- Email: [email protected]
Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
Ecological Modelling xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Ecological Modelling journal homepage: www.elsevier.com/locate/ecolmodel
Research Paper
Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model Jaeyoung Kim, Tongeun Lee, Dongil Seo ∗ Department of Environmental Engineering, Chungnam National University, 99, Daehak-ro, Yuseong-gu, Daejeon, 34134, Republic of Korea
a r t i c l e
i n f o
Article history: Received 27 June 2017 Received in revised form 26 October 2017 Accepted 29 October 2017 Available online xxx Keywords: The Han River Algal bloom Toxic cyanobacteria Environmental Fluid Dynamics Code Water quality modeling
a b s t r a c t The lower part of the Han River, which flows through Seoul, Korea, experienced excessive toxic cyanobacterial growth in 2015. Modeling of algal bloom occurrence patterns in the lower part of this river was performed using the Environmental Fluid Dynamics Code (EFDC) to understand algal dynamics and thus better develop management alternatives. For a 71 km long river section, 1175 horizontal 2D grid elements were developed. This grid system was determined adequate, as the maximum values of the Courant–Friedrichs–Lewy condition and orthogonality deviation were 0.5 and 20.1, respectively. Chlorophyll-a (Chl-a) was chosen as the primary indicator for the likelihood of algal blooms. Calibration and verification of EFDC were performed by comparing the model results to three years of field data collected from 2013 to 2015. Calibration accuracy was verified not only for physical variables, including the mean water level and temperature, but also for other water quality variables in various locations of the study area. To improve the prediction accuracy of Chl-a, three dominant groups of algae were considered: diatoms, green algae, and cyanobacteria. The optimum growth temperature ranges were selected based on field data for the study area. It was found necessary to apply different maximum growth rates for algal groups for the upstream and downstream regions of the study area to appropriately reflect field observations. This result indicates that more than three algal groups need to be included to improve Chl-a calibration accuracy for the study area, yet the current EFDC model can consider only up to three phytoplankton groups. Although this problem could be overcome by assigning different maximum growth rates for different regions, it may be necessary to improve EFDC so that it can include more phytoplankton groups. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The lower section of the Han River, which passes through Seoul City, Republic of Korea (hereafter Korea), experienced a relatively severe algal bloom in 2015. The Korean Government developed a four- level algal warning system by measuring Chlorophyll-a (Chla) concentrations and the number of toxic cyanobacteria cells per milliliter. During July and September in 2015, Chl-a concentrations and the number of toxic cyanobacteria were measured at levels greater than 25 mg/m3 and 5000 cells/ml, respectively. A second level algae warning based on the Korean system was declared in several locations of the river. An analysis that clearly understands the causes and effects of algal blooms and water pollution within a water body is often very difficult to achieve due to the many factors involved. For example, Chapra (1997) reported that a sufficient level of nutrients in
∗ Corresponding author. E-mail addresses: [email protected], [email protected] (D. Seo).
the water body for phytoplankton to feed on, an appropriate water temperature for algal growth, and an adequate amount of light for photosynthesis are all essential conditions for an algal bloom to occur. Reynolds (2006) reported the growth of algae is affected by competition between different algal groups, predation in the food chain, the degree of pollution, and the flow characteristics of the river are important factors. In the case of the lower Han River, many factors can affect algal blooms. It would be difficult to conclude that only one or a few factors would be responsible. Various methods may be applied to control algal blooms in surface waters. Physical methods, such as dilution or mechanical mixing may be used, but require enormous amounts of energy. Chemical methods such as algaecides or coagulants can be used, but are usually applied to relatively small and stagnant water bodies and are not used for rivers. Biological methods, such as the introduction of predators or constructed wetlands also can be used. All of these methods only deal with surficial symptoms of algal blooms and only last a limited period of time. Therefore, it is necessary to identify feasible water quality management alternatives that can be applied efficiently (Anderson, 2009).
