- Email: [email protected]
Development of a water quality modeling system for river pollution index and suspended solid loading evaluation
Journal of Hydrology 478 (2013) 89–101
Contents lists available at SciVerse ScienceDirect
Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol
Development of a water quality modeling system for river pollution index and suspended solid loading evaluation Y.C. Lai a, Y.T. Tu a, C.P. Yang b, R.Y. Surampalli c, C.M. Kao a,⇑ a
Institute of Environmental Engineering, National Sun Yat-Sen University, Kaohsiung 804, Taiwan Center for Teaching Excellence, National Pingtung University of Science & Technology, Pingtung County, Taiwan c Department of Civil Engineering, University of Nebraska, Lincoln, NE, USA b
a r t i c l e
i n f o
Article history: Received 13 August 2012 Received in revised form 13 November 2012 Accepted 21 November 2012 Available online 1 December 2012 This manuscript was handled by Peter Laurent Charlet, Editor-in-Chief, with the assistance of Hossein Ghadiri, Associate Editor Keywords: Ammonia nitrogen (NH3–N) River Pollution Index (RPI) River water quality Suspended solid (SS) Watershed management
s u m m a r y The Kaoping River Basin is the largest and most extensively used watershed in Taiwan. In the upper catchment, the non-point source (NPS) pollutants including suspended solid (SS) and ammonia nitrogen (NH3–N) are two major water pollutants causing the deterioration of Kaoping River water quality. Because SS is one of the four parameters involving in the River Pollution Index (RPI) calculation, it needs to be carefully evaluated to obtain the representative water quality index. The main objective of this study was to develop a water quality modeling system to obtain representative SS and RPI values for water quality evaluation. In this study, a direct linkage between the RPI calculation and a water quality model [Water Quality Analysis Simulation Program (WASP)] has been developed. Correlation equations between Kaoping River flow rates and SS concentrations were developed using the field data collected during the high and low flows of the Kaoping River. Investigation results show that the SS concentrations were highly correlated with the flow rates. The obtained SS equation and RPI calculation package were embedded into the WASP model to improve interactive transfers of required data for water quality modeling and RPI calculation. Results indicate that SS played an important role in RPI calculation and SS was a critical factor during the RPI calculation especially for the upper catchment in the wet seasons. This was due to the fact that the soil erosion caused the increase in the SS concentrations after storms. In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Results demonstrate that the integral approach could develop a direct linkage among river flow rate, water quality, and pollution index. The introduction of the integrated system showed a significant advance in water quality evaluation and river management strategy development. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction Compared to point source pollution, nonpoint-source (NPS) pollution is more diffuse and harder to isolate and control (Markku et al., 2010). The NPS pollution is defined legally as any pollution not originating from a statutory point source, which can be carried over by rainfall and irrigation water to enter a water body dispersedly via surface runoff. It has been well-documented that NPS pollution, such as nutrient runoff and atmospheric deposition, contribute significant pollutant loading to water bodies. The NPS pollutants can cause the deterioration of water quality through the release of suspended solids (SSs), nutrients, pesticides, fertilizer, and other sources of inorganic and organic matter. Because many factors, such as topography, soil characteristics, and rainfall intensity, affect their quantity and quality, variety types of NPS simulation models have been used to account for ⇑ Corresponding author. Tel.: +886 7 5254413; fax: +886 7 5254449. E-mail address: [email protected] (C.M. Kao). 0022-1694/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhydrol.2012.11.050
the integrated effects of the hydrological cycle and land cover in relation to nutrient yield (Kavvas et al., 2006; Papanicolaou and Abaci, 2005; Xu et al., 2010). With the aid of various environmental models, the improvement of the estimation and control of NPS pollution in watershed has been enhanced greatly in recent years. To assess the effectiveness of land-use-management policy, some applications seek improved modeling approaches for predicting the water quality effects of storm events as a function of land topography, land cover, and land use leading to the development of various applications (Chen et al., 2006; Ouyang et al., 2009). These models, for assessing NPS loads in the agricultural field, generally simulate rainfall, erosion, runoff sediment, temperature, wind speed, atmosphere pressure, and NPS processes (Lin et al., 2010; Luo et al., 2012). The Kaoping River Basin, located in the southeast region of Taiwan, is the largest and the most intensively utilized river basin in Taiwan. It is 171 km in length, drains a catchment of more than 3625 km2, and has a mean flow of 239 m3/s. Fig. 1 shows the location of Kaoping River Basin, Kaoping River, its catchment, and three
90
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 1. Kaoping River Basin and three major reaches.
major reaches (Chi-San Creek, I-Liao Creek, and Lao-Non Creek). Although the mean annual rainfall in this river basin is close to 3000 mm, over 90% appears in the wet season. The period of high flow rate in the stream usually occurs in the late spring and summer due to the impacts of monsoon and typhoon (Lin, 2010; Lai, 2010). NPS SS pollutants due to the severe storm events would cause significant adverse impacts on the river water quality of Kaoping River, and also cause the increased turbidity of the river water at the intake location of the downstream water treatment plant. Thus, the NPS SS pollution should be effectively evaluated and controlled. Taiwan Environmental Protection Administration (TEPA) has developed a River Pollution Index (RPI) classification system for river water quality evaluation based on the purpose of water usage and degree of protection for each stream section (TEPA,
2002). The RPI involves four parameters: dissolved oxygen (DO), biochemical oxygen demand (BOD), SS, and ammonia nitrogen (NH3–N), each of which is ultimately converted to a four-state quality sub-index (1, 3, 6, and 10). The overall index is then divided into four pollution levels. Table 1 presents the equation for RPI calculation and criteria for the four RPI classes (good, slightly polluted, moderate polluted, and gross polluted). Table 2 shows the classification system for the Kaoping River developed by TEPA. Basically, the upstream is classified as good water quality and mid to downstream is classified as slightly polluted quality. Thus, the highest degree of protection is given to the upstream section. The concentrations of some major water quality indicators (e.g., SS, NH3–N) are much higher than the Kaoping River water quality criteria (good or slightly polluted) developed by TEPA (TEPA, 2009). Among the four parameters, SS is one of the important factors and
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101 Table 1 The equation for RPI calculation and criteria for the four RPI classes. Items
DO (mg/L) BOD5 (mg/L) SS (mg/L) NH3–N (mg/L) Index scores (Si) Sub-index P Sub-index ¼ 14 4i¼1 Si
Ranks Nonpolluted
Slightly polluted
Moderate polluted
Gross polluted
>6.5 <3.0 <20 <0.5 1 <2
4.6–6.5 3.0–4.9 20–49 0.5–0.99 3 2.0–3.0
2.0–4.5 5.0–15.0 50–100 1.0–3.0 6 3.1–6.0
<2.0 >15 >100 >3.0 10 >6.0
Table 2 The classification system for the Kaoping River.
a
Water quality category
Class Aa upstream reacha
Class B mid to downstream reacha
DO (mg/L) BOD (mg/L) NH3–N (mg/L) Total phosphorus (mg/L) Suspended solid (mg/L)
P6.5 1 0.1 0.02
P5.5 2 0.3 0.05
<25
<25
91
WASP model was developed to obtain an immediate water quality evaluation in this study. In this study, integration of RPI into WASP was performed. Such linked model and index applications can enhance the accuracy and flexibility in describing the water quality conditions of Kaoping River, and are expected to produce better information about the river system, thus providing more confidence to decision-maker. The main objective of this study was to couple these models and use the combined function to simulate the effect of DO, SS, BOD, and NH3–N loadings from the watershed on the downstream river water quality. During the evaluation, water quality outputs from the WASP modeling were used for RPI calculation. To effectively obtain representative SS concentration for RPI calculation, relationship between SS and flow rate was evaluated. The major tasks included the following: (1) investigate and identify the current contributions of pollutants to the Kaoping River pollution, (2) develop SS estimation equation using correlations between SS and river flow rates, (3) develop a direct linkage between the RPI calculation and the WASP model containing SS equation, and (4) predict water column DO, SS, BOD, and NH3–N loading and evaluate their impacts on river water quality using the integrated modeling system. 2. Materials and methods
Illustrated in Fig. 1.
