SWQM - A simple river water quality model for assessment of urban ...

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
SWQM - A simple river water quality model for assessment of
urban wastewater discharges
M. Schütze 1* , F. Reußner 2 , J. Alex 1
1 ifak e. V. Magdeburg, Werner-Heisenberg-Str. 1, 39106 Magdeburg, Germany
2 TU Darmstadt, Petersenstr. 13, 64287 Darmstadt, Germany
* Corresponding author, email: manfred.schuetze@ifak.eu
ABSTRACT
This paper presents a river water quality model (“Simple Water Quality Model – SWQM”),
which has been set up specifically for its use within the context of assessment of urban
discharges from combined sewer overflows and wastewater treatment plants into receiving
rivers. It includes the relevant processes necessary for description of dissolved oxygen and
ammonia concentrations in the river, whilst, at the same time, it has been deliberately kept
simple in its complexity in order to facilitate its application also under practical (data-scarce)
conditions. pH modeling is included via alkalinity and carbonate balances, thus allowing the
evaluation of NH3 concentrations (which may be toxic to aquatic life). An example
application is presented and references to application to water quality modeling using SWQM
of real rivers are given.
KEYWORDS
Impact; River water quality; Simple Water Quality Model; SWQM; urban discharges
INTRODUCTION
In many urban wastewater systems, wastewaters are finally discharged into rivers as receiving
water bodies. These include continuous effluents of wastewater treatment plants and
intermittent discharges of sewer systems (rainwater discharges; combined sewer overflows).
In the last decades, a paradigm shift could be observed with regard to their assessment –
whilst in the past, solely end-of-pipe standards have been applied (emission-based approach),
increasingly environmental quality objectives, considering the impacts on receiving water
quality (immission-based approach), are being applied. Such development has been driven by
several initiatives, including the European Water Framework Directive and the development
of the Urban Pollution Management Manual in the UK. A review of various approaches will
be presented elsewhere at this conference (Blumensaat, 2011).
For the proper assessment of the impacts of urban discharges (rainwater discharges; combined
sewer overflows, wastewater treatment effluent), river water quality modeling is required.
Approaches applied so far in this context range from simple mixing to static modeling of
selected events and, even further, to the application of complex dynamic models for historic
time series. Guidelines in this context include the Urban Pollution Management Manual
(FWR,1998), suggesting the selection of individual events and then, if necessary, simulating
them using a set of detailed simulation models. The German document BWK-M3
(BWK,2004) suggests simplified static modeling of a single discharge for nine block rain
events. For those cases where such a simplified approach does not appear to be feasible, the
BWK-M7 guideline recommends setting up a detailed model and using this for simulation.
Blumensaat et al. (2011, this conference) provide an overview of guidelines currently applied
in Central Europe and provide related references.
Schütze et al . 1

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
For practical application, however, neither approach appears to be satisfying: Whilst, on the
one hand, simple approaches might be quick to apply, they lack from features necessary to be
considered when analyzing multiple discharges (e.g. several discharge locations into the same
receiving water body, …), on the other hand, approaches of high complexity and data demand
may not be feasible in “daily life applications” by consenting authorities. The present
contribution attempts to fit this gap by suggesting a water quality model covering the most
relevant processes and by embedding it not only in two software environments, but also, at
the same time, in a consenting framework.
REQUIREMENTS ON A WATER QUALITY MODEL RIPE FOR
PRACTICAL APPLICATION
For river water quality modeling, Rauch et al. (1998) suggest an application-specific model
set up. This was also the intention of the IWA working group compiling the River Water
Quality Model No. 1 (RWQM) (Reichert et al., 2001), which serves as a baseline for
application-specific submodels. Besides many other packages, also the QUAL2 family of
models is widely applied. However, for practical application within the context of the daily
routine processes of authorities responsible for consenting procedures, a river model should
comply with requirements such as:
• Consideration of all relevant processes and water quality objectives:
Even though the importance of particular process may vary from case study to case
study, a generally applicable model approach should comprise of the relevant
processes, without “overloading” the model by non-essential processes. Within the
central European context, assessment of urban discharges is usually carried out with
regard to Dissolved Oxygen and ammonia (NH3-N), the latter being toxic to fish life.