https://doi.org/10.1016/j.ecolmodel.2017.10.015 0304-3800/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10 2
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
Water quality modeling is a powerful tool for the comprehensive interpretation of complex water quality interactions (Ambrose et al., 2009). Many different models have been applied to water bodies to analyze which countermeasures may be effective for mitigating algal blooms (Chen and Mynett, 2006; Wong et al., 2007; Los et al., 2008; Leon et al., 2011; Seo et al., 2012; Camacho et al., 2014; Jian et al., 2014). The Environmental Fluid Dynamics Code (EFDC) (Hamrick, 1992) is a versatile surface water model that has been used widely for various water body types such as rivers, lakes, wetlands, dams, estuaries, and coasts for environmental assessment and management. Park et al. (2005) applied the EFDC water quality model in Gwangyang Bay, Korea. Li et al. (2011) applied the concept of water age using EFDC to understand the effect of the migration of dissolved substances on the water quality of Lake Taihu, China. Wu and Xu (2011) and Tang et al. (2016) used EFDC to predict the occurrence of algal blooms in Daoxiang and Taihu lakes in China, respectively. Seo and Song (2015) conducted a three-dimensional hydrodynamic and water quality modeling analysis using EFDC in the Youngsan River, the fourth largest river in Korea. Yin and Seo (2016) analyzed an optimal grid selection for EFDC water quality modeling of the Ara Navigation Channel in Korea. Lee et al. (2017) conducted a study to evaluate the impact of installing of a new sewage treatment plant on the water quality of the Galing River, Malaysia. Although many studies have been carried out, an accu-
rate prediction of the occurrence of algal blooms has been difficult to achieve. Since algae are not simple substances but living organisms, their population dynamics are difficult to explain with simple mass balance equations or chemical processes. Sufficient data is rarely available to estimate boundary conditions, define the parameters of water quality variables, or solve the various algal growth related problems required to perform the modeling. This study aims to improve the prediction accuracy for harmful algal bloom in the lower Han River, Seoul, Korea so that the model can be used in decision making processes in developing management alternatives. The growth-related parameters of algae groups were selected from the literature and from field observations in the study area, and then further adjusted in the calibration of EFDC. To improve model prediction accuracy for algal blooms in this study, different maximum growth rates were applied for algal groups in regions of the river influenced by different water quality conditions. 2. Materials and methods 2.1. Study area The Han River is the second largest river in Korea. It is 483 km long and its basin area is 34,428 km2 . This study focuses on a lower 71 km section of the river that passes through Seoul from the Pal-
Fig. 1. Monitoring stations and important in-stream structures in the study area.
Fig. 2. Fraction of algae in upstream and downstream.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
dang Dam to the Jeonryu water level station as shown in Fig. 1. There are five water level stations, fourteen water quality monitoring stations, and nine algae monitoring stations within the study site. Two fixed weirs, the Jamsil Weir (hereafter Weir 1) and Singok Weir (hereafter Weir 2) have been installed to maintain the water level and prevent salt water intrusion into the river, as Weir 2 is affected by tidal movement from the western coast of Korea. Between the two weirs, four major wastewater treatment plants in Seoul are operational and discharge effluent directly into the Han River.
2.2. Chl-a and cyanobacteria in the Han River during the period between 2013–2015 Fig. 2 shows the algae fractions observed in the study site between 2013 and 2015, both in the upstream region between Paldang Dam and Weir 1 and in the downstream region between Weir 1 and Weir 2. Diatoms were shown to be dominant during most of each year studied. Green algae and cyanobacteria growth only became significant during each summer. In 2015, cyanobacteria became dominant in the summer, and severe algal blooms occurred between August and September. Some of the most common toxinproducing cyanobacteria include the N-fixing genera: Anabaena, Aphanizomenon, Cylindrospermopsis, Lyngbya, Nodularia, Oscillatoria, and Trichodesmium; and non-N-fixing genera: Microcystis and
3
Planktothrix (Paerl and Otten, 2013). The Ministry of Environment developed the Korean Algae Warning System in 1997, which was updated in 2015. In the new Korean Algae Warning System, the cell numbers of Anabaena, Microcystis, Aphanizomenon, and Oscillatoria are used to trigger different warning levels from 2016 onwards (National Institute of Environmental Research, 2017). Nitrogen to phosphorus (N/P) ratios of twenty or greater generally indicate a phosphorus limited system for algal growth while N/P ratios of five or less indicate nitrogen limited systems (Thomann and Mueller, 1987). As bodies of freshwater become enriched with nutrients, especially phosphorus, there is often a shift in the phytoplankton community towards a dominance by cyanobacteria (Smith, 1986; Trimbee and Prepas, 1987; Watson et al., 1997; Paerl and Huisman, 2009; O’Neil et al., 2012). In 2015, the N/P ratio in the Han River was 70 upstream of Weir 1 and 25 downstream of Weir 1. The nutrient limiting algal growth in the system was identified as phosphorus (Fig. 3). Fig. 4 shows the number of cyanobacteria cells and total phosphorus (TP) concentrations upstream of Weir 1 in 2015. In this region, phosphorus concentrations ranged from 0.02 to 0.08 mg/L, and the maximum toxic cyanobacteria number was about 10,000 cells/ml, while the maximum non-toxic cyanobacteria number was about 39,000 cells/ml. On the other hand, the concentration of phosphorus in the downstream region, between Weir 1 and Weir 2, ranged from 0.05 to 0.43 mg/L. In this area, toxic cyanobacteria were dominant and
Fig. 3. N/P ratio of the lower Han River.
