Journal of Environmental Management 90 (2009) 18241832

Contents lists available at ScienceDirect

Journal of Environmental Management

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

An innovative modeling approach using Qual2K and HEC-RAS integration to

assess the impact of tidal effect on River Water quality simulation
Chihhao Fan a, *, Chun-Han Ko b, Wei-Shen Wang c
a

Department of Safety, Health, and Environmental Engineering, Ming Chi University of Technology, 84 Gung-Juan Rd, Taishan, Taipei County 24301, Taiwan
School of Forest and Resources Conservation, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd, Da-An District, Taipei 10617, Taiwan
c
Graduate School of Biochemical Engineering, Ming Chi University of Technology, 84 Gung-Juan Rd, Taishan, Taipei County 24301, Taiwan
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 27 December 2007
Received in revised form
4 August 2008
Accepted 24 November 2008
Available online 1 January 2009

Water quality modeling has been shown to be a useful tool in strategic water quality management. The
present study combines the Qual2K model with the HEC-RAS model to assess the water quality of a tidal
river in northern Taiwan. The contaminant loadings of biochemical oxygen demand (BOD), ammonia
nitrogen (NH3-N), total phosphorus (TP), and sediment oxygen demand (SOD) are utilized in the Qual2K
simulation. The HEC-RAS model is used to: (i) estimate the hydraulic constants for atmospheric reaeration constant calculation; and (ii) calculate the water level prole variation to account for concentration changes as a result of tidal effect. The results show that HEC-RAS-assisted Qual2K simulations
taking tidal effect into consideration produce water quality indices that, in general, agree with the
monitoring data of the river. Comparisons of simulations with different combinations of contaminant
loadings demonstrate that BOD is the most import contaminant. Streeter-Phelps simulation (in combination with HEC-RAS) is also performed for comparison, and the results show excellent agreement with
the observed data. This paper is the rst report of the innovative use of a combination of the HEC-RAS
model and the Qual2K model (or Streeter-Phelps equation) to simulate water quality in a tidal river. The
combination is shown to provide an alternative for water quality simulation of a tidal river when
available dynamic-monitoring data are insufcient to assess the tidal effect of the river.
2008 Elsevier Ltd. All rights reserved.

Keywords:
Water quality modeling
Qual2K
Streeter-Phelps
HEC-RAS
Tidal river

1. Introduction
The prudent utilization of water resources has been an important issue in public policy for decades, and much effort has been
expended in developing effective water management strategies to
ensure sufcient high-quality water supplies (Ning et al., 2001;
Elshorbagy and Ormsbee, 2006; Eatherall et al., 1998; Horn et al.,
2004). In this regard, water quality modeling is increasingly
recognized as a useful tool for acquiring valuable information for
optimal water quality management.
Over the years, several water quality models have been developed for various types of water bodies (such as rivers, lakes,
reservoirs, estuaries, and so on). Some of these models have
included basic water quality indices (such as dissolved oxygen and
biochemical oxygen demand), whereas others have included
more sophisticated water quality criteria (such as eutrophication
levels and toxicity impacts). For example, Thayer and Krutchkoff
(1967) applied a 3-dimensional algorithm to advanced water
quality spatial analysis, and Pelletier et al. (2006) conrmed the

* Corresponding author. Tel.: 886229089899x4656; fax: 886229080783.

E-mail address: dillon@mail.mcut.edu.tw (C. Fan).
0301-4797/$ see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2008.11.011

applicability and exibility of the QUAL2Kw framework for simulation of river water quality.
However, Lindenschmidt (2006) has suggested that complex
models are not necessarily the most useful models. This is because
abundant monitoring data are required for the estimation, calibration, and verication of model parameters, and some complex
simulation models can involve multiple parameters that have never
been previously measured or reported. In some instances, these
parameters have been estimated by other algorithms. For these
reasons, the use of complicated models for water quality simulation
is problematical, and the simulated results might not be as reliable
as they are purported to be.
Given these circumstances, the utilization of basic water quality
models has become a preferable option, and several successful
examples have been reported in the literature (Mahamah,1998; Drolc
and Koncan, 1996). However, the application of basic water quality
models has certain limitations. In particular, transport phenomena in
rivers are often estimated with simplied assumptions, and parsimonious models can be incapable of accurate assessments of
complex hydrodynamics (such as tidal effects in rivers).
Against this background, the present study investigates the
applicability of a basic model to water quality simulation of a tidal
river. The Qual2K model was selected for the study because of its

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

wide applicability and ease of use. The model calculates water

quality indices using various combinations of contaminant loadings
including biochemical oxygen demand (BOD), ammonia nitrogen
(NH3-N), total phosphorus (TP), sediment oxygen demand (SOD),
oxygen consumption by algae growth, and so on. Because Qual2K
does not consider tidal effects in its calculations, the HEC-RAS
model is employed in this study to assess the impact of tidal effect
on water quality simulation. Using these models, simulated results
are compared with monitoring data from the Keelung River in
northern Taiwan. This river was selected for the study because
more than two million people live in its catchment area and the
river water is employed for domestic water supply.
This paper represents the rst report of this innovative use of
combined models for water quality simulation in a tidal river.
2. Research setting
The Keelung River is a major river in the Taipei area of Taiwan. The
river, which originates from the Snow Mountain of Taiwan at an
elevation of approximately 3000 m, is 80 km in length and has
a watershed area of approximately 501 square kilometres. The population living in the rivers watershed has increased markedly in

