.0, the kA value is very low. It increased
t
X i¼1
m
Y ðuÞ ¼ (12) strongly from 0.16 d1 to 1.17 d1 as the pH value
j¼1
n increased from 6.0 to 6.5 and stayed increasing until pH
7.5. This indicates that AOB of the Tay Ninh River are
sensitive to pH lower than 6.0. In comparison to AOB, the
[(Fig._4)TD$IG]
62 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65

Fig. 4. Measured and simulated ammonium, nitrite and nitrate concentrations.

oxidation of NOB is rarely inhibited at low pH values. The two biological parameters (oxidation rate and lag time) are
kN at pH 5.3 is almost high as the kN at pH 6.5 and only a bit considered. For the simulation of ammonium oxidation
lower than the kN at pH 7.0. At pH 8.5, NOB, in contrast to and nitrite oxidation in detail and following the biological
AOB, are almost completely inhibited. behaviour, Eqs. (1)–(3) should be applied. In these
The simulation results fit well to the interpretation in equations, the effect of water temperature, DO, pH,
Section 3.1 and indicate that Eqs. (7)–(9) can be used to substrate concentration and nitrifying bacteria population
simulate the nitrification in the experiment, whereby only are taken into account. In most of the existing inland water
 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 63

Table 2
Results of the parameter optimization of the simulation of ammonium, nitrite and nitrate.

Parameter Description Value Parameter Description Value

k A5 Rate AOB at pH 5 0.00 kN5 Rate NOB at pH 5 0.00

kA5.3 Rate AOB at pH 5.3 0.02 kN5.3 Rate NOB at pH 5.3 2.05
kA5.6 Rate AOB at pH 5.6 0.06 kN5.6 Rate NOB at pH 5.6 2.07
k A6 Rate AOB at pH 6 0.16 kN6 Rate NOB at pH 6 2.08
kA6.5 Rate AOB at pH 6.5 1.17 kN6.5 Rate NOB at pH 6.5 2.10
k A7 Rate AOB at pH 7 2.03 kN7 Rate NOB at pH 7 2.33
kA7.5 Rate AOB at pH 7.5 2.79 kN 7.5 Rate NOB at pH 7.5 2.17
k A8 Rate AOB at pH 8 2.56 kN 8 Rate NOB at pH 8 1.10
kA8.5 Rate AOB at pH 8.5 2.60 kN 8.5 Rate NOB at pH 8.5 0.03
tL_A Lag time of AOB 4.00 tL_B Lag time of NOB 7.00
w Shape parameter 10.00 w Shape parameter 10.00