2.1. Water sampling and analysis
plays an important role in RPI value calculation in most of the large rivers in Taiwan. This is due to the fact that most of the watersheds in Taiwan have poor geologic formations. Most of the geologic structures contain sandstones, mudstones, shales, and conglomerates with easy-to-collapse features. Thus, sediment concentrations increase significantly after storms and heavy rains. The increased SS concentrations from the upstream section of the river after storms usually cause the deteriorations of water quality and also cause the increase in RPI value. However, the increase in RPI value could not be accurately estimated from pollution loading of most reaches due to the increased SS loads from the upper catchments. Thus, appropriate SS estimation method should be developed to effectively estimate SS concentrations after heavy rains and storms. The Water Quality Simulation Program (WASP) was selected for the water quality modeling. WASP is a surface water quality model developed by US Environmental Protection Agency (EPA) (Yang et al., 2007; Geza et al., 2009). WASP is capable of simulating the fate and transport of solutes in up to three dimensions in the steady-state and dynamic mode (Caruso, 2004; Xu et al., 2007; Fan et al., 2009; Lin et al., 2011). However, WASP model could not effectively simulate the SS loading in the river. Thus, the correlation between river flow rate and SS concentrations need to be evaluated to obtain representative RPI value when flow conditions are changed. Because SS is one of the parameters involving in the RPI calculation, it needs to be carefully evaluated. However, no river water quality models could effectively link with the RPI for efficient and immediate water quality evaluation. Furthermore, SS concentrations are significantly affected by river flow rates because the soil formation has an easily collapse feature in the upper catchments of most rivers in Taiwan. Currently, no river water quality models could effectively link with the RPI for efficient and immediate water quality evaluation and decision-making. Moreover, the commonly used WASP model could not be used for SS loading simulation and SS distribution during the river water quality modeling. Thus, a direct linkage between the RPI calculation and the
Fig. 1 presents the 12 water quality monitoring stations (W1 to W12) in the Kaoping River Basin. The W1 to W3, W4 to W6, and W7 to W9 were located in Lao-Non Creek, I-Liao Creek, and ChiSan Creek sub-Basins, respectively. The W10 to W12 were located in the downstream section of the Kaoping River. Water samples were collected monthly from 2009 to 2011. Flow velocity and flow rate data were collected from different flow monitoring stations located in the basin. Water samples were taken by the grab method developed by TEPA (NIEA, 2004). Samples were placed on ice until transferred to the appropriate sampling bottles and were kept refrigerated until analyzed. Water samples were sent for analyses within 24 h after sampling. Temperature, DO, and pH were analyzed immediately in the field. In the laboratory, water samples were analyzed for SS, NH3–N, and BOD. The laboratory analysis followed the Standard Methods for the examination of water and wastewater (APHA, 2005). Ion Chromatograph (Dionex) for inorganic nutrient and anion analyses, an Orion Ross pH meter for pH measurements, and an Orion DO meter (Model 840) for DO measurements. SS and BOD analyses were performed in accordance with the methods in Standard Methods (APHA, 2005). 2.2. Model description and application The WASP modeling employs the conservation of mass and momentum equations to determine the river hydraulic characteristics (e.g., depth, velocity, top width, flow rate, and cross-sectional area) (Ambrose et al., 2001; Yang et al., 2007; Kuo et al., 2006; Lai et al., 2011). The continuity equation (Eq. (1)) and momentum equation (Eq. (2)) used in the WASP model are as follows:
@Q 1 @Q þ ¼ qs @t B @x
ð1Þ
@Q @ Qw @z Q jQ j ¼0 þ þ 2 þ gA @t @x A @x K
ð2Þ
where z is the water surface elevation, Q the flow rate, B the wetted cross sectional width, A the wetted cross sectional area, t the time, x the distance along channel, K the conveyance of the channel, g the
92
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 2. Model grids of the Kaoping River Basin.
gravitational acceleration, and q is the side discharge per unit channel length. The water quality modeling starts from the EUTRO code of WASP, and the EUTRO model is developed to predict the water quality with respect to nutrients, plankton, DO, bacteria decay, and reactive pollutants in water column (Kima et al., 2004; Sehnert and Lindenschmidt, 2009). In this study, the model was used for simulating a 1-year period in 2011. Different interacting systems are developed comprising ammonia, nitrate, phosphate, phytoplankton, BOD, DO, organic nitrogen,
and organic phosphorus (Canu et al., 2004; Zhang et al., 2008). The nutrient interactions of the state variables are described as follows: Nitrogen cycle Organic nitrogen
V s3 ð1 fD7 ÞC 7 D
ð3Þ
T20 Sk1 ¼ k71 hT20 71 C 7 K 12 h12 C 1 Gpl C 4 P NH3 aNC
ð4Þ
Sk7 ¼ Dp1 C 4 aNC þ k71 hT20 71 C 7 Ammonia
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
93
Fig. 3. The simulation process using RPI and WASP.
Nitrate
Sk2 ¼
K 12 hT20 12 C 1
Gpl C 4 ð1 PNH3 ÞANC
ð5Þ
where aNC is the ratio nitrogen to carbon, K71 and K12 the nitrate and organic nitrogen denitrification rate at 20 °C, respectively, H12 and H71 the temperature coefficients, KNIT the half saturation constant for the oxygen limitation of nitrification, PNH3 the preference for ammonia uptake term, and Vs3 is the organic matter settling velocity. DO balance BOD
C6 v s 3ð1 fD 5Þ 5 C5 G5 Sk5 ¼ aoc K 1D C 4 K D hT20 D D 4 K BOD þ C 6 32 3 ðT20Þ C2 K 2D h2D K NO 14 K NO 3 þ C 6 ð6Þ DO
C6 Sk6 ¼ K 2 ðC s C 6 Þ K d hT20 C5 d K BOD þ C 6 64 C6 SOD ðT20Þ C1 K 12 hT20 h 12 14 D s K NIT þ C 6 32 48 14 32 þ ð1 PNH 3Þ C 4 k1R hT20 þ Gp 1 1R C 4 12 14 12 12
ð7Þ
where K1D, Kd, and K1R are the phytoplankton respiration rate, deoxygenation rate, and endogenous respiration rate at 20 °C, respectively, aoc the rate of oxygen to carbon, hd and h1R the temperature coefficients, Ka the reaction rate, Cs the saturation degree of DO, and KBOD is the half saturation constant for oxygen limitation (Canu et al., 2004). In the WASP modeling, the studied water system is segmented into a series of completely mixed water cells. In this study, the Kaoping River and its three major reaches were divided into 51 segments from upstream of each creek to the downstream section of Kaoping River. Each segment had a length of 2.2 km. Fig. 2 shows the model grids of the Kaoping River Basin. During the model simulation, the time step of 360 s was used. The WASP/EUTRO model was configured for mass transport calculation. The driving forces in
this model included upstream boundary conditions at segments 1, 19, and 29 and downstream boundary condition at segment 51, which were imposed for the model calibration (Fig. 2). 2.3. Linkage of SS sub-module and RPI equation in WASP model To obtain the correlations between Kaoping River flow rate and SS concentrations, water samples were collected from 12 sampling stations for SS measurement. Flow rate were monitored at water monitoring stations at each sampling event. Flow velocity and flow rate were measured following the methods described in NIEA (2004). To determine the correlation between the flow rate and SS concentrations, water sample collection and flow rate measurement were performed during the dry and wet seasons. The collected water quality and hydrological data were analyzed to evaluate the correlation between SS and flow rate (Cencic and Rechberger, 2008). The obtained correlation was used for the development of SS sub-model and was then linked to the WASP model for direct SS estimation. To obtain the immediate RPI value and determine the river water quality status, RPI index was embedded into the WASP model for direct RPI calculation. Fig. 3 presents the simulation process using RPI and WASP. Fig. 4 presents the process of the modeling procedure repeated for each segment until all segments met the requirements. The developed decision-making process could be used for the pollutant loading evaluation. The source code of the SS equation and RPI index package were embedded in the WASP coupling platform to improve the interactive transfer of water quality information to the models. In the simulation process, SS concentrations under different flow conditions were simulated using the developed SS and flow rate equations, and the results were stored in a data file in WASP model. This file was then retrieved for RPI calculation and other application. 3. Results and discussion 3.1. Field sampling and data analyses Tables 3 and 4 show the averaged results for DO, BOD, NH3–N, SS, and calculated RPI values for different water quality sampling
94
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
WASP modeling
RPI calculation If all segments are OK Judgment
End
If all segments are not OK Locate the operated segment
Data input and Modeling
Evaluation of the modeling results
Re-modeling
RPI recalculation
No Judgment
OK Fig. 4. The process of the modeling procedure repeated for each segment.