1825

recent years in association with signicant urban and economic

development. Unfortunately, these developments have been
accompanied by marked increases in contaminant loadings in the
river, which, in combination with an inadequate sewerage system,
has resulted in serious water pollution problems in the Keelung River.
To estimate the contaminant loadings in the Keelung River, the
watershed area of the river has been divided into 41 pollution subdivisions (K1K41) in accordance with topographical and
geographical characteristics. In each sub-division, BOD, NH3-N, and
TP discharges due to domestic and industrial uses have been estimated under conditions of 60% sewerage-system coverage. The
geographical location of the Keelung River and the pollution subdivisions of the watershed are shown in Fig. 1. The BOD, NH3-N, and
TP loadings in each pollution sub-division are presented in Table 1.
According to the water quality monitoring protocol followed by
the Taiwan Environment Protection Administration (EPA), BOD is
the most important contaminant for analysis. Domestic sewage has
been identied as the most signicant source of BOD pollution,
whereas the wastewater from pig farms is the least signicant
source. During the years between 2000 and 2006, BOD discharge
due to domestic sewage, woodland non-point sources, industrial
wastewater, and pig-farms wastewater were 41,698 kg/day,

Fig. 1. Keelung River watershed: (a) geographical location; (b) pollution sub-divisions and locations of water quality monitoring stations.

1826

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

Table 1
Estimation of BOD, NH3-N, and TP discharge in pollution sub-divisions in Keelung
River watershed.
Sub-Division K1
BOD
NH3N
TP

K13

K14 K15

K24

K25 K26

377 1811 341

70 315 66
25
115 22

Sub-Division K34
BOD
NH3N
TP

K4

2365 1688 358 5039

431 296 335
911
158
110 23
334

Sub-Division K23
BOD
NH3N
TP

K3

K5

K6

K7

K8

1019 4199 695 2387 1322 1106 2623 7388

197 763 125 434 241 198 485 1332
68 280 46
159
88
74
174 492

Sub-Division K12
BOD
NH3N
TP

K2

53
105
39

K35
676
9
3

647
115
42

K36 K37
111
18
7

428
68
24

K16
86
15
5
K27
333
56
21
K38
221
31
11

K17
28
11
1
K28
122
19
7
K39
156
26
10

K18
887
160
58
K29
746
132
49
K40
191
31
12

K19

K9

K10 K11
2
0
0

K20

14 590
1 105
0
39

K31

K32 K33

318 1710 542 446

63 295 84
82
20 104 31
29
K41
261
44
16

unit: kg/day.

921 kg/day, 691 kg/day, and 189 kg/day, respectively, and accounted
for 95.9%, 2.1%, 1.6%, and 0.4% of the total BOD discharge in the
watershed, respectively (Fan and Wang, 2006).
To improve the water quality of Keelung River, eight sewagepumping stations (S1S8) had been installed along the river to
intercept and pump domestic sewage to a wastewater treatment
plant. These sewage-pumping stations had intercepted the BOD
discharge and river water quality had promptly improved as
a consequence. The operational capacities of these sewage-pumping stations had thus become critical to river water quality.
3. Models
3.1. Selection of models
The applications of a variety of water quality models including
QUAL2E, QUAL2K, and WASP have been reported in the literature
(Kannel et al., 2007; Palmieri and de Carvalho, 2006; Kim et al.,
2004; Park and Lee, 2002). As previously noted, the Qual2K was
chosen for the present study because of its popularity and ease of
application. Detailed descriptions and applications of this model
are available in the literature (Park and Lee, 2002; Chapra and
Pelletier, 2003). In addition, because the Qual2K model does not
consider tidal effects in its calculations, the HEC-RAS model was
employed to assess the impact of tidal effect on water quality
simulation.
3.2. Qual2K model
In the Qual2K model, the mass balance for a constituent (C) in an
element (i) is expressed as,

Q out;i
dCi
Q
Q
E
C i1 Ci1  Ci
i1 Ci1  i Ci 
dt
Vi
Vi
Vi i
Vi
Ei
Wi
Si
Ci1 Ci
Vi
Vi

HEC-RAS, which was developed by the United States Army Corps

of Engineers, has been applied extensively in calculating the
hydraulic characteristics of rivers (Carson, 2006; Pappenberger
et al., 2005). It is an integrated program to calculate water surface
proles using the following energy equation (Brunner, 2002):

K21 K22

33 1028 575 576

10 197 116 104
1
68 37
37
K30

3.3. HEC-RAS model

in which: Ci, Vi, Qi, Ei, Wi represent the concentration, volume,

efuent, dispersion coefcient, and external constituent loading of
element i, respectively; Si represents the sources and sinks of the
constituent due to reactions and mass transfer mechanisms in
element i; and Qout,i represents ow abstraction from element i.
To simulate the water quality, the degradation constant of each
contaminant and hydraulic characteristics of ow velocity and
water depth are required.