This is also in line with current guidelines. As ammonia (NH3-N) is a key parameter
for the assessment, it should be considered directly. Therefore it is necessary to
consider pH in the water quality model as well, as pH usually is different in various
types of urban discharges.
• Consideration of different characteristics of sewer and treatment plant discharges:
As sewage discharges and treatment plant effluent have different characteristics, such
as different content of easily biodegradable matter, these have to be considered in an
appropriate way.
• Limited data requirements:
For application on a wider scale (such as within a consent setting context), the model
should not require (much) more data than would be readily available. Whilst data on
river geometry and profiles might still be available from digital cadastres of the
responsible government agencies, data on river water quality might have been sampled
only on a rare basis if at all they are available.
• Limited number of parameters, avoiding ambiguity:
In order to a river water quality model to be applied also by non-specialists, it should
have a limited number of parameters, which can be set easily, in a non-ambigious way.
Often, users tend to use default values.
• Easy to use, also by non-experts:
Though this requirement may sound trivial, non-trivial handing of the model
application might render the model unsuitable for practical application. This refers not
only to the nature of the user interface, but also to the procedure of preparing the input
data, of performing simulation runs and of processing output.
2 SWQM - A simple river water quality model

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
THE SIMPLE WATER QUALITY MODEL (SWQM)
Based on these requirements, the Simple Water Quality Model (SWQM) has been developed,
which describes five processes (degradation of organic matter, nitrification, sedimentation,
photosynthesis and reaeration) of seven fractions, allowing water quality to be assessed with
regard to dissolved oxygen (DO) and ammonia (NH3-N). A sixth process, gas exchange (CO2)
with the atmosphere, has been included in the model description, but has, for simplicity, so far
been “switched off” by corresponding parameter settings. As opposed to the RWQM with its
23 processes, 24 fractions and 107 parameters, SWQM is much reduced in complexity, yet
includes the main processes and parameters (see Table 1 and Table 2):
Table 1. Constituents being modeled by the Simple Water Quality Model (SWQM)
Name Description
SO Dissolved oxygen
SNH Ammonium + ammonia (NH4+NH3)
SI Soluble inert COD fraction, including the non-settled part of the particulate inert COD fraction (XI)
SS Soluble biodegradable COD fraction, including the non-settled part of the particulate biodegradable
COD fraction (XS) and XH.
SCON Conservative (user-defined) substance, i.e. not subject to any degradation process
SALK Alcalinity (HCO3 equivalent)
SIC Dissolved carbonates (inorganic carbon) (SALK+CO2)
Organic matter is considered here in two fractions. It should be noted, however, that “SI” and
“SS” here have a slightly different meaning than the fractions of identical names in the
Activated Sludge Models of wastewater treatment plants (see also Figure 1). Partitioning of
organic matter into two fractions allows considering combined sewer overflow discharges
(usually having higher fraction of inert material) in a different way than wastewater treatment
plant effluent discharges (with usually its readily biodegradable fractions being reduced
during wastewater treatment).
The soluble inert fraction of COD has been combined in the model with the non-sedimented
portion of the particulate inert fraction to form the inert fraction SISWQM of the SWQM,
assuming that the remaining portion of the particular inert matter sediments to the river bed
and, thus, does not contribute to degradation of organic matter. Sediment oxygen demand is
considered separately. In a similar way, the biodegradable soluble fraction and a portion of
the biodegradable particulate matter are combined to form the biodegradable fraction SSSWQM
of the SWQM. A similar approach of splitting organic demand into two fractions had been
suggested by Lijklema (1996), using however BOD as a base for describing organic matter.