and hydrology measurement events in the dry (from November to April) and wet (from May to October) seasons, respectively, during the investigation period from 2009 to 2011. Results of the hydrology investigation show that the averaged flow rates at all monitoring stations were significantly increased in wet seasons (up to 53.4 m3/s in Kaoping River). Results show that the averaged NH3–N concentrations varied from 0.04 to 0.11 mg/L in three reaches in dry seasons. However, the averaged NH3–N concentrations varied from 0.19 to 0.28 mg/L in three reaches in wet seasons. Results also show that the SS concentrations increased from below
71 mg/L (varied from 17 to 71 mg/L) to up to 3,475 mg/L (varied from 2318 to 3475 mg/L) in water samples collected in dry and wet seasons, respectively. This indicates that the loads of NPS NH3–N and SS pollution to the Kaoping River would be also increased during the wet seasons (up to 3475 mg/L in average in Lao-Non Creek). Furthermore, the soil erosion mostly occurred in the wet seasons caused significant high SS concentrations. Results show that lower BOD and NH3–N concentrations were observed at sampling stations located in upper catchments compared to the concentrations observed at sampling stations located
95
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101 Table 3 Average concentrations of water sampling events performed in the dry seasons. Sampling location
Chi-San Creek
Lao-Non Creek
I-Liao Creek
Kaoping River
DO (mg/L) NH3–N (mg/L) SS (mg/L) BOD (mg/L) RPI Water quality Flow rate (m3/s)
4.2 0.11 17 2.5 2.25 Slightly polluted 7.2
5.2 0.02 36 1.6 2 Slightly polluted 12.8
4.4 0.04 21 1.5 2.75 Slightly polluted 3.1
3.7 0.75 71 5.8 5.25 Moderate polluted 29.1
Table 4 Average concentrations of water sampling events performed in the wet seasons.
Table 5 Parameters used in the WASP modeling process.
Sampling location
Chi-San Creek
Lao-Non Creek
I-Liao Creek
Kaoping River
Description
Parameter
Range
Fixed or estimated by calibration
DO (mg/L) NH3–N (mg/L) SS (mg/L) BOD (mg/L) RPI Water quality
7.3 0.28 2867 1.1 3.25 Moderate polluted 17.1
7.5 0.19 2618 0.6 3.25 Moderate polluted 25.6
7.1 0.26 3475 0.1 3.25 Moderate polluted 10.1
6.4 0.34 2384 2.4 3.75 Moderate polluted 53.4
Nitrification Nitrification rate constant
K12C
0.06
KNIT
0.05– 0.15 1.08– 1.20 –
K20C K20T
– –
0.03 1.04
KNO3
–
0.01
K1C
1.4– 2.6 0.98– 1.072 0.05– 0.35 1.045– 1.1 0.02– 0.1 – – – –
2.01
0.07
KDT
0.05– 0.3 –
KBOD
–
0.3
K71C
0.02– 0.2 1.02– 1.3
0.03
0.01– 0.4 1.045– 1.2
0.03
0.25– 0.5 –
0.3
3
Flow rate (m /s)
Nitrification rate temperature constant Half saturation constant for nitrification Denitrification Denitrification rate constant Denitrification rate temperature constant Half saturation constant for denitrification Phytoplankton Phytoplankton growth rate constant Phytoplankton growth rate temperature constant Algal respiration rate constant Algal respiration rate temperature constant Phytoplankton death rate constant Zooplankton grazing rate Oxygen to carbon rate Phosphorus to carbon rate Nitrogen to carbon rate BOD Oxidation of BOD rate constant Oxidation of BOD rate temperature constant BOD half-saturation constant
Fig. 5. Correlations between SS concentrations and flow rate measurements in water samples [Low flow rate (flow rate: 0–50 m3/s) Y = 0.028X 0.37; High flow rate (flow rate: 50–650 m3/s) Y = 0.13X 21.36, Y = flow rate (m3/s or CMS), X = SS concentration (mg/L)].
in lower catchments. The averaged BOD concentrations varied from 1.5 to 2.5 mg/L in three reaches compared to 5.8 mg/L in Kaoping River in dry seasons. In wet seasons, the averaged BOD concentrations varied from 0.1 to 1.1 mg/L in three reaches compared to 2.4 mg/L in Kaoping River. This was due to the fact that most of the land use patterns were mainly forest and agricultural uses, which produced had less pollution levels compared to land use patterns of human activities in the lower catchments. Because the lower catchments have been developed to domestic and industrial areas (human activity categories), higher BOD and NH3–N concentrations were detected at sampling stations located in this region (W10 to W12). The variability of pollutant concentrations is usually affected by two different forcing functions including the hydrodynamic flow rate of the watershed and the impacts of
Organic nitrogen Mineralization of dissolved ON rate constant Mineralization of dissolved ON rate temperature constant Organic phosphorus Mineralization of dissolved OP rate constant Mineralization of dissolved OP rate temperature constant
K12T
K1T K1RC K1RT K1D K1G OCRB PCRB NCRB KDC
K71T
K83C K83T
Fraction of dead and respired phytoplankton Fraction of ON from algal death FON Fraction of OP from algal death
FOP
1.04 0.01
1.066 0.008 1.08 0.11 0.80 2.67 0.028 0.200
1.040
1.04
1.04
0.3
ON: organic nitrogen; OP: organic phosphorus.
human activities. Due to the variability of the two forcing functions, variable BOD and NH3–N concentrations were observed. Higher DO values were observed in the upper catchments due to the natural turbulence (natural reaeration) in three reaches (varied from 7.1 to 7.5 mg/L in wet seasons), which enhanced the oxy-
96
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 6. Comparison of time-variable simulated results and observed BOD at water monitoring stations.
gen mass transfer between the river and atmosphere. This caused higher DO concentrations in the river. Moreover, lower DO concen-
trations in downstream sections (3.7 mg/L in Kaoping River) during the dry seasons were mainly due to the higher pollution loading in
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
97
Fig. 7. Comparison of time-variable simulated results and observed NH3–N at water monitoring stations.
lower catchments and lower oxygen solubility in the dry seasons. Results from the RPI evaluation indicate that the water quality changed from slightly polluted in dry seasons to moderate polluted
in wet seasons in three reaches. Results from Tables 3 and 4 reveal that SS was an important parameter among the four parameters in RPI calculation. Results indicate that without the dilution of high
98
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 8. Comparison of time-variable simulated results and observed DO at water monitoring stations.
flow of river water and soil erosion in dry seasons, low DO and high BOD values became the determinant factors in the RPI calculation.
Fig. 5 presents the correlation between the measured SS and flow rates. Results indicate that the SS and flow rate values of water samples collected from the three major reaches and main
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Kaoping River had a significant correlation [R2 values of 0.9 and 0.87 (Eq. (8) and (9))]. Thus, SS concentrations in river water can be calculated using the following equation:
99
value could then be calculated for water quality evaluation. The equations were embedded into the integrated WASP modeling for SS simulation to obtain more representative SS data and RPI values.
Low flow rate ðflow rate : 0—50 m3 =sÞ Y ¼ 0:028X 0:37 R2 ¼ 0:90
ð8Þ
High flow rate ðflow rate : 50—650 m3 =sÞ Y ¼ 0:13X 21:36 R2 ¼ 0:87
ð9Þ
where Y = flow rate (m3/s) and X = SS concentration (mg/L). In the Kaoping River Basin, most of the SS concentrations were caused by the soil erosion of the upper catchment during the wet seasons. Thus, the obtained empirical equations could be used for immediate SS calculation when flows have significant changes. The RPI
3.2. Water quality modeling In this study, the WASP modeling was performed for time-variable river water quality simulation. Table 5 shows the parameters and input values used in the WASP modeling. Figs. 6–8 present the comparisons of observed BOD, NH3–N, and DO data and time-variable simulated results at different water sampling stations in 2011. The WASP modeling performed a good match compared to the observed data. Results suggest that the monthly BOD and NH3–N loadings were affected by the flows and agricultural activities.
Fig. 9. Comparison of time-variable simulated results and observed SS at water monitoring stations.
100
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
In this study, the SS equation assisted WASP modeling was applied for time-variable river water quality simulation. The WASP model was then used for RPI simulation. Fig. 9 presents the comparisons of observed SS data and time-variable simulated results at different water sampling stations in 2011. Results indicate that the SS equation assisted modeling had reasonable matches with the observed data compared to the simulated SS results using the empirical flows derived SS equations. Results indicate that the SS equation assisted WASP modeling was able to include the influence of soil erosion on the increase in SS concentrations. Thus, the modified WASP model provided a good tool to effectively simulate the impacts of SS pollution on river water quality.