Y2 Z2

a2 V22
2g

Y1 Z1

a1 V12
2g

he

(2)

in which: Y represents water depth; Z represents channel elevation;

is a velocity weighting coefcient;
V represents average velocity; a
he represents energy head loss; g represents gravitational acceleration; and subscripts 1 and 2 denote cross sections 1 and 2,
respectively.
To execute the model, details of cross sections of the river and
upstream ow rate are required. Using the energy conservation
equation, the velocity and water depth of the given cross section are
calculated.
4. Methodology
Dynamic simulation of water quality was not applicable in the
present study because there was a deciency of data on water
quality monitoring and hydraulic characteristics with regard to the
Keelung River. Because domestic sewage had been identied as the
major source of pollution, the contaminant loadings of BOD, NH3-N,
TP, and SOD were included in the Qual2K model, and other
contaminant loadings were assumed to be negligible. Variations in
such environmental factors as wind, heat exchange, and solar
radiation were also assumed to be negligible. The temperature was
assumed to be constant at 20  C. The water quality indices (DO,
BOD, NH3-N, and TP) were calculated.
For simulation of water quality, the river was divided into 71
continuous ow stirred tank reactors (CSTRs) in series. The schematic segmentation of the Keelung River and the locations of
domestic sewage discharge from 41 pollution sub-divisions are
shown in Fig. 2. In this gure, the locations of eight sewagepumping stations with operational capacity are also presented.
In accordance with common practice, monitoring data were
used to calibrate the required parameters of the water quality
model at the beginning of the study. To minimize the error between
simulation results and observed data, the constants of BOD and
NH3-N degradation had to be adjusted. With the calibrated
constants, the simulation still deviated from the observed data
signicantly in the downstream (010 km) region. Using the BOD
degradation constant (k1) as an example, the utilization of the
calibrated constant (0.91 day1) did not provide a BOD value that
was in agreement with the monitored data, which implies that
parameter determination through calibration might not have
calculated the concentration accurately in the downstream region
(010 km) of the Keelung River.
Empirical degradation constants for BOD, NH3-N, TP and SOD
were therefore used. These degradation constants had been evaluated in previous studies. The values for BOD degradation
constants (k1) are summarized in Table 2. The k1 used in the study
was 0.16 day1 (TEPA, 1997). The NH3-N degradation constant used
in the study was 0.03 day1 (SSOTCG, 2005). The SOD degradation
coefcients ranged from 0.1 to 2.03 gO2/m2 day (Huang, 2005;
SSOTCG, 2005).
To calculate the water quality, the ow velocity and water depth
were required. They can be calculated using Eqs. (3) and (4), as
proposed by Leopold and Maddock (1953):

U aQ b

(3)

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

K38~K41

K34
K33, K35

K31

K30

Station S8
K28

Collection
S-D:K27
Capacity:
657,000CMD

K27

K25, K26

Station S7
K24

Collection
S-D:K24
Capacity:
17,000CMD

K21, K22

K20
K19

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

1827

K36,K37

Station S4
Collection
S-D:K15

K32

K15,K16

Capacity:
150,000CMD

K09,K11

Station S1
K29

Collection
S-D:K07
Capacity:
9,000CMD

K07

K05,K06

Station S6
Collection
S-D:K23
Capacity:

K03,K04

1,730CMD

K23
K02

K01

Station S5
Collection
S-D:K18

41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71

Station S3
Collection
S-D:K13
Capacity:
18,500CMD
K13,K14
K12

K08,K10

Station S2
Collection
S-D:K08
Capacity:
18,000CMD

Capacity:
2,000CMD
K17,K18

Fig. 2. Modeling segmentation, sewage-pumping station locations, and pollution discharges in Keelung River.

H cQ d

(4)

in which: Q is the ow rate (m3/s); and a, b, c, and d are hydraulic

characteristic constants.
The hydraulic characteristic constants can be obtained from: (i)
reported studies; (ii) Manning equation calculation; or (iii) HECRAS calculations. In this study, the empirical hydraulic characteristic constants were adopted from Chapra and Pelletier (2003). For
the Manning equation calculation, hydraulic parameters of riverbed

slopes, coefcients of roughness, and wetting periphery were

collected or estimated to calculate the hydraulic constants. These
two methods were included as calculation modules in the Qual2K
model. Details can be obtained from the literature (Park and Lee,
2002; Huang, 2005).
For the HEC-RAS model, steady ow water surface prole was
computed. The average tidal level (0.03 m above mean sea level) at
the river mouth was selected as the downstream boundary
condition because the water quality monitoring data, which were

1828

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

Table 2
Literature values for BOD Degradation Constant (k1).

Table 3
Flow rates, ow velocities, and water depths in 71 CSTRs of Keelung Rivera.

K1 (20?)