COD
COD soluble COD particulate
SI SS XI XS
inert biodegradable inert biodegradable
SI SISWQM SWQM
Non-sedimented part
SS SWQM
Non-sedimented part
Figure 1. Fractionation of COD in Activated Sludge Models and in SWQM
XBH
Schütze et al . 3

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
SCON has been introduced as an (optional) user-definable conservative substance, i. e. a
fraction not being subject to any degradation processes. For example, it could be used to
include phosphorous in a simplistic way when necessary. SALK and SIC are used in order to
calculate pH, which is required for determination of the ammonia (NH3-N) part of SNH, using
the well-known relation by Emerson et al. (1975). The difference of these two fractions (SIC
– SALK) represents the CO2 contents of the liquid phase. SWQM therefore allows also to
model CO2 and its exchange with the atmosphere (Alex, 2009a).
Table 2 provides a summary of the biochemical transformations considered within SWQM
using the Petersen-Gujer notation as established in wastewater treatment modeling and also
used in the RWQM descriptions. As SWQM also considers pH and carbonate, it allows the
determination of NH3–N (toxic to aquatic life) rather than just using NH3+NH4 as a lumped
constituent.
Table 2. SWQM - the Simple Water Quality Model
Among the processes modelled by SWQM, reaeration is crucial one for the DO balance.
Although a wide variety of relationships describing the reaeration rate (most of them using
average values of water depth and velocity) has been described in the literature (see, for
example, Chapra (1997) and Bowie et al. (1985) for reviews, the modified approach by Wolf
(ATV-DVWK, 2002) has been applied here. For wide ranges of velocities and water depths,
this approach renders reaeration rates similar to other approaches. For very low water levels,
however, most equations in the literature, due to their exponential nature, result in very
(unrealistically) high reaeration coefficients. A correction term introduced by Wolf – hence
the term “modified approach by Wolf” - ensures that also for low water depths, seemingly
realistic reaeration coefficients k2 will be determined:
k
2
40 0
( 3 + ) v
k
=
h
St
1.
7
0
. 7
0.
5
+
h
⎧
with h0
= ⎨
⎩
h
− 0.
588
( 10 −13.
5*
h)
for h ≥ 0.
5
else
(kSt: roughness coefficient [m -1/3 /s], v: flow velocity [m/s], h: water depth [m])
As opposed to the many other model implementations of river water quality models where
often average values for water depth and velocity are used in the water quality module, both
implementations of the SWQM reported in this paper (see next section) use the actual values
4 SWQM - A simple river water quality model
[ 1/
d]

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
of water depth and velocity in each river section. This is of relevance when considering
wastewater discharges and adds to greater accuracy, as discharges lead to different values of
water flows and levels in particular during the period of CSO discharges.
The description of the nitrification process corresponds to other water quality models. The
stoichiometric coefficient YA-64/14 has, for YA =0.24 (see, e.g. Gujer et al., 1999) the value
-4.33, as it is also in many other water quality models. The (small) influence of nitrification
on alkalinity is described by the factor -2/14*facSALK. However, as facSALK has been set to
zero, this influence is neglected here.
Sediment oxygen demand is, in line with other water quality models (e.g. Lijklema, 1996,
QUAL2E; Bowie et al., 1992), represented in simplified form by a single coefficient (SOD).
By division of this coefficient by the water depth h, the effect of reduced influence of
sediment oxygen demand on oxygen balance of the entire water phase has been considered.
The coefficient iC_COD allows consideration of the influence on alkalinity.
Photosynthesis has, depending on the locally prevailing conditions, major influence on the
oxygen balance. However, in general, detailed data, describing solar irradiation, cloudiness,
shadowing of plants on the river banks etc. are not available for the specific river section.
Therefore, in order to ensure the applicability of the model also in data-scarce applications,
the impact of photosynthesis is considered in a simplified way through the parameter alpha
and the variable beta which considers the sun’s elevation (being a function of date, time and
geographical location).