quality models, the effects of water temperature and DO separate oxidation of ammonium to nitrite und further to
are simulated using the popular modified Arrhenius nitrate is omitted. The dependency of nitrification on the
(commonly called the theta approach) and the Michae- substrate availability, the amount of bacteria, as well as on
lis-Menten approach (Bicknell et al., 2005; Chapra et al., the pH-value is also not taken into account. The results of
2012; Cole and Wells, 2017; Deltares, 2011; DHI, 2003; our study emphasize the importance of considering the pH
Hipsey et al., 2006; Martin and Wool, 2002; Neitsch et al., effect on nitrification modelling of river basins. Moreover,
2011). While the dependency of nitrification on water the nitrifying bacteria population should be simultaneous-
temperature and DO has received intensive attention by ly simulated. In doing so, nitrification can be simulated
the watershed modelling community, the effect of realistically and sustainable water management can be
substrate, pH and bacteria on this process has been mostly carried out successfully.
neglected. Different oxidation rates for AOB and NOB at
different pH values determined in this study emphasize [152_TD$IF]Acknowledgements
the importance of taking into account the effect of pH in
the nitrification modelling of river basins. This study is part of the German-Vietnamese research
project TAPIOKA. The authors would like to thank the
4. Conclusions Bundesministerium für Bildung und Forschung (BMBF,
Grants No.: 02WA0991), Germany, and the Ministry of
A laboratory microcosm experiment was carried out to Science of Technology (MOST), Vietnam, for funding the
investigate the nitrification process in the water column of project.
the Tay Ninh River at various pH values under optimum DO
References
and at optimum water temperature. The results show that
the actual ammonium oxidation rate of the Tay Ninh River Admiraal, W., Botermans, Y., 1989. Comparison of nitrification rates in
are very low (appear to be zero) and that the ammonium three branches of the lower river Rhine. Biogeochemistry 8 (2), 135–
oxidation is completely inhibited at pH 5.0 while the nitrite 151, http://dx.doi.org/10.1007/BF00001317.
Annuar, M.S.M., Tan, I.K.P., Ibrahim, S., Ramachandran, K.B., 2008. A
oxidation, in contrast to the ammonium oxidation, is
kinetic model for growth and biosynthesis of medium-chain-length
strongly inhibited at pH 8.5. The actual nitrite oxidation poly-(3-hydroxyalkanoates) in Pseudomonas putida. Braz. J. Chem.
rate in the water column of the river could not be examined Eng. 25 (2), 217–228, http://dx.doi.org/10.1590/S0104-
in the experiment due to the substrate deficiency. We also 66322008000200001.
Augustin, J.-C., Carlier, V., 2000. Mathematical modelling of the growth
found out that the NOB in the Tay Ninh River is insensitive rate and lag time for Listeria monocytogenes. Int. J. Food Microbiol. 56
to low pH and that pH values of 7.5 and 7.0 are optimal for (1), 29–51, http://dx.doi.org/10.1016/S0168-1605(00)00223-3.
the activity of AOB and NOB, respectively. Using the Bae, W., Baek, S., Chung, J., Lee, Y., 2001. Optimal operational factors for
nitrite accumulation in batch reactors. Biodegradation 12 (5), 359–
temporal change in concentrations of ammonium and 366, http://dx.doi.org/10.1023/A:1014308229656.
nitrate, nitrification rates at different pH values of Bansal, M.K., 1976. Nitrification in natural streams. J. Water Pollut.
ammonium oxidation and nitrite oxidation were deter- Control Fed. 48 (10), 2380–2393.
Baty, F., Delignette-Muller, M.-L., 2004. Estimating the bacterial lag time:
mined. They are determining constants in the modelling of which model, which precision? Int. J. Food Microbiol. 91 (3), 261–
the nitrification process as part of water quality modelling 277, http://dx.doi.org/10.1016/j.ijfoodmicro.2003.07.002.
for a sustainable water management. Bicknell, B.R., Imhoff, J.C., Kittle Jr., J.L., Jobes, T.H., Donigian Jr., A.S., 2005.
Hydrological Simulation Program – FORTRAN: HSPF Version
Nowadays, existing water quality models, which
12.2 User’s Manual. AQUA TERRA Consultants, Mountain View, Cali-
include the nitrification process, are increasingly applied fornia.
to simulate impacts of different treatment facilities on Boyd, C.E., 2000. Water Quality: An Introduction. Springer.
Brion, N., Billen, G., 2000. Wastewater as a source of nitrifying bacteria in
point sources and diffuse sources and to manage the water
river systems: the case of the River Seine downstream from Paris.
quality sustainably (Le et al., 2012). In such models, Water Res. 34 (12), 3213–3221, http://dx.doi.org/10.1016/S0043-
biochemical processes such as nitrification are mathemat- 1354(00)00075-0.
ically represented and simulated. Many of them simplify Brunner, G., 2008. HEC-RAS River Analysis System User’s Manual: Version
4.0. US Army Corps of Engineers 747 pp. (accessed 05.04.14).
the two-step nitrification process as a one-step process, Buchanan, R.L., Whiting, R.C., Damert, W.C., 1997. When is simple good
whereby ammonium is directly converted to nitrate. The enough: a comparison of the Gompertz, Baranyi, and three-phase
64 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65