References

0.160.21
0.07
0.100.20
0.15
0 < H < 2.4 m 0.3(H/8)0.434
H >> 2.4 m 0.3
H: water depth (m)

Wool et al., 2001

TCEPB, 2005
TEPA, 1997
SSOTCG, 2005

Chapra, 1997

utilized for comparison with the simulation results, are obtained

under average tidal conditions. A total of 155 hydraulic cross
section proles of the Keelung River from the website of the Water
Resources Agency (Taiwan) and ow rate were applied to the HECRAS model to calculate the ow velocities and water depths. For
a given hydraulic cross section, the ow velocity and water depth
were calculated at a specic ow rate. Similar calculations were
conducted at seven comparable ow rates for the given cross
section. All the calculated ow velocities and water depths at
various ow rates were regressed (using Eqs. (2) and (3)) to obtain
the hydraulic characteristic constants of a, b, c, and d for the specic
cross section. The calculation was performed on all cross sections,
and all of these hydraulic constants were then employed in the
subsequent water quality simulation. If a CSTR contained several
hydraulic cross section proles, the hydraulic constants (a, b, c, and
d) of all cross section proles within that CSTR were averaged to
represent the hydraulic characteristics. In the regression, the R
square values for parameters a and b ranged from 0.542 to 1.0, with
an average of 0.973 and a standard deviation of 0.059. The R square
values for parameters c and d ranged from 0.930 to 1.0, with an
average of 0.985 and a standard deviation of 0.015.
Having obtained the constants of a, b, c, and d, the ow rate of
the river was substituted into Eqs. (3) and (4) to calculate the water
depth and ow velocity. The water depth and ow velocity were
applied to Eqs. (5)(7) (below) to assess the atmospheric re-aeration constant at a given cross section (Liu et al., 2005; Chapra and
Pelletier, 2003; Chapra, 1997; OConnor and Dobbins, 1956):

U 0:5
OConnor-Dobbins : k2 3:93 1:5
H

Hm < 0:61

U 0:67  
H m > 0:61 m and Hm
H1:85

> 3:45U 2:5

(5)

Owens-Gibbs : k2 5:32

U
Churchill : k2 5:026 1:67
H

(6)

situation other than the above

(7)

in which: U represents ow velocity (m/s); and H represents water

depth (m).
Simulations were conducted using contaminant loadings,
contaminant degradation constants, and re-aeration constants
from the three respective methods. All of the calculated ow rates,
ow velocities, and water depths are presented in Table 3. Fig. 3
presents two examples of the cross section proles with water
depths calculated with HEC-RAS.
5. Results and discussion
5.1. Qual2K simulation
Fig. 4 shows the Qual2K simulations of water quality using BOD,
NH3-N, SOD, and TP as the contaminant loadings. For comparison,

CSTR
Q, cms
U, m/s
H, m

1
1.66
0.73
0.49

2
1.94
0.74
0.48

3
1.94
0.66
0.60

4
5
2.10 2.10
0.59 0.53
0.58 0.61

6
7
2.34 2.28
0.55 0.59
0.77 0.79

8
2.28
0.81
0.35

9
2.68
0.79
0.70

10
2.68
0.44
0.90

11
2.68
0.36
1.15

12
2.68
0.38
1.53

CSTR
Q, cms
U, m/s
H, m

13
1.92
0.37
0.90

14
2.12
0.14
1.06

15
1.97
0.27
0.60

16
1.97
0.21
0.88

17
1.97
0.30
1.00

18
1.97
0.25
1.54

19
1.97
0.10
1.94

20
2.03
0.29
0.96

21
2.03
0.26
1.73

22
2.03
0.12
1.78

23
2.03
0.17
1.13

24
2.97
0.20
1.23

CSTR
Q, cms
U, m/s
H, m

25
2.97
0.10
1.28

26
2.97
0.15
0.92

27
3.35
0.13
0.86

28
4.00
0.07
1.39

29
4.00
0.15
1.13

30
4.00
0.06
1.71

31
4.00
0.05
3.49

32
4.00
0.50
2.00

33
4.32
0.23
1.37

34
4.32
0.27
0.81

35
4.32
0.16
1.50

36
5.10
0.11
1.36

CSTR
Q, cms
U, m/s
H, m

37
5.27
0.18
1.05

38
39
5.65 5.65
0.29 0.13
1.94 1.41

40
5.65
0.07
2.91

41
5.65
0.05
3.54

42
5.65
0.05
3.54

43
5.65
0.25
1.83

44
6.76
0.22
1.68

45
6.76
0.18
1.25

46
6.76
0.17
1.50

47
7.48
0.09
1.95

48
7.86
0.06
3.17

CSTR
Q, cms
U, m/s
H, m

49
8.09
0.05
3.87

50
8.09
0.06
3.26

51
8.09
0.05
3.92

52
9.41
0.04
3.11

53
9.41
0.04
2.82

54
9.41
0.03
3.41

55
9.98
0.03
2.56

56
9.98
0.03
4.43

57
9.98
0.02
9.62

58
10.30
0.02
10.75

59
10.30
0.02
7.98

60
10.30
0.02
10.65

CSTR
Q, cms
U, m/s
H, m

61
10.30
0.02
8.50

62
12.05
0.02
7.01

63
12.05
0.02
8.37

64
12.05
0.02
8.26

65
12.05
0.02
7.24

66
12.89
0.02
7.96

67
12.89
0.02
6.15

68
13.39
0.02
6.86

69
13.39
0.02
10.83

70
13.39
0.02
9.45

71
13.39
0.02
5.36

a
The ow rate in each CSTR is calculated by accumulating in-stream ow rates
with discharges of surface runoff and domestic sewage. The ow velocities and
water depths are calculated by HEC-RAS simulation.