Decay of organic matter reduces, under consumption of oxygen, the contents of SS, whilst
generating inert material SI, as described by the factor fXI. fXI is determined as the product of
the corresponding factors fXI,ASM3 YSTO, YA of the Activated Sludge Model No. 3 (Henze et
al.,2000), thus describing the stoichiometry of the conversion of SS to XI. Furthermore,
nitrogen is generated within this decay process (cf. parameter iNSS). Since the XI fraction is
not modelled explicitly within SWQM, it is recommended to use a value of iNSS lower than its
original value in the ASM3 (iNSS = iNSS,original – fXI*fNXI= 0.03 – 0.2*0.04 = 0.022). The
parameter iNSS also describes the influence of the decay process on the alkalinity. However, in
order to ensure compatibility with other models, iNSS has been, at present, set to zero.
INTEGRATION WITH HYDRAULIC MODELLING
After the foundations of the water quality model had been laid, it has been linked to hydraulic
modelling of the water flows and levels in the river. Using different software implementations
(described in the subsequent section), SWQM has been linked to both (hydrodynamic and
hydrologic), approaches. Implementations have been done for the conventional means of
pollutant transport (being based on the concept of continuously stirred tank reactors (CSTR))
as well as using a Lagrangian approach to pollutant transport modelling, thus avoiding errors
frequently introduced by numerical dispersion (which can be quite large, Muschalla and Alex,
2010), on two software platforms: SIMBA and the BlueM package (Bach et al., 2009). The
following section describes a sample implementation of SWQM using hydrodynamic flow
modeling combined with the conventional (CSTR-based) approach to pollutant transport
modeling.
Schütze et al . 5

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
EXAMPLE APPLICATION
As an example of application of SWQM, a hypothetical river described in the literature
(Schütze et al., 2002) has been modelled. It consists of a 40 km stretch of a river with
identical profile in all river sections. It has a base flow (from upstream) of 1500 l/s. The
wastewater treatment plant effluent has been determined by simulation of adaptations of the
example sewer system (A128 example) and wastewater treatment plant of the Simba
simulation system (see ifak, 2009); its average flow amounts to 318 l/s. In a similar way, CSO
discharges have been obtained by simulation. Figure 3a illustrates the simulation set up.
Baseflow is assumed to enter the river at Section E0, whilst CSO discharges are into Section
E6 and wastewater treatment effluent discharges into Section E11. Each section is assumed to
have a length of 1 km. Hydrodynamic simulation was carried out using an extended version of
SWMM, which has been integrated into SIMBA, thus allowing arbitrary biochemical
transformation processes (such as the SWQM) to be modelled. As wastewater treatment
modelling took place using the Activated Sludge Model No. 1 (Henze et al., 2000), modelling
results (using the pollutant fractions of ASM1) had to be converted into the variable set of
SWQM. Appropriate converter functions have been implemented. In a similar way, sewer
system modelling results had been converted from sewer system variables (COD, TKN) into
the SWQM variables in order to describe the discharges into the river in an appropriate way.
1000
0 1 2 3 4 5
Time [d]
6 7 8 9 10
6 SWQM - A simple river water quality model
[l/s]
7000
6000
5000
4000
3000
2000
Selected river sections
link E07 flow
link E13 flow
link E20 flow
link E27 flow
link E34 flow
link E40 flow
Figure 3. River hydraulics a) Model setup b) flow results in various river sections
Figure 3b illustrates the flows in the river system, with the CSO discharges into Section E6
being clearly visible. Figure 4a displays DO concentrations at selected river sections vs. time.
The small influence of the diurnal variation of discharges as well as of photosynthetic activity
can be seen. DO sag curves are clearly visible, also showing that, for the given river and
discharge characteristics, the minimum DO concentration lies far downstream of the
discharge location. Finally, Figure 5 illustrates the evaluation of river water quality, based on
simulation of a real one year rainfall series. Here, the evaluation according to the Urban
Pollution Management Manual (FWR, 1998), counting exceedances of certain threshold
values of critical DO and NH3-N concentrations over exposure times of 1 hour, 6 hours and 1
day, indicates that the river has problems with low oxygen concentrations, showing itself as
frequent short-duration occurrence of critically low DO concentrations. A similar analysis
with regard to NH3-N concentrations yields that no problems with regard to ammonia are
present (not shown here). The evaluation routine has been set up in a buffered way, hence
enabling it to process arbitrarily long time series (ifak, 2009).