Fig. 4. Toxic and non-toxic cyanobacteria in upstream and downstream areas.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
4
Fig. 5. Conceptual diagram of the EFDC model.
its maximum count was about 40,000 cells/ml. Many researchers have reported that summer phytoplankton numbers are dominated by cyanobacteria at phosphorus concentration between approximately 0.1 and 1.0 mg/L (Trimbee and Prepas, 1987; Jensen et al., 1994; Watson et al., 1997; Downing et al., 2001; O’Neil et al., 2012). It seems that toxic cyanobacteria grow better in the high phosphorus concentrations in the summer period than non-toxic cyanobacteria. It is suggested that the growth pattern of each algal group can be different and may vary depending on the region of study. 2.3. Model development and boundary conditions The EFDC three-dimensional hydrodynamic and water quality model was used to simulate water quality and algal blooms in the study area. The model includes water quality, sediment transport, and toxicity modules as shown in Fig. 5. The EFDC computational grid was developed by the curvilinear orthogonal coordinate method. 1175 horizontal grid points were selected by a trial and error method. In the selected grid system, the Courant–Friedrichs–Lewy (CFL) condition (Courant et al., 1928) was estimated to be less than 0.50 and the average value of orthogonality deviation was 3.14. Calibration and verification of the model to observed data were performed for the three years period from 2013 to 2015. As shown in Fig. 1, the boundary conditions included thirteen tributaries along with the headwater from the discharge of the Paldang Dam, five water intake systems, five wastewater treatment plants, and open boundary conditions that reflect the tidal movement along the west coast. The necessary field data for this study were obtained from Korean governmental database management systems. The
Table 1 Summary of the kinetic coefficients employed in the current model application. Parameter
Value
Optimal temperature for algal growth (◦ C) Basal metabolism rate for algae [day−1 ] Predation rate for algae [day−1 ] Half-saturation constant for nitrogen uptake of algae [mg L−1 ] Half-saturation constant for phosphorus uptake of algae [mg L−1 ] Half-saturation constant for silica uptake of diatoms [mg L−1 ] Algae settling rate [m day−1 ] Decay rate of organic carbon [day−1 ] Decay rate of organic phosphorus [day−1 ] Decay rate of organic nitrogen [day−1 ] Settling velocity of particulate organic matter [m day−1 ] Benthic flux rate of phosphate [g m−2 day−1 ] Benthic flux rate of ammonia nitrogen [g m−2 day−1 ] Benthic flux rate of nitrate [g m−2 day−1 ]
25, 13–15, 25a 0.06 0.1 0.01 0.001 0.05 0.215 0.005, 0.075, 0.1b 0.005, 0.075, 0.1b 0.005, 0.075, 0.015b 2.19 0.003 0.15 0.03
a
For cyanobacteria, diatoms, and green algae, respectively. b For refractory particulate, labile particulate, and dissolved organic matter, respectively.