water quality monitoring data collected on 5 September 2003 are

also presented in Fig. 4 for comparison. It is noteworthy that the
observed data showed an improved outcome with full operation of
S1 and S4 sewage-pumping stations. In the simulation, the required
hydraulic constants were either adopted from reported empirical
studies or obtained from the Manning equation calculation.
The Qual2K simulations deviated from the monitored data for
DO concentration. For BOD, employment of empirical hydraulic
constants resulted in a comparable trend with the monitored data,
whereas calculation with the Manning equation led to an overestimation in BOD concentration. For the Keelung River, in which
hydro-geological characteristics vary signicantly, the hydraulic
calculation module in Qual2K might not capture the characteristics
needed to provide accurate simulation.
To improve the simulation, the HEC-RAS-calculated hydraulic
constants were employed. Fig. 4 also shows the Qual2K simulation
using HEC-RAS-calculated hydraulic constants for comparison. A
comparable trend of calculated results with monitoring data in DO
concentration can be observed, and BOD simulation again deviated
from monitoring data in the downstream region.
In Qual2K, water quality is calculated by considering transport
phenomena as well as the chemical and biological reactions
occurred in the river, whereas hydraulic calculation is performed
with empirical constants or the Manning equation. For the Keelung
River, the parameters used in the Manning equation were estimated, and calculation might not be accurate. Deviation might also
occur for calculations using empirical values. Therefore, the utilization of empirical constants or the Manning equation might not be
capable of describing the hydraulic characteristics of the Keelung
River, and discrepancies between simulations and monitored data
were observed.
With the HEC-RAS model, the required hydraulic constants were
calculated, and the simulation was improved compared with the
Qual2K simulation using either the Manning equation or empirical
constants. However, the HEC-RAS-assisted Qual2K simulation still
deviated from the monitored data, and the discrepancy resulting
from the tidal effect (that is, water level variation due to tidal effect)

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

1829

Fig. 3. Examples of HEC-RAS calculated water depths and cross section proles of Keelung River.

was not considered. Detailed tidal effect calculation is presented in

Section 5.3.
5.2. HEC-RAS-assisted Qual2K simulations with different
contaminant loadings
Fig. 5 presents DO and BOD simulations from Qual2K using
various combinations of contaminant loadings and HEC-RAScalculated hydraulic constants. Concentrations of NH3-N and TP are
simulated and presented in Fig. 6. The water quality monitoring
data are again presented for comparison.
HEC-RAS-assisted Qual2K simulation using contaminant loadings of BOD, NH3-N, SOD, and TP resulted in the lowest DO
concentration (0.5 mg/L) at the river mouth, whereas simulation
using BOD as the only contaminant loading led to a relatively higher
DO concentration (1.37 mg/L). For BOD, the HEC-RAS-assisted

Qual2K simulations using various combinations of contaminant

loadings resulted in comparable aqueous concentrations. This is
because the rst order BOD degradation was assumed, and inclusion
of other contaminant loadings does not affect BOD kinetic calculation. Among the HEC-RAS-assisted Qual2K simulations, concentration differences of DO and BOD at the river mouth (maximum
difference of 0.9 mg/L for DO, and 0.5 mg/L for BOD) were observed,
but these simulations were, in general terms, comparable.
As shown in Fig. 5, inclusion of TP as a contaminant loading did
not cause a noticeable difference in DO concentration. This was
because there was minimal TP discharge in the Keelung River
watershed compared with contaminant loadings of BOD and NH3-N.
Moreover, TP is usually considered as the limiting nutrient for
aquatic bioactivity, and it is largely removed through biodegradation or sedimentation. In the Keelung River watershed, the massive

2003/09/05

2003/09/05

10

BOD+NH3-N+SOD+TP
BOD+NH3-N+SOD
BOD+NH3-N
BOD
SP eq
Observed
00~04 5yr ave.

10
8

DO mg/L

DO mg/L

8
6

6
4

4
2
2
0
0
0

10

20

30

40

50

10

10

20

30

40

50

60

20

30

40

50

60

60
12

HEC-RAS calculated a,b,c, and d

Empirical values of a,b,c, and d
Manning euqation calculated
Observed-DO
00~04 5yr ave.-DO

BOD mg/L

10
8
6

10

BOD mg/L

12

8
6
4

4
2
2
0
0
0

10

20

30

40

50

60

Distance from River Mouth(km)

Fig. 4. Qual2K water quality simulations of Keelung River using empirical approach,
Manning equation, and HEC-RAS for hydraulic characteristics estimation.

Distance from River Mouth(km)

Fig. 5. Qual2K and Streeter-Phelps simulations of DO and BOD (using HEC-RAScalculated re-aeration constants) of water quality of Keelung River without considering
tidal effect.

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

2003/09/05
10000
BOD+NH3-N+SOD+TP (no tidal effect)
BOD+NH3-N+SOD+TP (with tidal effect)
Observed
00~04 5yr ave.