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
[mg/l]
8
7
6
5
4
3
2
1
Concentrations in selected river sections
link E07 SO
link E13 SO
link E20 SO
link E27 SO
link E34 SO
link E40 SO
0
0 1 2 3 4 5
Time [d]
6 7 8 9 10
0
0 1 2 3 4 5
Time [d]
6 7 8 9 10
Schütze et al . 7
[mg/l]
0.06
0.05
0.04
0.03
0.02
0.01
Concentrations in selected river sections
link E07 NH3
link E13 NH3
link E20 NH3
link E27 NH3
link E34 NH3
link E40 NH3
Figure 4. SWQM application for example river a) DO results b) ammonia (NH3) results
Frequency
Results of UPM evaluation for Dissolved Oxygen - River Section E13
15
10
5
0
1 hour
6 hours
Duration
1 day
1 month
3 months
1 year
Return period
Figure 5. SWQM application for example river a) UPM evaluation for DO in Section E13
Sensitivity analyses of the parameters have been carried out and also some recommendations
for setting of the parameter values, considering also the outcomes of an analytic study of the
SWQM (Alex, 2009b) have been given elsewhere (Reussner et al., 2010).
OTHER IMPLEMENTATIONS
Other implementations of SWQM include its integration with hydrologic flow routing and a
Lagrange-based pollutant transport model, using SIMBA as software platform: The
Lagrangian modeling approach to pollutant transport, considering parcels of water travelling
along the river stretch, avoids problems of numerical dispersion (as frequently observed in
modeling practice) (Muschalla and Alex, 2010). An application of this to the Nítra river in
Slovakia is reported by Černochová et al. (2010), using SWQM for modeling of the Nitra
river in Slovakia
Another implementation of SWQM consists its integration within a Lagrange-based pollutant
transport model, within in the BlueM modeling environment. BlueM uses OpenMI as a
modeling interface between different subsystems of the urban water system (Reußner et al.,
2009). This integration of sewer system and river modeling forms part of a suggested
consenting procedure (Brehmer et al., 2009) for urban discharges, allowing consideration of
dynamic discharges from CSOs and wastewater treatment plants also of several urban

12 nd International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15 September 2011
catchments into the same river. Within this context, SWQM has been applied to three river
systems within the German Federal state of Hessen.
CONCLUSION
In summary it can be stated that a simple water quality model has been developed, and its
applicability to practical case studies been illustrated. It is hoped that this paper stimulates
discussion and encourages application of river modeling as a means to assess impacts of
urban discharges to the riverine environment.
References
Alex, J. (2009a): Modellierung von Masse-Algenkulturen. Internal report. ifak Magdeburg
Alex, J. (2009b): Stability, spatial and temporal discretisation selection for numerical water quality simulation
models. Internal report. ifak – Institut für Automation und Kommunikation e. V. Magdeburg
ATV-DVWK (2002): Handbuch ATV-DVWK-Gewässergütemodell; ATV-Arbeitsgruppe GB-4.2; Hennef/Sieg
Bach, M., Froehlich, F., Heusch, S., Hübner, C., Muschalla, D., Reußner, F., Ostrowski, M. W. (2009): BlueM -
a free software package for integrated river basin management, in: Fohrer, N., Schmalz, B., Hörmann,
G., Bieger, K. (eds.): Hydrologische Systeme im Wandel, 109-115. ISBN: 978-3-941089-54-9
Blumensaat, F., Staufer, P., Heusch, S., Reussner, F., Rieckermann, J., Schmuck, S., Schütze, M., Seiffert, S.,
Gruber, G., Zawilski, M. (2011): Water-quality based assessment of urban drainage impacts in Europe -
where do we stand today? Submitted to 12th ICUD, Porto Alegre, September 2011
Bowie GL, Mills WB, Porcella,DB, Campbell CL, Pagenkopf JR, Rupp GL, Johnson KM, Chan PWH, Gherini
SA, Chamberlin CE (1985): Rates, Constants, and Kinetic Formulations in Surface Water Quality
Modeling. U.S. Environmental Protection Agency. Athens Georgia, EPA/600/3-85/040
Brehmer, I., Reußner, F., Schütze, M., Muschalla, D., Ostrowski, M. (2009): Weiterentwicklung des hessischen
„Leitfadens zum Erkennen ökologisch kritischer Gewässerbelastungen durch Abwassereinleitungen –
Entwicklung einer simulationsgestützten Planungsmethodik“; Korrespondenz Abwasser, 56, 4, 382-384
BWK (2004). Ableitung von immissionsorientierten Anforderungen an Misch- und Niederschlagswassereinleitungen
unter Berücksichtigung örtlicher Verhältnisse. Bund der Ingenieure für Wasserwirtschaft,
Abfallwirtschaft und Kulturbau (BWK) e.V..