3.3. RPI calculation In this study, RPI equation was embedded in the WASP model for RPI calculation and water quality evaluation. Fig. 10 presents the calculated RPI values using WASP model containing RPI module (equation described in Table 1). In Fig. 10, RPI values for each monitoring station were simulated using the simulated water quality results of BOD, DO, NH3–N, and SS values for the whole year. Results show that in the dry seasons, SS played an important role in RPI calculation. SS was a critical factor during the RPI calculation especially for the upper catchment. Because of the discharge of domestic and industrial wastewaters into the river, higher BOD
Fig. 10. Comparison of time-variable simulated results and observed RPI at water monitoring stations.
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
and NH3–N concentrations were observed in the downstream sections of the river. Thus, BOD and NH3–N became the determinant factors in the downstream section of the Kaoping River for the RPI calculation. In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Higher river flow rates also caused the turbulence of the flow and increase in DO concentrations in the upper sections of the river. Therefore, all four parameters (BOD, DO, NH3–N, and SS) would affect the RPI calculation in the downstream sections. 4. Conclusions In this study, SS estimation equation and RPI package were embedded into the WASP model to assist the river water quality simulation of the Kaoping River. A direct linkage between the RPI calculation and the WASP model was developed to obtain an immediate water quality evaluation. The major findings of this study include the following: (1) Water quality monitoring results indicate that SS played an important role in RPI calculation and SS was a critical factor during the RPI calculation especially for the upper catchment in wet seasons. This was due to the fact that the soil erosion caused the increase in the SS concentrations after storms. (2) In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Higher river flow rates also caused the turbulence of the flow and increase in DO concentrations in the upper section of the river. Thus, NH3–N and DO affected the RPI value in the upper sections of the river. (3) SS concentrations were highly correlated with the flow rates of the Kaoping River. The obtained SS and flow rate equations were as follows: Y = 0.028X 0.37 for low flow rate (flow rate: 0–50 m3/s), Y = 0.13X 21.36 for high flow rate (flow rate: 50–650 m3/s) (Y = flow rate, m3/s; X = SS concentration, mg/L). The equations were embedded in the integrated WASP modeling for SS simulation to obtain more representative SS data and RPI values (4) During the evaluation, water quality outputs from the WASP modeling were used for RPI calculation. Thus, an integral decision-making system combining SS estimation equation, RPI package, and water quality model for river water quality evaluation was developed. (5) This study demonstrates that the introduction of SS estimation equation and RPI package into WASP simulation has been shown to be a significant advance in estimating water quality in Kaoping River. The developed decision-making modeling concept could be easily adopted to other similar rivers. Thus, WASP modeling assisted RPI estimation provides a good tool to effectively simulate the impacts of pollution on river water quality. The integrated system will be useful in developing appropriate water quality management strategies for the improvement of river water quality, and this makes it easier for decision-makers to evaluate alternative water quality management plans.
Acknowledgements This study was funded by Taiwan National Science Council and Taiwan Environmental Protection Administration (TEPA). Addi-
101
tional thanks to the personnel of TEPA for their assistance throughout this project. References Ambrose, B., Wool, T.A., Martin, J.L., 2001. The Water Quality Analysis Simulation Program, WASP6, User Manual, US EPA, Athens, GA. APHA (American Public Health Association), 2005. Standard Methods for the Examination of Water and Waste Water, 22th ed. APHA-AWWA-WPCF, Washington, DC. Canu, D.M., Solidoro, C., Umgiesser, G., 2004. Erratum to modeling the responses of the Lagoon of Venice ecosystem to variations in physical forcing. Ecol. Model. 175, 197–216. Caruso, B.S., 2004. Modeling metals transport and sediment/water interactions in a mining impacted mountain stream. Water Resources Assoc. 40, 1603–1615. Cencic, O., Rechberger, H., 2008. Material flow analysis with software STAN. Sustain. Environ. Res. 18, 3–7. Chen, S.W., Chao, A.C., Chen, T.Y., Kao, C.M., Lai, Y.C., Lin, C.E., 2006. Application of water quality modeling on river basin management. WSEAS Trans. Math. 5, 1078–1083. Fan, C., Ko, C.H., Wang, W.S., 2009. An innovated modeling approach using Qual2K and HEC-RAS integration to assess the impact of tidal effect on river water quality simulation. J. Environ. Manage. 90, 1824–1832. Geza, M., Poeter, E.P., McCray, J.E., 2009. Quantifying predictive uncertainty for a mountain watershed model. J. Hydrol. 376, 170–181. Kavvas, M.L., Yoon, J., Chen, Z.Q., Liang, L., Dogrul, E.C., Ohara, N., Aksoy, H., 2006. Watershed environmental hydrology model: environmental module and its application to a California watershed. J. Hydrol. Eng. 11, 450–464. Kima, D., Wanga, Q., Soriala, G.A., Dionysioua, D.D., Timberlake, D., 2004. A model approach for evaluating effects of remedial actions on mercury speciation and transport in a lake system. Sci. Total Environ. 327, 1–15. Kuo, J.T., Lung, W.S., Yang, C.P., Liu, W.C., Yang, M.D., Tang, T.S., 2006. Eutrophication modelling of reservoirs in Taiwan. Environ. Model. Softw. 21, 829–844. Lai, Y.C., 2010. Development of Kaoping River Management Strategies. Ph.D. Dissertation, National Sun Yat-Sen University, Kaohsiung, Taiwan. Lai, Y.C., Yang, C.P., Hsieh, C.Y., Wu, C.Y., Kao, C.M., 2011. Evaluation of non-point source pollution and river water quality using a multimedia two-model system, Taiwan. J. Hydrol. 409, 583–595. Lin, C.E., 2010. Development of River Water Quality and Sediment Management Strategies. Ph.D. Dissertation, National Sun Yat-Sen University, Kaohsiung, Taiwan. Lin, C.E., Kao, C.M., Jou, C.J., 2010. Preliminary identification of watershed management strategies for the Houjing River in Taiwan. Water Sci. Technol. 62, 1667–1675. Lin, C.E., Chen, C.T., Kao, C.M., Hong, A., Wu, C.Y., 2011. Development of the sediment and water quality management strategies for the Salt-water River, Taiwan. Mar. Pollut. Bullet. 63, 528–534. Luo, H., Li, M., Xu, R., Fu, X., Huang, G., Huang, X., 2012. Pollution characteristics of urban surface runoff in a street community. Sustain. Environ. Res. 22, 61–68. Markku, P., Eila, T., Manna, K., Jari, K., Jarmo, L., Rittu, N., Sirkka, T., 2010. YIHMA-A tool for allocation of measures to control erosion and nutrient loading from Finnish agricultural catchment. Ecosyst. Environ. 138, 306–317. NIEA (National Institute of Environmental Analysis, Taiwan EPA), 2004. Method of River Flow Rate Analysis, NIEA W022.51C. Ouyang, W., Wang, X., Hao, F., Srinivasan, R., 2009. Temporal-spatial dynamics of vegetation variation on non-point source nutrient pollution. Ecol. Model. 220, 2702–2713. Papanicolaou, T., Abaci, O., 2005. An integrated watershed model. In: Managing Watersheds for Human and Natural Impacts: Engineering, Ecological, and Economic Challenges, Managing Watersheds for Human and Natural Impacts: Engineering, Ecological, and Economic Challenges – Proceedings of the 2005 Watershed Management Conference, p. 189. Sehnert, C., Lindenschmidt, K.E., 2009. The effect of upstream river hydrograph characteristics on zinc deposition, dissolved oxygen depletion and phytoplankton growth in a flood detention basin system. Quatern. Int. 208, 158–168. TEPA, Taiwan Environmental Protection Administration, 2002. Development of Non-Point Source Pollutant Remedial Strategy, Taipei, Taiwan. TEPA, Taiwan Environmental Protection Administration, 2009. Project for Aerial Photo Interpretation and Analysis of Water Pollution Sources in Kaoping River, Annual Report, Taipei, Taiwan. Xu, Z., Godrej, A.N., Grizzard, T.J., 2007. The hydrological calibration and validation of a complexly-linked watershed–reservoir model for the Occoquan watershed, Virginia. J. Hydrol. 345, 167–183. Xu, K., Milliman, J.D., Xu, H., 2010. Temporal trend of precipitation and runoff in major Chinese River since 1951. Glob. Planet. Change 73, 219–232. Yang, C.P., Kuo, J.T., Lung, W.S., Lai, J.S., Wu, J.T., 2007. Water quality and ecosystem modeling of tidal wetlands. J. Environ. Eng. 133, 711–721. Zhang, M.L., Shen, Y.M., Guo, Y., 2008. Development and application of a eutrophication water quality model for river networks. J. Hydrodyn. 20, 719– 726.