NH3-N ug/L

8000
6000
4000
2000
0
0

10

10

20

30

40

50

60

20

30

40

50

60

1400

TP ug/L

1200
1000
800
600
400
200
0

Distance from River Mouth (km)

Fig. 6. Qual2K simulations of NH3-N and TP (using HEC-RAS-calculated re-aeration
constants) of Keelung River.

ow rate prevented settlement of the TP. The substantial amount of

BOD discharge consumed most of the DO and minimal TP was
biodegraded aerobically. All of the discharged TP was accumulated
in a slightly increasing trend along the river. This explanation is
supported by the TP concentration prole in Fig. 6.
SOD has been considered anther important oxygen sink in
several studies (Lindenschmidt, 2006; Tyagi et al., 1999). A
comparison of DO simulations with and without SOD contaminant
loading exhibited a similar general trend. A concentration difference
(ranging from 0.05% to 20.43%, with an average of 7.43%) occurred
between 30 and 12 km and 8.5 and 3.6 km from the river mouth.
Fig. 5 also shows comparable DO simulations with and without
consideration of NH3-N. Compared with the BOD degradation
constant, the NH3-N degradation constant was relatively small, and
NH3-N degradation was less competitive than BOD degradation.
Neglecting NH3-N as a contaminant loading in water quality
simulation did not have a signicant impact on DO concentration.

contaminants within the tidal section were re-calculated by

assuming the complete mixing of river and tidal water in each CSTR.
With Qual2K, the contaminant concentrations in the river water
were obtained through model simulation. For the tidal water, the
contaminant aqueous concentrations at the river mouth (0.9 mg/L
for DO, 2.5 mg/L for BOD, 4.75 mg/L for NH3-N, and 0.412 mg/L for
TP) were adopted. For the water volumes, the river water volume
was estimated through HEC-RAS calculation. The tidal river volume
was calculated as the volume difference between the HEC-RAS
calculated river water surface proles using no limit and average
tidal level as the downstream boundary conditions. Fig. 7 shows
the HEC-RAS calculated water level proles of the Keelung River as
well as its riverbed elevation prole.
The DO and BOD concentrations (taking the tidal effect into
consideration) are presented in Fig. 8, and the calculated NH3-N
and TP concentrations with tidal effect taken into consideration are
presented in Fig. 6. It is apparent that the calculated aqueous
concentrations taking tidal effect into consideration agreed with
the observed data. At the river mouth, the downstream BOD
concentration was reduced from 8.68 mg/L to 2.64 mg/L, and the
DO concentration was increased from 0.32 mg/L to 0.88 mg/L.
For the simulation of a tidal river, having more monitoring data
always leads to better modeling results. Unfortunately, in many
countries, water quality monitoring work is either not conducted at
all or is performed only periodically. Incomplete hydraulic data and
insufcient water quality monitoring data make the application of
complicated models impracticable. In these circumstances, the
utilization of a basic model for water quality assessment becomes
the only option. In the present study, integrated utilization of HECRAS and Qual2K has been shown to be capable of assessing water
quality in the Keelung River. Through the two-stage HEC-RASassisted Qualt2K water quality assessment process (that is, aqueous
concentration calculation with better estimation on hydraulic
retention time and atmospheric re-aeration constant, followed by

50
River bed Elev.
Water Surface Elev.
(no downstream limit)
Water Surface Elev.(ave., +0.03)

40

Elevation(m)

1830

30
20
10
0
-10
-20
0

20

30

40

50

60

Elevation(m)

Employment of HEC-RAS for calculation of hydraulic constants

resulted only in improved estimation of aqueous contaminant
degradation due to aquatic kinetic reactions. Aqueous concentration change due to an increase in water volume as a result of tidal
effect was not considered. No signicant difference in DO concentration was observed, which was due to the fact that the river
mouth tidal water contained comparable amounts of DO, thus
resulting in little change in DO concentration. For BOD, the
concentration discrepancy decreased as the distance from the river
mouth increased.
In the Keelung River, the downstream riverbed is at a slope of 1/
6700, and the water quality within this section is affected signicantly by the tidal effect up to 31 km upstream. Water at the river
mouth is driven back upstream during high tide, and the increase in
water volume results in changing contaminant concentrations. To
account for this tidal effect, the aqueous concentrations of the

10

Distance from River Mouth (km)

5.3. HEC-RAS-assisted Qual2K simulation with tidal effect

-5
Bottom Elev.
Water Surface Elev.
(no downstream limit)
Water Surface Elev.(ave., +0.03)

-10

-15
0

10

20

30

Distance from River Mouth (km)

Fig. 7. Riverbed elevation prole and HEC-RAS calculated water surfaces using no
boundary limit and average tidal level as the boundary conditions in the Keelung
River downstream.

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

in which: k20 is the rate constant at 20  C; kT is the rate constant at

the temperature of T; and q is the temperature correction coefcient (being 1.047 and 1.028 for k1 and k2, respectively).
To calculate the DO decit, the DO saturation concentration
(Cs) at a specic temperature can be calculated with Eq. (11)
(APHA, 1985):

2003/09/05
BOD+NH3-N+SOD+TP
BOD+NH3-N+SOD
BOD+NH3-N
BOD
SP eq
Observed
00~04 5yr ave.