linear models for fitting bacterial growth curves. Food Microbiol. 4 ein hydrodynamisches Gewässergütemodell. (Dissertation)Tech-
(14), 313–326, http://dx.doi.org/10.1006/fmic.1997.0125. nische Universität Braunschweig, Braunschweig.
Cavari, B.Z., 1977. Nitrification potential and factors governing the rate of Le, T.T.H., Lorenz, M., Prilop, K., Lam, Q.D., Meon, G., 2012. An ecohydro-
nitrification in Lake Kinneret. Oikos 28 (2/3), 285–290. logical, ecohydraulic model system for water management of the
Chapra, S., Pelletier, G., Tao, H., 2012. QUAL2K: a modeling framework for saigon river system under tide effect. Austria 9th International Sym-
simulating river and stream water quality, version 2.12: documenta- posium on Ecohydraulics Vienna.
tion and users manual. Civil and Environmental Engineering Dept, Le, T.T.H., Zeunert, S., Lorenz, M., Meon, G., 2017. Multivariate statistical
Tufts University, Medford, MA. assessment of a polluted river under nitrification inhibition in the
Chapra, S.C., 1997. Surface Water-quality Modeling. McGraw-Hill, New tropics. Environ. Sci. Pollut. Res. 24 (15), 13845–13862, http://
York. dx.doi.org/10.1007/s11356-017-8989-2.
Chen, S., Ling, J., Blancheton, J.P., 2006. Nitrification kinetics of biofilm as Legates, D.R., McCabe, G.J., 1999. Evaluating the use of ‘‘goodness-of-fit’’
affected by water quality factors. Aquacult. Eng. 34 (3), 179–197, measures in hydrologic and hydroclimatic model validation. Water
http://dx.doi.org/10.1016/j.aquaeng.2005.09.004. Resour. Res. 35 (1), 233–241, http://dx.doi.org/10.1029/
Cirello, J., Rapaport, R.A., Strom, P.F., Matulewich, V.A., Morris, M.L., Goetz, 1998WR900018.
S., Finstein, M.S., 1979. The question of nitrification in the Passaic Martin, J.L., Wool, T., 2002. A Dynamic One-Dimensional Model of Hy-
River, New Jersey: analysis of historical data and experimental in- drodynamics and Water Quality EPD-RIV1, Version 1.0, User’s Manu-
vestigation. Water Res. 13 (6), 525–537, http://dx.doi.org/10.1016/ al. Georgia Environmental Protection Division Atlanta, Georgia
0043-1354(79)90048-4. 194 pp.
Cole, T., Wells, S., 2011. CE-QUAL-W2: A Two-Dimensional, Laterally McCutcheon, S., 1987. Laboratory and instream nitrification rates for
Averaged, Hydrodynamic and Water Quality Model, Version selected streams. J. Environ. Eng. 113 (3), 628–646.
3.71 User’s Manual. Department of Civil and Environmental Engi- Mochizuki, M., Hattori, T., 1987. Kinetic study of growth throughout the
neering, Portland State University. lag phase and the exponential phase of Escherichia coli. FEMS Micro-
Cole, T.M., Wells, S.A., 2017. CE-QUAL-W2: a two-dimensional, laterally biol. Lett. 45 (5), 291–296, http://dx.doi.org/10.1111/j.1574-
averaged, hydrodynamic and water quality model, version 4.0. Port- 6968.1987.tb02379.x.
land State University, Portland 828 pp. Müller, S., 2001. ATV Gewässergütemodell: Beschreibung der Modell-
Cooper, A.B., 1986. Developing management guidelines for river nitroge- bausteine. Deutsche Vereinigung für Wasserwirtschaft, Abwasser
nous oxygen demand. J. Water Pollut. Control Fed. (58), 845–852. und Abfall e.V..
Deltares, 2011. Delft3D Processes Library, Detailed description of Pro- Neitsch, S.L., Arnold, J.G., Kiniry, J.R., Williams, J.R., 2011. Soil and Water
cesses, Technical Reference Manual, Water Quality and Aquatic Ecol- Assessment Tool (SWAT): Theoretical Documentation. Version 2009.
ogy, Version: 4.00. Deltares 376 pp. Blackland Research Center, Texas Agriculture Experiment Station;
DHI, 2003. ECOLAB Water Quality Templates Scientific Description. DHI Grassland Soil and Water Research Laboratory, Agricultural Research
Water and Environment 58 pp. Service, Texas 494 pp.
Fettig, J., Pick, V., 2013. Final report of the joint research project ‘‘Treat- Ocampo-López, C., Colorado-Arias, S., Ramı́rez-Carmona, M., 2015.
ment of tapioca wastewater and sustainable water pollution control Modeling of microbial growth and ammonia consumption at different
management of key economic zones in southern Vietnam" (Schluss- temperatures in the production of a polyhydroxyalkanoate (PHA)
bericht, Verbundsvorhaben "Reinigung von Tapiokaabwasser und biopolymer. J. Appl. Res. Technol. 13 (5), 498–503, http://
nachhaltiges Gewässerschutzmanagement in der Haupt- dx.doi.org/10.1016/j.jart.2015.10.001.
wirtschaftszone in Südvietnam"). Hochschule Ostwestfalen-Lippe Pambrun, V., Paul, E., Spérandio, M., 2006. Modeling the partial nitrifica-
(in German). tion in sequencing batch reactor for biomass adapted to high ammo-
Fettig, J., Pick, V., Austermann-Haun, U., Blumberg, M., Phuoc, N.V., 2013. nia concentrations. Biotechnol. Bioeng. 95 (1), 120–131, http://
Treatment of tapioca starch wastewater by a novel combination of dx.doi.org/10.1002/bit.21008.