Water Information System (MOE, 2017) was used for water quality data and the Water Resources Management Information System (MOLIT, 2017) was used for water quantity data. Table 1 shows selected calibrated parameters used in this study. The parameters were initially chosen based on previous research results (Bowie et al., 1985; Cole and Wells, 2006; Shin et al., 2008) and then adjusted further to improve the calibration accuracy in the cur-
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
5
Table 2 Scenarios used in this study. Maximum Growth rate [day−1 ]
Cyanobacteria
(a) Overall value (b) Different value for each algal group (c) Different value for per region and algal group
2.5 2.5 Upper 3.5
Diatoms
Green Algae
1.0 Upper 1.5
Lower 2.5
3.0 Upper 4.0
Lower 1.1
Lower
rent study. The optimal temperatures for algal growth for the three different groups in Table 1 were selected based on field observations of the study area, as shown in Fig. 2. Although this study did not simulate sediment transport, nutrients released from bottom sediments were considered as shown in Table 1.
Table 3 AME and RMSE of the water level for selected locations of the Han River.
2.4. Development of scenarios for algae prediction
regarding algal growth not considered in the model. To increase the accuracy of the algal growth simulation in this study, different maximum growth rates for different algal groups were assigned for each region. The spatial variability reflects various factors affecting algal growth.
It is difficult to simulate the occurrence of algal blooms accurately for complex natural water systems. Many conditions affect the growth and loss of algae in a water body and there is often a lack of appropriate data. In EFDC, the overall dynamics and growth of algae in a cell are determined by Eqs. (1) and (2). EFDC can simulate up to three groups of algae. Because the number of simulated algae groups is limited, EFDC cannot reflect various factors
Station
WL1
WL2
WL3
WL4
AME [m] RMSE [m]
0.14 0.29
0.29 0.41
0.19 0.36
0.17 0.34
∂Bx ∂ WBx = (Px − BMx − PRx )Bx + (WSx · Bx ) + V ∂t ∂z
(1)
Bx = algal biomass of algal group x (g C m−3 )
Fig. 6. Calibration results of water level and water temperature for selected locations of the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
6
Table 4 AME and RMSE for temperature in selected locations of the Han River. Station
Q2
Q3
Q4
Q5
Q6
Q7
Q8
Q9
Q10
Q11
Q12
Q13
AME [◦ C] RMSE [◦ C]
0.63 0.85
0.69 0.86
0.83 1.04
0.99 1.16
1.07 1.33
1.10 1.36
1.10 1.37
1.32 1.71
1.55 1.94
1.48 1.85
1.37 1.66
1.68 2.07
t = time (day) Px = production rate of algal group x (day−1 ) BMx = basal metabolism rate of algal group x (day−1 ) PRx = predation rate of algal group x (day−1 ) WSx = positive settling velocity of algal group x (m day−1 ) WBx = external loads of algal group x (g C day−1 ) V = cell volume (m3 ) Px = PMx ·f 1 (N)·f 2 (I)·f 3 (T)
perature and light intensity were chosen from observations Fig. 2 for each group of algae and these were shown in Table 1. 3. Results and discussion 3.1. Hydrodynamic model calibration (2)
PMx = maximum growth rate under optimal conditions for algal group x (day−1 ) f1 (N) = effect of suboptimal nutrient concentration (0–1) f2 (I) = effect of suboptimal light intensity (0–1) f3 (T) = effect of suboptimal temperature (0–1) Three different calibrations were performed and compared to field observations to determine their accuracy. Table 2 shows the three calibration scenarios: a) the same maximum growth rate for all algae groups, b) different maximum growth rate for three algal groups and c) different growth rates for different regions (upper and lower stream areas from Weir 1 in Fig. 1). The optimum tem-
Calibration and verification of the hydrodynamic model were carried out using water level and water temperature data for the Han River collected for three years from 2013 to 2015. For water level, data from four monitoring stations in the study site were used for calibration. For water temperature, calibration was carried out for twelve water quality monitoring stations. Fig. 6 shows the calibration results for all water level stations and some selected stations for water temperature. Tables 3 and 4 show the results of the analysis using both absolute mean error (AME) and root mean square error (RMSE) to evaluate the accuracy of the simulated results. The mean values for the water level were 0.20 m and 0.35 m respectively. In the case of temperature, the mean values were 1.15 ◦ C and 1.43 ◦ C respectively.