10

DO mg/L

8
6





Cs 139:39 1:5757  105 Tk1  6:6423  107 Tk2




1:2438  1010 Tk3  8:6219  1011 Tk4


 0:5535S 0:031929  19:428Tk1 3867Tk2

4
2
0
0

10

10

20

30

40

50

60

20

30

40

50

60

12

BOD mg/L

10
8
6
4
2
0

Distance from River Mouth(km)

Fig. 8. Qual2K and StreeterPhelps simulations (using HEC-RAS-calculated re-aeration
constants) of water quality of Keelung River taking tidal effect into consideration.

concentration correction with tidal effect), the water quality of

a tidal river can be accurately calculated.
5.4. StreeterPhelps simulation
As noted above, BOD is the most important aqueous contaminant in the Keelung River, and the calculated DO concentration in
Fig. 8 agrees with the trend in monitoring data without considering
HH3-N, TP, and SOD. To verify the legitimacy of using BOD as the
only contaminant loading in the simulation, the StreeterPhelps
equation was utilized, and the simulation result was compared to
the monitoring data.
In the StreeterPhelps equation, DO and BOD are the only two
water quality indices considered. The equation calculates DO
concentration by considering the oxygen source and sink of
atmospheric re-aeration and BOD consumption, respectively
(Streeter and Phelps, 1925; Chapra, 1997).
By assuming steady state conditions and solving for the mass
balance equation, the basic StreeterPhelps equation yields:


k1 L0  k1 t
e
 ek2 t D0 ek2 t
k2  k1

L L0 ek1 t

(8)

(9)

in which: D is DO decit; Do is initial DO decit; k1 is the BOD

degradation constant; k2 is the atmospheric re-aeration constant; L
is the BOD concentration (mg/L); Lo is the ultimate BOD (mg/L); and
t is the hydraulic retention time.
The rate constants of k1 and k2 are usually determined at 20  C. If
necessary, these rate constants can be extrapolated to other
temperatures through the Arrhenius relationship:
T20

kT k20 q

1831

(10)

(11)

in which: Tk is the ow absolute temperature; and S is salinity (mg/L).

The salinity is estimated from observed electric conductivity.
In Fig. 5, the StreeterPhelps calculated DO concentrations are
presented. The simulation agreed with the monitoring data as well
as with the trend in the HEC-RAS-assisted Qual2K simulation. The
similarity in trend of DO concentration supported the contention
that the StreeterPhelps equation is almost as good as Qual2K in
calculating the water quality of the Keelung River.
For BOD, both Qual2K and the StreeterPhelps equation provide
similar estimations. A comparison of the StreeterPhelps-calculated result with the monitoring data revealed an overestimation of
50% in BOD concentration at the river mouth. This overestimation
with the StreeterPhelps simulation can be improved by taking the
tidal effect into account.
Fig. 8 shows the StreeterPhelps simulation with the HEC-RAS
tidal effect taken into consideration. This shows that the DO and
BOD simulations by the StreeterPhelps equation are in agreement
with monitoring data as well as with the simulation by the HECRAS-assisted Qual2k simulation. This consistency between the
Qual2K and StreeterPhelps simulations conrms the legitimacy of
selecting the StreeterPhelps equation as an alternative in the
water quality simulation of the Keelung River.
6. Conclusion
Water quality modeling has been shown to be a useful tool for
water resources management. However, the use of a basic model
for water quality simulation of a river becomes unavoidable when
the available data on hydraulic characteristics and water quality
monitoring are limited.
In this study, Qual2K simulations using various combinations of
contaminant loadings of BOD, NH3-N, SOD, and TP have been
compared with monitoring data of the Keelung River in northern
Taiwan. With HEC-RAS calculation included in the water quality
simulation, the atmospheric re-aeration constant was evaluated
more accurately, and the tidal effect on the river quality was also
accounted for. In general, simulation using a combination of Qual2K
and HEC-RAS was in good agreement with the monitored data
trends.
Because BOD has been shown to be the most important
contaminant loading in the Keelung River, the StreeterPhelps
equation was also employed to simulate the water quality of the
river. HEC-RAS was again introduced to account for the tidal effect.
Inclusion of HEC-RAS in the model simulation again improved the
water quality assessment signicantly.
This paper represents the rst report of an attempt to combine
HEC-RAS and Qual2K (or the StreeterPhelps equation) to simulate
the water quality of a tidal river. The study demonstrates that the
HEC-RAS-assisted two-stage water quality assessment process
(that is, aqueous concentration calculation with better estimation
on hydraulic retention time and atmospheric re-aeration constant,
followed by concentration correction with tidal effect) enables the
water quality of a tidal river to be accurately assessed. Given that