physical and biological processes. Water Sci. Technol. 68 (6), 1264– Pamment, N.B., Hall, R.J., Barford, J.P., 1978. Mathematical modeling of lag
1270, http://dx.doi.org/10.2166/wst.2013.354. phases in microbial growth. Biotechnol. Bioeng. 20 (3), 349–381,
Groeneweg, J., Sellner, B., Tappe, W., 1994. Ammonia oxidation in Nitro- http://dx.doi.org/10.1002/bit.260200304.
somonas at NH3 concentrations near km: effects of pH and tempera- Pauer, J., Auer, M.T., 2000. Nitrification in the water column and sediment
ture. Water Res. 28 (12), 2561–2566, http://dx.doi.org/10.1016/0043- of a hypereutrophic lake and adjoining river system. Water Res. 34
1354(94)90074-4. (4), 1247–1254, http://dx.doi.org/10.1016/S0043-1354(99)00258-4.
Grunditz, C., Dalhammar, G., 2001. Development of nitrification inhibition Peng, Y., Zhu, G., 2006. Biological nitrogen removal with nitrification and
assays using pure cultures of Nitrosomonas and Nitrobacter. Water denitrification via nitrite pathway. Appl. Microbiol. Biotechnol. 73 (1),
Res. 35 (2), 433–440, http://dx.doi.org/10.1016/S0043- 15–26, http://dx.doi.org/10.1007/s00253-006-0534-z.
1354(00)00312-2. Prats, C., 2008. Individual-based modelling of bacterial cultures in the
Gujer, W., 2010. Nitrification and me – a subjective review. Water Res. 44 study of the lag phase. (Dissertation)Universitat Politècnica de
(1), 1–19, http://dx.doi.org/10.1016/j.watres.2009.08.038. Catalunya, Barcelona.
Hagopian, D.S., Riley, J.G., 1998. A closer look at the bacteriology of Prosser, J.I. (Ed.), 1990. Autotrophic Nitrification in Bacteria: In: Advances
nitrification. Aquacult. Eng. 18 (4), 223–244, http://dx.doi.org/ in Microbial Physiology, vol. 30. .
10.1016/S0144-8609(98)00032-6. Rheinheimer, G., 1991. Mikrobiologie der Gewässer. 5 überarbeitete
Hellinga, C., van Loosdrecht, M., Heijnen, J.J., 1999. Model based design of Auflage. Gustav Fischer Verlag Jena.
a novel process for nitrogen removal from concentrated flows. Math. Rheinheimer, G., Hegemann, W., Raff, J., Sekoulov, I., 1988. Nitrogen cycle
Comput. Model. Dyn. Syst. 5 (4), 351–371, http://dx.doi.org/10.1076/ in water (Stickstoffkreislauf im Wasser. Stickstoffumsetzung in nat-
mcmd.5.4.351.3678. ürlichen Gewässern, in der Abwasserreinigung und Wasserversor-
Hipsey, M.R., Romero, J.R., Antenucci, J.P., Hamilton, D., 2006. Computa- gung) Oldenburg Verlag, München (in German).
tional Aquatic Ecosystem Dynamics Model: CAEDYM, v2.3 Science Rosso, L., Bajard, S., Flandrois, J.P., Lahellec, T.C., Fournaud, J., Veit, P., 1996.
Manual. Centre for Water Research. University of Western Australia. Differential growth of Listeria monocytogenes at 4 and 8 8C: conse-
Hsiao, S.S.-Y., Hsu, T.-C., Liu, J.-w., Xie, X., Zhang, Y., Lin, J., Wang, H., Yang, quences for the shelf life of chilled products. J. Food Prot. 59 (9), 944–
J.-Y.T., Hsu, S.-C., Dai, M., Kao, S.-J., 2014. Nitrification and its oxygen 949, http://dx.doi.org/10.4315/0362-028X-59.9.944.
consumption along the turbid Chang Jiang River plume. Biogeos- Schön, G.H., Engel, H., 1962. Der Einfluss des Lichtes auf Nitrosomonas
ciences 11 (7), 2083–2098, http://dx.doi.org/10.5194/bg-11-2083- Europaea Win. Arch. Microbiol. 42 (4), 415–428.
2014. Sebilo, M., Billen, G., Mayer, B., Billiou, D., Grably, M., Garnier, J., Mariotti,
Jiménez, E., Giménez, J.B., Ruano, M.V., Ferrer, J., Serralta, J., 2011. Effect of A., 2006. Assessing nitrification and denitrification in the Seine river
pH and nitrite concentration on nitrite oxidation rate. Bioresour. and estuary using chemical and isotopic techniques. Ecosystems 9 (4),
Technol. 102 (19), 8741–8747, http://dx.doi.org/10.1016/j.bior- 564–577, http://dx.doi.org/10.1007/s10021-006-0151-9.
tech.2011.07.092. Senzia, M.A., Mayo, A.W., Mbwette, T., Katima, J., Jørgensen, S.E., 2002.
Krause, P., Boyle, D.P., Bäse, F., 2005. Comparison of different efficiency Modelling nitrogen transformation and removal in primary faculta-
criteria for hydrological model assessment. Adv. Geosci. 5 (5), 89–97, tive ponds. Ecol. Model. 154 (3), 207–215, http://dx.doi.org/10.1016/
http://dx.doi.org/10.5194/adgeo-5-89-2005. S0304-3800(02)00018-2.
Laanbroek, H.J., Bollmann, A., 2011. Nitrification in inland waters: In: Stenstrom, M.K., Poduska, R.A., 1980. The effect of dissolved oxygen
Ward, B.B., Arp, D.J., Klotz, M.G. (Hrsg). Nitrification. American Society concentration on nitrification. Water Res. 14 (6), 643–649, http://
for Microbiology, Washington, DC, USA, pp. 385–403. dx.doi.org/10.1016/0043-1354(80)90122-0.
Le, T.T.H., 2014. Entwicklung eines detaillierten Modells zur prozessba- van Hulle, S., Volcke, E.I.P., Teruel, J.L., Donckels, B., van Loosdrecht, M.CM,
sierten Simulation der Nitrifikation in Gewässern und Einbindung in Vanrolleghem, P.A., 2007. Influence of temperature and pH on the
 T.T.H. Le et al. / Ecohydrology & Hydrobiology 19 (2019) 54–65 65