Contents lists available at SciVerse ScienceDirect
Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol
Development of a water quality modeling system for river pollution index and suspended solid loading evaluation Y.C. Lai a, Y.T. Tu a, C.P. Yang b, R.Y. Surampalli c, C.M. Kao a,⇑ a
Institute of Environmental Engineering, National Sun Yat-Sen University, Kaohsiung 804, Taiwan Center for Teaching Excellence, National Pingtung University of Science & Technology, Pingtung County, Taiwan c Department of Civil Engineering, University of Nebraska, Lincoln, NE, USA b
a r t i c l e
i n f o
Article history: Received 13 August 2012 Received in revised form 13 November 2012 Accepted 21 November 2012 Available online 1 December 2012 This manuscript was handled by Peter Laurent Charlet, Editor-in-Chief, with the assistance of Hossein Ghadiri, Associate Editor Keywords: Ammonia nitrogen (NH3–N) River Pollution Index (RPI) River water quality Suspended solid (SS) Watershed management
s u m m a r y The Kaoping River Basin is the largest and most extensively used watershed in Taiwan. In the upper catchment, the non-point source (NPS) pollutants including suspended solid (SS) and ammonia nitrogen (NH3–N) are two major water pollutants causing the deterioration of Kaoping River water quality. Because SS is one of the four parameters involving in the River Pollution Index (RPI) calculation, it needs to be carefully evaluated to obtain the representative water quality index. The main objective of this study was to develop a water quality modeling system to obtain representative SS and RPI values for water quality evaluation. In this study, a direct linkage between the RPI calculation and a water quality model [Water Quality Analysis Simulation Program (WASP)] has been developed. Correlation equations between Kaoping River flow rates and SS concentrations were developed using the field data collected during the high and low flows of the Kaoping River. Investigation results show that the SS concentrations were highly correlated with the flow rates. The obtained SS equation and RPI calculation package were embedded into the WASP model to improve interactive transfers of required data for water quality modeling and RPI calculation. Results indicate that SS played an important role in RPI calculation and SS was a critical factor during the RPI calculation especially for the upper catchment in the wet seasons. This was due to the fact that the soil erosion caused the increase in the SS concentrations after storms. In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Results demonstrate that the integral approach could develop a direct linkage among river flow rate, water quality, and pollution index. The introduction of the integrated system showed a significant advance in water quality evaluation and river management strategy development. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction Compared to point source pollution, nonpoint-source (NPS) pollution is more diffuse and harder to isolate and control (Markku et al., 2010). The NPS pollution is defined legally as any pollution not originating from a statutory point source, which can be carried over by rainfall and irrigation water to enter a water body dispersedly via surface runoff. It has been well-documented that NPS pollution, such as nutrient runoff and atmospheric deposition, contribute significant pollutant loading to water bodies. The NPS pollutants can cause the deterioration of water quality through the release of suspended solids (SSs), nutrients, pesticides, fertilizer, and other sources of inorganic and organic matter. Because many factors, such as topography, soil characteristics, and rainfall intensity, affect their quantity and quality, variety types of NPS simulation models have been used to account for ⇑ Corresponding author. Tel.: +886 7 5254413; fax: +886 7 5254449. E-mail address: [email protected] (C.M. Kao). 0022-1694/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhydrol.2012.11.050
the integrated effects of the hydrological cycle and land cover in relation to nutrient yield (Kavvas et al., 2006; Papanicolaou and Abaci, 2005; Xu et al., 2010). With the aid of various environmental models, the improvement of the estimation and control of NPS pollution in watershed has been enhanced greatly in recent years. To assess the effectiveness of land-use-management policy, some applications seek improved modeling approaches for predicting the water quality effects of storm events as a function of land topography, land cover, and land use leading to the development of various applications (Chen et al., 2006; Ouyang et al., 2009). These models, for assessing NPS loads in the agricultural field, generally simulate rainfall, erosion, runoff sediment, temperature, wind speed, atmosphere pressure, and NPS processes (Lin et al., 2010; Luo et al., 2012). The Kaoping River Basin, located in the southeast region of Taiwan, is the largest and the most intensively utilized river basin in Taiwan. It is 171 km in length, drains a catchment of more than 3625 km2, and has a mean flow of 239 m3/s. Fig. 1 shows the location of Kaoping River Basin, Kaoping River, its catchment, and three
90
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 1. Kaoping River Basin and three major reaches.
major reaches (Chi-San Creek, I-Liao Creek, and Lao-Non Creek). Although the mean annual rainfall in this river basin is close to 3000 mm, over 90% appears in the wet season. The period of high flow rate in the stream usually occurs in the late spring and summer due to the impacts of monsoon and typhoon (Lin, 2010; Lai, 2010). NPS SS pollutants due to the severe storm events would cause significant adverse impacts on the river water quality of Kaoping River, and also cause the increased turbidity of the river water at the intake location of the downstream water treatment plant. Thus, the NPS SS pollution should be effectively evaluated and controlled. Taiwan Environmental Protection Administration (TEPA) has developed a River Pollution Index (RPI) classification system for river water quality evaluation based on the purpose of water usage and degree of protection for each stream section (TEPA,
2002). The RPI involves four parameters: dissolved oxygen (DO), biochemical oxygen demand (BOD), SS, and ammonia nitrogen (NH3–N), each of which is ultimately converted to a four-state quality sub-index (1, 3, 6, and 10). The overall index is then divided into four pollution levels. Table 1 presents the equation for RPI calculation and criteria for the four RPI classes (good, slightly polluted, moderate polluted, and gross polluted). Table 2 shows the classification system for the Kaoping River developed by TEPA. Basically, the upstream is classified as good water quality and mid to downstream is classified as slightly polluted quality. Thus, the highest degree of protection is given to the upstream section. The concentrations of some major water quality indicators (e.g., SS, NH3–N) are much higher than the Kaoping River water quality criteria (good or slightly polluted) developed by TEPA (TEPA, 2009). Among the four parameters, SS is one of the important factors and
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101 Table 1 The equation for RPI calculation and criteria for the four RPI classes. Items
DO (mg/L) BOD5 (mg/L) SS (mg/L) NH3–N (mg/L) Index scores (Si) Sub-index P Sub-index ¼ 14 4i¼1 Si
Ranks Nonpolluted
Slightly polluted
Moderate polluted
Gross polluted
>6.5 <3.0 <20 <0.5 1 <2
4.6–6.5 3.0–4.9 20–49 0.5–0.99 3 2.0–3.0
2.0–4.5 5.0–15.0 50–100 1.0–3.0 6 3.1–6.0
<2.0 >15 >100 >3.0 10 >6.0
Table 2 The classification system for the Kaoping River.
a
Water quality category
Class Aa upstream reacha
Class B mid to downstream reacha
DO (mg/L) BOD (mg/L) NH3–N (mg/L) Total phosphorus (mg/L) Suspended solid (mg/L)
P6.5 1 0.1 0.02
P5.5 2 0.3 0.05
<25
<25
91
WASP model was developed to obtain an immediate water quality evaluation in this study. In this study, integration of RPI into WASP was performed. Such linked model and index applications can enhance the accuracy and flexibility in describing the water quality conditions of Kaoping River, and are expected to produce better information about the river system, thus providing more confidence to decision-maker. The main objective of this study was to couple these models and use the combined function to simulate the effect of DO, SS, BOD, and NH3–N loadings from the watershed on the downstream river water quality. During the evaluation, water quality outputs from the WASP modeling were used for RPI calculation. To effectively obtain representative SS concentration for RPI calculation, relationship between SS and flow rate was evaluated. The major tasks included the following: (1) investigate and identify the current contributions of pollutants to the Kaoping River pollution, (2) develop SS estimation equation using correlations between SS and river flow rates, (3) develop a direct linkage between the RPI calculation and the WASP model containing SS equation, and (4) predict water column DO, SS, BOD, and NH3–N loading and evaluate their impacts on river water quality using the integrated modeling system. 2. Materials and methods
Illustrated in Fig. 1.