Černochová, L., Schütze, M., Derco, J. (2010): River Water Quality Modelling of Nitra River Basin in Slovakia
(as a Part of Integrated Modelling). In 1st IWA Austrian National Young Water Professionals
Conference, 9.-11. June 2010, Austria.
Chapra, S. C. 1997. Surface water-quality modeling. McGraw-Hill
Emerson, K., Russo, R.E., Lund, R.E., Thursteon, R.V. (1975): Aquaeous ammonia equilibrium concentrations:
effects of pH and temperaure. J. Fish. Res. Bd. Can. 32, 2379-2383
FWR. (1998). "Urban Pollution Management Manual. 2nd Edition.", Foundation for Water Research, Marlow.
Henze, M., Gujer, W., Mino, T., van Loosdrecht, M. 2000. Activated Sludge Models ASM1, ASM2, ASM2d
and ASM3. Scientific and Technical Re-port No. 9, IWA London.
Lijklema, L., R. H. Aalderink, und H. de Ruiter. 1996. Procesbeschrijvingen DUFLOW Zuurstofhuishouding in
stromende en stagnante Water-systemen. Landbouwuniversiteit Wageningen
Muschalla, D., Alex, J. (2010): La pipe - A Lagrange-based quality and transport model, Workshop of Central
European Simulation Group (HSG); Lungern; 25.-27.11.2010
Rauch W, Aalderink H, Krebs P, Schilling W, Vanrolleghem P (1998a) Requirements for integrated wastewater
models - driven by receiving water objectives. Water Sci. Technol. 38,11,97-104 ###EDH???
Reichert, P., Borchardt, D., Henze, M., Rauch, W., Shanahan, P., Somlyódy, L., Vanrolleghem, P.A. (2001):
River Water Quality Model No. 1, Edited by IWA Task Group on River Water Qualty Monitornng,
IWA, Scientific and Technical Report No. 12
Reussner, F., Alex, J., Bach, M., Schütze, M., Muschalla, D. 2009. Basin-wide integrated modelling via OpenMI
considering multiple urban catchments. Wat. Sci. Tech. 60, 5, 1241-1248.
Reußner, F., Schütze, M., Ostrowski, M. (2010): Dokumentation des Forschungsvorhabens „Modifizierung des
Leitfadens zum Erkennen kritischer Gewässerbelastungen durch Abwassereinleitungen - Entwicklung
einer simulationsgestützten Analyse- und Planungsmethodik“. Dokumentation des Werkzeugpakets zur
integrierten hydraulischen und stofflichen Modellierung für die immissionsorientierte
Nachweisführung, Technische Universität Darmstadt
Schütze, M., Butler, D., Beck, M.B. (2002): Modelling, Simulation and Control of Urban Wastewater Systems.
Springer Verlag; ISBN 1-85233-553-X
8 SWQM - A simple river water quality model