Fig. 7. Water quality calibration results for scenarios (a) and (b) in the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
7
Fig. 7. (Continued)
3.2. Algae and water quality calibration 3.2.1. Scenarios (a) and (b); overall and different growth rate for each algal group The water quality calibration was performed at 12 monitoring stations. Calibrated water quality variables include total organic carbon (TOC), TP, phosphorous-phosphate (PO4 -P), total nitrogen (TN), nitrogen-ammonia (NH3 -N), nitrogen-nitrate (NO3 -N), dis-
solved oxygen (DO) and Chl-a. Chl-a was chosen as a primary indicator for overall algal bloom for all the scenario in this study as this is only available in the national water quality monitoring database system (MOE, 2017). Fig. 7 shows the calibration results for scenarios (a) and (b) in selected locations and for four water quality variables. It seems that the calibration results for TP and TN are relatively more accurate than other water quality variables in this study. In the case of TOC and Chl-a, calibration results for
Table 5 AME analysis of the water quality calibration results. AME
Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Average
Scenario (a)
Scenario (b) 3
Scenario (c) 3
TOC [mg/L]
Chl-a [mg/m ]
TOC [mg/L]
Chl-a [mg/m ]
TOC [mg/L]
Chl-a [mg/m3 ]
0.37 0.37 0.40 0.67 0.39 0.48 0.47 0.62 0.60 0.66 1.00 0.61 0.55
3.11 5.63 6.36 7.57 7.71 7.23 7.65 10.73 13.11 14.00 17.34 16.50 9.75
0.36 0.35 0.36 0.75 0.32 0.44 0.46 0.70 0.35 0.46 1.32 0.60 0.54
2.92 5.25 6.10 6.49 6.12 4.71 4.70 7.90 10.70 12.28 16.54 16.19 8.33
0.37 0.36 0.37 0.68 0.33 0.38 0.39 0.60 0.30 0.34 1.10 0.45 0.47
3.03 5.12 5.73 6.43 5.91 4.96 4.96 6.94 10.40 11.86 16.85 15.69 8.16
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
8 Table 6 RMSE analysis of the water quality calibration results. RMSE
Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Average
Scenario (a)
Scenario (b)
Scenario (c)
TOC [mg/L]
Chl-a [mg/m3 ]
TOC [mg/L]
Chl-a [mg/m3 ]
TOC [mg/L]
Chl-a [mg/m3 ]
0.67 0.63 0.63 1.15 0.57 0.63 0.62 0.91 0.73 0.78 1.35 0.71 0.78
3.98 8.01 10.40 11.47 10.24 8.95 9.63 14.01 16.89 17.43 22.67 20.91 12.88
0.66 0.62 0.60 1.20 0.50 0.61 0.59 0.96 0.46 0.57 1.59 0.69 0.75
3.89 7.88 10.43 11.19 9.24 6.47 6.71 12.60 16.34 17.92 23.60 23.80 12.51
0.66 0.62 0.60 1.11 0.51 0.54 0.54 0.83 0.40 0.44 1.40 0.54 0.68
4.06 7.20 9.54 10.35 8.50 6.01 6.18 10.93 15.57 16.94 22.38 22.65 11.69
upstream areas were more accurate than downstream areas. Scenario (b) showed better reproducibility for TOC and Chl-a results, although greater errors result in the downstream area especially for high Chl-a concentrations. As illustrated in Fig. 4, the growth of the three algal groups in the study area was different in each region. These results suggest that it is necessary to consider growth characteristics separately for each group to achieve better accuracy in the modeling of Chl-a.