2.1. Water sampling and analysis
plays an important role in RPI value calculation in most of the large rivers in Taiwan. This is due to the fact that most of the watersheds in Taiwan have poor geologic formations. Most of the geologic structures contain sandstones, mudstones, shales, and conglomerates with easy-to-collapse features. Thus, sediment concentrations increase significantly after storms and heavy rains. The increased SS concentrations from the upstream section of the river after storms usually cause the deteriorations of water quality and also cause the increase in RPI value. However, the increase in RPI value could not be accurately estimated from pollution loading of most reaches due to the increased SS loads from the upper catchments. Thus, appropriate SS estimation method should be developed to effectively estimate SS concentrations after heavy rains and storms. The Water Quality Simulation Program (WASP) was selected for the water quality modeling. WASP is a surface water quality model developed by US Environmental Protection Agency (EPA) (Yang et al., 2007; Geza et al., 2009). WASP is capable of simulating the fate and transport of solutes in up to three dimensions in the steady-state and dynamic mode (Caruso, 2004; Xu et al., 2007; Fan et al., 2009; Lin et al., 2011). However, WASP model could not effectively simulate the SS loading in the river. Thus, the correlation between river flow rate and SS concentrations need to be evaluated to obtain representative RPI value when flow conditions are changed. Because SS is one of the parameters involving in the RPI calculation, it needs to be carefully evaluated. However, no river water quality models could effectively link with the RPI for efficient and immediate water quality evaluation. Furthermore, SS concentrations are significantly affected by river flow rates because the soil formation has an easily collapse feature in the upper catchments of most rivers in Taiwan. Currently, no river water quality models could effectively link with the RPI for efficient and immediate water quality evaluation and decision-making. Moreover, the commonly used WASP model could not be used for SS loading simulation and SS distribution during the river water quality modeling. Thus, a direct linkage between the RPI calculation and the
Fig. 1 presents the 12 water quality monitoring stations (W1 to W12) in the Kaoping River Basin. The W1 to W3, W4 to W6, and W7 to W9 were located in Lao-Non Creek, I-Liao Creek, and ChiSan Creek sub-Basins, respectively. The W10 to W12 were located in the downstream section of the Kaoping River. Water samples were collected monthly from 2009 to 2011. Flow velocity and flow rate data were collected from different flow monitoring stations located in the basin. Water samples were taken by the grab method developed by TEPA (NIEA, 2004). Samples were placed on ice until transferred to the appropriate sampling bottles and were kept refrigerated until analyzed. Water samples were sent for analyses within 24 h after sampling. Temperature, DO, and pH were analyzed immediately in the field. In the laboratory, water samples were analyzed for SS, NH3–N, and BOD. The laboratory analysis followed the Standard Methods for the examination of water and wastewater (APHA, 2005). Ion Chromatograph (Dionex) for inorganic nutrient and anion analyses, an Orion Ross pH meter for pH measurements, and an Orion DO meter (Model 840) for DO measurements. SS and BOD analyses were performed in accordance with the methods in Standard Methods (APHA, 2005). 2.2. Model description and application The WASP modeling employs the conservation of mass and momentum equations to determine the river hydraulic characteristics (e.g., depth, velocity, top width, flow rate, and cross-sectional area) (Ambrose et al., 2001; Yang et al., 2007; Kuo et al., 2006; Lai et al., 2011). The continuity equation (Eq. (1)) and momentum equation (Eq. (2)) used in the WASP model are as follows:
@Q 1 @Q þ ¼ qs @t B @x
ð1Þ
@Q @ Qw @z Q jQ j ¼0 þ þ 2 þ gA @t @x A @x K
ð2Þ
where z is the water surface elevation, Q the flow rate, B the wetted cross sectional width, A the wetted cross sectional area, t the time, x the distance along channel, K the conveyance of the channel, g the
92
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 2. Model grids of the Kaoping River Basin.
gravitational acceleration, and q is the side discharge per unit channel length. The water quality modeling starts from the EUTRO code of WASP, and the EUTRO model is developed to predict the water quality with respect to nutrients, plankton, DO, bacteria decay, and reactive pollutants in water column (Kima et al., 2004; Sehnert and Lindenschmidt, 2009). In this study, the model was used for simulating a 1-year period in 2011. Different interacting systems are developed comprising ammonia, nitrate, phosphate, phytoplankton, BOD, DO, organic nitrogen,
and organic phosphorus (Canu et al., 2004; Zhang et al., 2008). The nutrient interactions of the state variables are described as follows: Nitrogen cycle Organic nitrogen
V s3 ð1 fD7 ÞC 7 D
ð3Þ
T20 Sk1 ¼ k71 hT20 71 C 7 K 12 h12 C 1 Gpl C 4 P NH3 aNC
ð4Þ
Sk7 ¼ Dp1 C 4 aNC þ k71 hT20 71 C 7 Ammonia
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
93
Fig. 3. The simulation process using RPI and WASP.
Nitrate
Sk2 ¼
K 12 hT20 12 C 1
Gpl C 4 ð1 PNH3 ÞANC
ð5Þ
where aNC is the ratio nitrogen to carbon, K71 and K12 the nitrate and organic nitrogen denitrification rate at 20 °C, respectively, H12 and H71 the temperature coefficients, KNIT the half saturation constant for the oxygen limitation of nitrification, PNH3 the preference for ammonia uptake term, and Vs3 is the organic matter settling velocity. DO balance BOD
C6 v s 3ð1 fD 5Þ 5 C5 G5 Sk5 ¼ aoc K 1D C 4 K D hT20 D D 4 K BOD þ C 6 32 3 ðT20Þ C2 K 2D h2D K NO 14 K NO 3 þ C 6 ð6Þ DO
C6 Sk6 ¼ K 2 ðC s C 6 Þ K d hT20 C5 d K BOD þ C 6 64 C6 SOD ðT20Þ C1 K 12 hT20 h 12 14 D s K NIT þ C 6 32 48 14 32 þ ð1 PNH 3Þ C 4 k1R hT20 þ Gp 1 1R C 4 12 14 12 12
ð7Þ
where K1D, Kd, and K1R are the phytoplankton respiration rate, deoxygenation rate, and endogenous respiration rate at 20 °C, respectively, aoc the rate of oxygen to carbon, hd and h1R the temperature coefficients, Ka the reaction rate, Cs the saturation degree of DO, and KBOD is the half saturation constant for oxygen limitation (Canu et al., 2004). In the WASP modeling, the studied water system is segmented into a series of completely mixed water cells. In this study, the Kaoping River and its three major reaches were divided into 51 segments from upstream of each creek to the downstream section of Kaoping River. Each segment had a length of 2.2 km. Fig. 2 shows the model grids of the Kaoping River Basin. During the model simulation, the time step of 360 s was used. The WASP/EUTRO model was configured for mass transport calculation. The driving forces in
this model included upstream boundary conditions at segments 1, 19, and 29 and downstream boundary condition at segment 51, which were imposed for the model calibration (Fig. 2). 2.3. Linkage of SS sub-module and RPI equation in WASP model To obtain the correlations between Kaoping River flow rate and SS concentrations, water samples were collected from 12 sampling stations for SS measurement. Flow rate were monitored at water monitoring stations at each sampling event. Flow velocity and flow rate were measured following the methods described in NIEA (2004). To determine the correlation between the flow rate and SS concentrations, water sample collection and flow rate measurement were performed during the dry and wet seasons. The collected water quality and hydrological data were analyzed to evaluate the correlation between SS and flow rate (Cencic and Rechberger, 2008). The obtained correlation was used for the development of SS sub-model and was then linked to the WASP model for direct SS estimation. To obtain the immediate RPI value and determine the river water quality status, RPI index was embedded into the WASP model for direct RPI calculation. Fig. 3 presents the simulation process using RPI and WASP. Fig. 4 presents the process of the modeling procedure repeated for each segment until all segments met the requirements. The developed decision-making process could be used for the pollutant loading evaluation. The source code of the SS equation and RPI index package were embedded in the WASP coupling platform to improve the interactive transfer of water quality information to the models. In the simulation process, SS concentrations under different flow conditions were simulated using the developed SS and flow rate equations, and the results were stored in a data file in WASP model. This file was then retrieved for RPI calculation and other application. 3. Results and discussion 3.1. Field sampling and data analyses Tables 3 and 4 show the averaged results for DO, BOD, NH3–N, SS, and calculated RPI values for different water quality sampling
94
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
WASP modeling
RPI calculation If all segments are OK Judgment
End
If all segments are not OK Locate the operated segment
Data input and Modeling
Evaluation of the modeling results
Re-modeling
RPI recalculation
No Judgment
OK Fig. 4. The process of the modeling procedure repeated for each segment.