kinetics of the Sharon nitritation process. J. Chem. Technol. Biotech- Wrage, N., Velthof, G., van Beusichem, M., Oenema, O., 2001. Role of
nol. 82 (5), 471–480, http://dx.doi.org/10.1002/jctb.1692. nitrifier denitrification in the production of nitrous oxide. Soil Biol.
Volcke, E., 2006. Modelling, analysis and control of partial nitritation in a Biochem. 33 (12-13), 1723–1732, http://dx.doi.org/10.1016/S0038-
SHARON reactor. (Dissertation)Ghent University, Ghent. 0717(01)00096-7.
Wett, B., Rauch, W., 2003. The role of inorganic carbon limitation in Xia, X.H., Yang, Z.F., Huang, G.H., Zhang, X.Q., Yu, H., Rong, X., 2004.
biological nitrogen removal of extremely ammonia concentrated Nitrification in natural waters with high suspended-solid content––a
wastewater. Water Res. 37 (5), 1100–1110, http://dx.doi.org/ study for the Yellow River. Chemosphere 57 (8), 1017–1029, http://
10.1016/S0043-1354(02)00440-2. dx.doi.org/10.1016/j.chemosphere.2004.08.027.
Wetzel, R., 2001. Limnology, Lake and River Ecosystems, 3rd ed. Elsevier Yongming, X., 1988. A study of nitrification of the Toujiang River. Int. J.
Academic Press, San Diego. USA. Environ. Stud. 31 (4), 297–303, http://dx.doi.org/10.1080/
Wheaton, F.W., Hochheimer, J.N., Kaiser, G.E., Krones, M.J., Libey, G.S., 00207238908710434.
Easter, C.C., 1994. Nitrification filter principles. In: Timmons, M.B., Zwietering, M.H., Jongenburger, I., Rombouts, F.M., van’ t Riet, K., 1990.
Losordo, T.M. (Hrsg.). Aquaculture Water Reuse Systems: Engineering Modeling of the bacterial growth curve. Appl. Environ. Microbiol. 56
Design and Management, pp. 101–126. (6), 1875–1881.