1832

C. Fan et al. / Journal of Environmental Management 90 (2009) 18241832

many basic water quality models in common use are incapable of

calculating the water quality of a tidal river, the introduction of
HEC-RAS into water quality simulation to account for the tidal
effect has been shown to be a signicant advance in estimating
water quality in tidal rivers.
References
APHA, 1985. Standard Methods for the Examination of Water and Wastewater, 15th
ed. American Public Health Association, Washington DC.
Brunner, G.W., 2002. HEC-RAS, River Analysis System Hydraulic Reference Manual.
Hydrological Engineering Center, US Army Corps of Engineers, Davis, CA.
Carson, E.C., 2006. Hydrologic modeling of ood conveyance and impacts of historic
overbank sedimentation on West Fork Blacks Fork, Uinta Mountains, northeastern Utah, USA. Geomorphology 75, 368383.
Chapra, S.C., 1997. Surface WaterQuality Modeling. McGraw-Hill Inc, New York.
Chapra, S.C., Pelletier, G.J., 2003. QUAL2K: a Modeling Frame Work for Simulating
River and Stream Water Quality: Documentation and Users Manual. Civil and
Environmental Department, Tufts University, Medford, MA.
Drolc, A., Koncan, J.Z., 1996. Water quality modeling of the river SAVA, Slovenia.
Water Research 30, 25872592.
Eatherall, A., Boorman, D.B., Williams, R.J., Kowe, R., 1998. Modeling in-stream water
quality in LOIS. Science of the Total Environment 210/211, 499517.
Elshorbagy, A., Ormsbee, L., 2006. Object-oriented modeling approach to surface
water quality management. Environmental Modeling and Software 21, 689698.
Fan, C., Wang, W.-S., 2006. Application of Streeter-and-Phelps equation to the aquatic
environment management a case study based on water quality monitoring
data of Keelung river, Taiwan. Environmental Informatics Archive 4, 3549.
Horn, A.L., Rueda, F.J., Hormann, G., Fohrer, N., 2004. Implementing river water
quality modeling issues in mesoscale watershed models for water policy
demands-an overview on current concepts, decits, and future tasks. Physics
and Chemistry of the Earth 29, 725737.
Huang, T.Y., 2005. QUAL2K applied on the derivation of the optimal sanitary sewer
collection rates in the non-tidal inuenced section of Damshui river. Master
thesis, National United University (in Chinese).
Kannel, P.R., Lee, S., Lee, Y.S., Kanel, S.R., Pelletier, G.J., 2007. Application of automated QUAL2Kw for water quality modeling and management in the Bagmati
river, Nepal. Ecological Modeling 202, 503517.
Kim, D., Wang, Q., Soriala, G.A., Dionysioua, D.D., Timberlakeb, D., 2004. A model
approach for evaluating effects of remedial actions on mercury speciation
and transport in a lake system. Science of the Total Environment 327, 115.

Leopold, L.B., Maddock, T., 1953. The hydraulic geometry channels and some
physiographic implications. Geological survey Professional paper 252,
Washington D.C.
Lindenschmidt, K.E., 2006. The effect of complexity on parameter sensitivity and
model uncertain in river water quality modeling. Ecological Modeling 190,
7286.
Liu, W.C., Liu, S.Y., Hsu, M.S., Kuo, A.Y., 2005. Water quality modeling to determine
minimum instream ow for sh survival in tidal rivers. Journal of Environmental Management 52, 5566.
Mahamah, D.S., 1998. Simplifying assumptions in water quality modeling. Ecological Modeling 109, 295300.
Ning, S.K., Chang, N., Yang, L., Chen, H.W., Hsu, H.Y., 2001. Assessing pollution
prevention program by QUAL2E simulation analysis for the Kao-Ping river
Basin, Taiwan. Journal of Environmental Management 61, 6176.
OConnor, D., Dobbins, W., 1956. Mechanism of Reaeration in Natural Streams.
Transactions of American Society of Civil Engineers 123, 641684.
Palmieri, V., de Carvalho, R.J., 2006. Qual2e model for the Corumbatai River.
Ecological Modeling 198, 269275.
Pappenberger, F., Beven, K., Horritt, M., Blazkova, S., 2005. Uncertainty in the calibration of effective roughness parameters in HEC-RAS using inundation and
downstream level observations. Journal of Hydrology 302, 4669.
Park, S.S., Lee, Y.S., 2002. Water quality modeling study of the Nakdong river, Korea.
Ecological Modeling 152, 6675.
Pelletier, G.J., Chapra, S.C., Tao, H., 2006. QUAL2Kw-a framework for modeling water
quality in stream and rivers using a genetic algorithm for calibration. Environmental Modeling and Software 21, 419425.
SSOTCG, 2005. Water Pollution Monitoring, Evaluation and Reduction Plan in Keelung
River. Sewage System Ofce, Taipei City Government, Taipei, Taiwan (in Chinese).
Streeter, H.W., Phelps, E.B., 1925. A Study of Pollution and Natural Purication of the
Ohio, vol. 146. Public Health Bulletin, 175 pp.
TCEPB, 2005. Detailed Design and Water Quality Planning in Ching-Mei Creek and
Keelung River. Taipei County Environmental Protection Bureau, Taipei, Taiwan
(in Chinese).
TEPA, 1997. Management on Tan-Sui River Watershed Pollution Remediation System.
Taiwan Environmental Protection Administration, Taipei, Taiwan (in Chinese).
Thayer, H.P., Krutchkoff, R.G., 1967. Stochastic model for BOD and DO in Streams.
Journal of the Sanitary Engineering Division, American Society of Civil Engineers ASCE SA3, 5972.
Tyagi, B., Gakkhar, S., Bhargava, D.S., 1999. Mathematical modeling of stream
DO-BOD accounting for settleable BOD and periodically varying BOD source.
Environmental Modeling and Software 14, 461471.
Wool, T.A., Ambrose, R.B., Martin, J.L., Comer, E.A., 2001. Water Quality Analysis
Simulation Program (WASP) Version 6.0 Draft Users Manual. US Environmental
Protection Agency Region 4, Atlanta, GA.