and hydrology measurement events in the dry (from November to April) and wet (from May to October) seasons, respectively, during the investigation period from 2009 to 2011. Results of the hydrology investigation show that the averaged flow rates at all monitoring stations were significantly increased in wet seasons (up to 53.4 m3/s in Kaoping River). Results show that the averaged NH3–N concentrations varied from 0.04 to 0.11 mg/L in three reaches in dry seasons. However, the averaged NH3–N concentrations varied from 0.19 to 0.28 mg/L in three reaches in wet seasons. Results also show that the SS concentrations increased from below
71 mg/L (varied from 17 to 71 mg/L) to up to 3,475 mg/L (varied from 2318 to 3475 mg/L) in water samples collected in dry and wet seasons, respectively. This indicates that the loads of NPS NH3–N and SS pollution to the Kaoping River would be also increased during the wet seasons (up to 3475 mg/L in average in Lao-Non Creek). Furthermore, the soil erosion mostly occurred in the wet seasons caused significant high SS concentrations. Results show that lower BOD and NH3–N concentrations were observed at sampling stations located in upper catchments compared to the concentrations observed at sampling stations located
95
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101 Table 3 Average concentrations of water sampling events performed in the dry seasons. Sampling location
Chi-San Creek
Lao-Non Creek
I-Liao Creek
Kaoping River
DO (mg/L) NH3–N (mg/L) SS (mg/L) BOD (mg/L) RPI Water quality Flow rate (m3/s)
4.2 0.11 17 2.5 2.25 Slightly polluted 7.2
5.2 0.02 36 1.6 2 Slightly polluted 12.8
4.4 0.04 21 1.5 2.75 Slightly polluted 3.1
3.7 0.75 71 5.8 5.25 Moderate polluted 29.1
Table 4 Average concentrations of water sampling events performed in the wet seasons.
Table 5 Parameters used in the WASP modeling process.
Sampling location
Chi-San Creek
Lao-Non Creek
I-Liao Creek
Kaoping River
Description
Parameter
Range
Fixed or estimated by calibration
DO (mg/L) NH3–N (mg/L) SS (mg/L) BOD (mg/L) RPI Water quality
7.3 0.28 2867 1.1 3.25 Moderate polluted 17.1
7.5 0.19 2618 0.6 3.25 Moderate polluted 25.6
7.1 0.26 3475 0.1 3.25 Moderate polluted 10.1
6.4 0.34 2384 2.4 3.75 Moderate polluted 53.4
Nitrification Nitrification rate constant
K12C
0.06
KNIT
0.05– 0.15 1.08– 1.20 –
K20C K20T
– –
0.03 1.04
KNO3
–
0.01
K1C
1.4– 2.6 0.98– 1.072 0.05– 0.35 1.045– 1.1 0.02– 0.1 – – – –
2.01
0.07
KDT
0.05– 0.3 –
KBOD
–
0.3
K71C
0.02– 0.2 1.02– 1.3
0.03
0.01– 0.4 1.045– 1.2
0.03
0.25– 0.5 –
0.3
3
Flow rate (m /s)
Nitrification rate temperature constant Half saturation constant for nitrification Denitrification Denitrification rate constant Denitrification rate temperature constant Half saturation constant for denitrification Phytoplankton Phytoplankton growth rate constant Phytoplankton growth rate temperature constant Algal respiration rate constant Algal respiration rate temperature constant Phytoplankton death rate constant Zooplankton grazing rate Oxygen to carbon rate Phosphorus to carbon rate Nitrogen to carbon rate BOD Oxidation of BOD rate constant Oxidation of BOD rate temperature constant BOD half-saturation constant
Fig. 5. Correlations between SS concentrations and flow rate measurements in water samples [Low flow rate (flow rate: 0–50 m3/s) Y = 0.028X 0.37; High flow rate (flow rate: 50–650 m3/s) Y = 0.13X 21.36, Y = flow rate (m3/s or CMS), X = SS concentration (mg/L)].
in lower catchments. The averaged BOD concentrations varied from 1.5 to 2.5 mg/L in three reaches compared to 5.8 mg/L in Kaoping River in dry seasons. In wet seasons, the averaged BOD concentrations varied from 0.1 to 1.1 mg/L in three reaches compared to 2.4 mg/L in Kaoping River. This was due to the fact that most of the land use patterns were mainly forest and agricultural uses, which produced had less pollution levels compared to land use patterns of human activities in the lower catchments. Because the lower catchments have been developed to domestic and industrial areas (human activity categories), higher BOD and NH3–N concentrations were detected at sampling stations located in this region (W10 to W12). The variability of pollutant concentrations is usually affected by two different forcing functions including the hydrodynamic flow rate of the watershed and the impacts of
Organic nitrogen Mineralization of dissolved ON rate constant Mineralization of dissolved ON rate temperature constant Organic phosphorus Mineralization of dissolved OP rate constant Mineralization of dissolved OP rate temperature constant
K12T
K1T K1RC K1RT K1D K1G OCRB PCRB NCRB KDC
K71T
K83C K83T
Fraction of dead and respired phytoplankton Fraction of ON from algal death FON Fraction of OP from algal death
FOP
1.04 0.01
1.066 0.008 1.08 0.11 0.80 2.67 0.028 0.200
1.040
1.04
1.04
0.3
ON: organic nitrogen; OP: organic phosphorus.
human activities. Due to the variability of the two forcing functions, variable BOD and NH3–N concentrations were observed. Higher DO values were observed in the upper catchments due to the natural turbulence (natural reaeration) in three reaches (varied from 7.1 to 7.5 mg/L in wet seasons), which enhanced the oxy-
96
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 6. Comparison of time-variable simulated results and observed BOD at water monitoring stations.
gen mass transfer between the river and atmosphere. This caused higher DO concentrations in the river. Moreover, lower DO concen-
trations in downstream sections (3.7 mg/L in Kaoping River) during the dry seasons were mainly due to the higher pollution loading in
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
97
Fig. 7. Comparison of time-variable simulated results and observed NH3–N at water monitoring stations.
lower catchments and lower oxygen solubility in the dry seasons. Results from the RPI evaluation indicate that the water quality changed from slightly polluted in dry seasons to moderate polluted
in wet seasons in three reaches. Results from Tables 3 and 4 reveal that SS was an important parameter among the four parameters in RPI calculation. Results indicate that without the dilution of high
98
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Fig. 8. Comparison of time-variable simulated results and observed DO at water monitoring stations.
flow of river water and soil erosion in dry seasons, low DO and high BOD values became the determinant factors in the RPI calculation.
Fig. 5 presents the correlation between the measured SS and flow rates. Results indicate that the SS and flow rate values of water samples collected from the three major reaches and main
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
Kaoping River had a significant correlation [R2 values of 0.9 and 0.87 (Eq. (8) and (9))]. Thus, SS concentrations in river water can be calculated using the following equation:
99
value could then be calculated for water quality evaluation. The equations were embedded into the integrated WASP modeling for SS simulation to obtain more representative SS data and RPI values.
Low flow rate ðflow rate : 0—50 m3 =sÞ Y ¼ 0:028X 0:37 R2 ¼ 0:90
ð8Þ
High flow rate ðflow rate : 50—650 m3 =sÞ Y ¼ 0:13X 21:36 R2 ¼ 0:87
ð9Þ
where Y = flow rate (m3/s) and X = SS concentration (mg/L). In the Kaoping River Basin, most of the SS concentrations were caused by the soil erosion of the upper catchment during the wet seasons. Thus, the obtained empirical equations could be used for immediate SS calculation when flows have significant changes. The RPI
3.2. Water quality modeling In this study, the WASP modeling was performed for time-variable river water quality simulation. Table 5 shows the parameters and input values used in the WASP modeling. Figs. 6–8 present the comparisons of observed BOD, NH3–N, and DO data and time-variable simulated results at different water sampling stations in 2011. The WASP modeling performed a good match compared to the observed data. Results suggest that the monthly BOD and NH3–N loadings were affected by the flows and agricultural activities.
Fig. 9. Comparison of time-variable simulated results and observed SS at water monitoring stations.
100
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
In this study, the SS equation assisted WASP modeling was applied for time-variable river water quality simulation. The WASP model was then used for RPI simulation. Fig. 9 presents the comparisons of observed SS data and time-variable simulated results at different water sampling stations in 2011. Results indicate that the SS equation assisted modeling had reasonable matches with the observed data compared to the simulated SS results using the empirical flows derived SS equations. Results indicate that the SS equation assisted WASP modeling was able to include the influence of soil erosion on the increase in SS concentrations. Thus, the modified WASP model provided a good tool to effectively simulate the impacts of SS pollution on river water quality.
3.3. RPI calculation In this study, RPI equation was embedded in the WASP model for RPI calculation and water quality evaluation. Fig. 10 presents the calculated RPI values using WASP model containing RPI module (equation described in Table 1). In Fig. 10, RPI values for each monitoring station were simulated using the simulated water quality results of BOD, DO, NH3–N, and SS values for the whole year. Results show that in the dry seasons, SS played an important role in RPI calculation. SS was a critical factor during the RPI calculation especially for the upper catchment. Because of the discharge of domestic and industrial wastewaters into the river, higher BOD
Fig. 10. Comparison of time-variable simulated results and observed RPI at water monitoring stations.
Y.C. Lai et al. / Journal of Hydrology 478 (2013) 89–101
and NH3–N concentrations were observed in the downstream sections of the river. Thus, BOD and NH3–N became the determinant factors in the downstream section of the Kaoping River for the RPI calculation. In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Higher river flow rates also caused the turbulence of the flow and increase in DO concentrations in the upper sections of the river. Therefore, all four parameters (BOD, DO, NH3–N, and SS) would affect the RPI calculation in the downstream sections. 4. Conclusions In this study, SS estimation equation and RPI package were embedded into the WASP model to assist the river water quality simulation of the Kaoping River. A direct linkage between the RPI calculation and the WASP model was developed to obtain an immediate water quality evaluation. The major findings of this study include the following: (1) Water quality monitoring results indicate that SS played an important role in RPI calculation and SS was a critical factor during the RPI calculation especially for the upper catchment in wet seasons. This was due to the fact that the soil erosion caused the increase in the SS concentrations after storms. (2) In the wet seasons, higher river flow rates caused the discharges of NPS pollutants (NH3–N and SS) into the upper sections of the river. Higher river flow rates also caused the turbulence of the flow and increase in DO concentrations in the upper section of the river. Thus, NH3–N and DO affected the RPI value in the upper sections of the river. (3) SS concentrations were highly correlated with the flow rates of the Kaoping River. The obtained SS and flow rate equations were as follows: Y = 0.028X 0.37 for low flow rate (flow rate: 0–50 m3/s), Y = 0.13X 21.36 for high flow rate (flow rate: 50–650 m3/s) (Y = flow rate, m3/s; X = SS concentration, mg/L). The equations were embedded in the integrated WASP modeling for SS simulation to obtain more representative SS data and RPI values (4) During the evaluation, water quality outputs from the WASP modeling were used for RPI calculation. Thus, an integral decision-making system combining SS estimation equation, RPI package, and water quality model for river water quality evaluation was developed. (5) This study demonstrates that the introduction of SS estimation equation and RPI package into WASP simulation has been shown to be a significant advance in estimating water quality in Kaoping River. The developed decision-making modeling concept could be easily adopted to other similar rivers. Thus, WASP modeling assisted RPI estimation provides a good tool to effectively simulate the impacts of pollution on river water quality. The integrated system will be useful in developing appropriate water quality management strategies for the improvement of river water quality, and this makes it easier for decision-makers to evaluate alternative water quality management plans.
Acknowledgements This study was funded by Taiwan National Science Council and Taiwan Environmental Protection Administration (TEPA). Addi-
101
tional thanks to the personnel of TEPA for their assistance throughout this project. References Ambrose, B., Wool, T.A., Martin, J.L., 2001. The Water Quality Analysis Simulation Program, WASP6, User Manual, US EPA, Athens, GA. APHA (American Public Health Association), 2005. Standard Methods for the Examination of Water and Waste Water, 22th ed. APHA-AWWA-WPCF, Washington, DC. Canu, D.M., Solidoro, C., Umgiesser, G., 2004. Erratum to modeling the responses of the Lagoon of Venice ecosystem to variations in physical forcing. Ecol. Model. 175, 197–216. Caruso, B.S., 2004. Modeling metals transport and sediment/water interactions in a mining impacted mountain stream. Water Resources Assoc. 40, 1603–1615. Cencic, O., Rechberger, H., 2008. Material flow analysis with software STAN. Sustain. Environ. Res. 18, 3–7. Chen, S.W., Chao, A.C., Chen, T.Y., Kao, C.M., Lai, Y.C., Lin, C.E., 2006. Application of water quality modeling on river basin management. WSEAS Trans. Math. 5, 1078–1083. Fan, C., Ko, C.H., Wang, W.S., 2009. An innovated modeling approach using Qual2K and HEC-RAS integration to assess the impact of tidal effect on river water quality simulation. J. Environ. Manage. 90, 1824–1832. Geza, M., Poeter, E.P., McCray, J.E., 2009. Quantifying predictive uncertainty for a mountain watershed model. J. Hydrol. 376, 170–181. Kavvas, M.L., Yoon, J., Chen, Z.Q., Liang, L., Dogrul, E.C., Ohara, N., Aksoy, H., 2006. Watershed environmental hydrology model: environmental module and its application to a California watershed. J. Hydrol. Eng. 11, 450–464. Kima, D., Wanga, Q., Soriala, G.A., Dionysioua, D.D., Timberlake, D., 2004. A model approach for evaluating effects of remedial actions on mercury speciation and transport in a lake system. Sci. Total Environ. 327, 1–15. Kuo, J.T., Lung, W.S., Yang, C.P., Liu, W.C., Yang, M.D., Tang, T.S., 2006. Eutrophication modelling of reservoirs in Taiwan. Environ. Model. Softw. 21, 829–844. Lai, Y.C., 2010. Development of Kaoping River Management Strategies. Ph.D. Dissertation, National Sun Yat-Sen University, Kaohsiung, Taiwan. Lai, Y.C., Yang, C.P., Hsieh, C.Y., Wu, C.Y., Kao, C.M., 2011. Evaluation of non-point source pollution and river water quality using a multimedia two-model system, Taiwan. J. Hydrol. 409, 583–595. Lin, C.E., 2010. Development of River Water Quality and Sediment Management Strategies. Ph.D. Dissertation, National Sun Yat-Sen University, Kaohsiung, Taiwan. Lin, C.E., Kao, C.M., Jou, C.J., 2010. Preliminary identification of watershed management strategies for the Houjing River in Taiwan. Water Sci. Technol. 62, 1667–1675. Lin, C.E., Chen, C.T., Kao, C.M., Hong, A., Wu, C.Y., 2011. Development of the sediment and water quality management strategies for the Salt-water River, Taiwan. Mar. Pollut. Bullet. 63, 528–534. Luo, H., Li, M., Xu, R., Fu, X., Huang, G., Huang, X., 2012. Pollution characteristics of urban surface runoff in a street community. Sustain. Environ. Res. 22, 61–68. Markku, P., Eila, T., Manna, K., Jari, K., Jarmo, L., Rittu, N., Sirkka, T., 2010. YIHMA-A tool for allocation of measures to control erosion and nutrient loading from Finnish agricultural catchment. Ecosyst. Environ. 138, 306–317. NIEA (National Institute of Environmental Analysis, Taiwan EPA), 2004. Method of River Flow Rate Analysis, NIEA W022.51C. Ouyang, W., Wang, X., Hao, F., Srinivasan, R., 2009. Temporal-spatial dynamics of vegetation variation on non-point source nutrient pollution. Ecol. Model. 220, 2702–2713. Papanicolaou, T., Abaci, O., 2005. An integrated watershed model. In: Managing Watersheds for Human and Natural Impacts: Engineering, Ecological, and Economic Challenges, Managing Watersheds for Human and Natural Impacts: Engineering, Ecological, and Economic Challenges – Proceedings of the 2005 Watershed Management Conference, p. 189. Sehnert, C., Lindenschmidt, K.E., 2009. The effect of upstream river hydrograph characteristics on zinc deposition, dissolved oxygen depletion and phytoplankton growth in a flood detention basin system. Quatern. Int. 208, 158–168. TEPA, Taiwan Environmental Protection Administration, 2002. Development of Non-Point Source Pollutant Remedial Strategy, Taipei, Taiwan. TEPA, Taiwan Environmental Protection Administration, 2009. Project for Aerial Photo Interpretation and Analysis of Water Pollution Sources in Kaoping River, Annual Report, Taipei, Taiwan. Xu, Z., Godrej, A.N., Grizzard, T.J., 2007. The hydrological calibration and validation of a complexly-linked watershed–reservoir model for the Occoquan watershed, Virginia. J. Hydrol. 345, 167–183. Xu, K., Milliman, J.D., Xu, H., 2010. Temporal trend of precipitation and runoff in major Chinese River since 1951. Glob. Planet. Change 73, 219–232. Yang, C.P., Kuo, J.T., Lung, W.S., Lai, J.S., Wu, J.T., 2007. Water quality and ecosystem modeling of tidal wetlands. J. Environ. Eng. 133, 711–721. Zhang, M.L., Shen, Y.M., Guo, Y., 2008. Development and application of a eutrophication water quality model for river networks. J. Hydrodyn. 20, 719– 726.