3.2.2. Scenario (c): different maximum growth rate per each region and algal group Fig. 8 shows the simulation results for scenario (c). Calibration results for NH4 -N, NO3 -N, PO4 -P, and DO were added to show the complete range of water quality variables considered in this study. The calibration accuracy for TOC and Chl-a showed greater improvement than that for TN and TP. Tables 5 and 6 show a statistical analysis of TOC and Chl-a calibration results. The two tables show that the prediction errors become smaller as the growth
Fig. 8. Water quality calibration results for scenario (c) in the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
9
Fig. 8. (Continued)
rates of the algal groups become more specific. This result suggests that it may be necessary to include more algal groups or genera to improve the prediction accuracy as shown in Fig. 4. Chapra et al. (2017) also suggested using four phytoplankton functional groups by dividing cyanobacteria group to N-fixers and non-Nfixers to improve prediction accuracy of cyanobacterial harmful algal blooms in his QUALIDAD model. Though it was possible to improve the calibration accuracy by subdividing the growth rates for different conditions, calibration of Chl-a in the downstream area, for example in Q13, did not seem to reflect the field data, especially for more extreme conditions. This is interpreted as the limit of data that can be reflected by the model, as well as the effects of other factors on the growth of algae that EFDC does not consider. 4. Conclusion This study was carried out to simulate the changes in water quality associated with algal blooms in the Han River using the EFDC model. For the simulation period from 2013 to 2015, calibration of the model was performed with respect to water level, water temperature, and water quality variables in various locations of the study area. The aim of this study was to improve the reproducibility of algal growth as reflected by Chl-a using growth-rate-related parameters of multiple algal groups applied to the Han River. The dominant groups of algae in the study area were diatoms, green algae, and cyanobacteria. We observed that non-toxic
cyanobacteria were more significant upstream, where the phosphorus concentration was lower. On the other hand, toxic cyanobacteria were more significant downstream, which had a higher concentration of phosphorus. The results of this study suggest that the degree of algal growth differs depending on the environment and region. Therefore, to reflect this phenomenon better, the maximum growth rates of algal groups were classified for different sections and entered into the model accordingly. We analyzed three scenarios to compare the prediction accuracy of Chl-a. The application of different maximum growth rates for different algal groups was found to be more reproducible than was applying the same growth rate. Moreover, accuracy was the highest when the maximum growth rate for algal groups was used separately for upstream and downstream areas of Weir 1 of the study area. Although the overall model accuracy was improved compared to the original model, the error was larger downstream of Weir 1. For extremely high concentrations of Chl-a, algal growth was not accurately reflected by the model. This implies that other factors not considered in EFDC affect the growth of algae. This result suggests that the current algal simulation method in EFDC may need to be modified to improve the prediction accuracy, especially for large harmful algal blooms. The predicted results for water quality and the occurrence of algal blooms in the Han River made using these assumptions are considered significant, and continued research to confirm these results is needed.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
Acknowledgement This study was supported by Water Circulation Policy Division of Seoul Metropolitan Government. References Ambrose Jr., R.B., Wool, T.A., Barnwell Jr., T.O., 2009. Development of Water Quality Modeling in the United States. Environ. Eng. Res. 14, 200–210. Anderson, D.M., 2009. Approaches to monitoring, control, and management of harmful algal blooms (HABs). Ocean Coast. Manag. 52 (7), 342–347. Bowie, G.L., Mills, W.B., Porcella, D.B., Campbell, C.C., Pagenkopf, J.R., Rupp, G.L., Johnson, K.M., Chan, R.W.H., Gherini, S.A., Chamberlin, C.E., 1985. Rates, Constants and Kinetics Formulations in Surface Water Quality Modeling, second ed. Environment Research Laboratory, Environmental Protection Authority, EPA/600/385/040. Camacho, R.A., Martin, J.L., Watson, B., Paul, M.J., Zheng, L., Stribling, J.B., 2014. Modeling the factors controlling phytoplankton in the St. Louis Bay estuary, Mississippi and evaluating estuarine responses to nutrient load modifications. J. Environ. Eng. 141 (3), 04014067. Chapra, S.C., Boehlert, B., Fant, C., Bierman Jr., V.J., Henderson, J., Mills, D., Mas, D.M.L., Rennels, L., Jantarasami, L., Martinich, J., Strzepek, K.M., Paerl, H.W., 2017. Climate change impacts on harmful algal blooms in U.S. freshwaters: a screening-level assessment. Environ. Sci. Technol. 51 (16), 8933–8943. Chapra, S.C., 1997. Surface Water-Quality Modeling. McGraw-Hill Companies, Inc. Chen, Q., Mynett, A.E., 2006. Modeling algal blooms in the Dutch coastal waters by integrated numerical and fuzzy cellular automata approaches. Ecol. Model. 199 (1), 73–81. Cole, T.M., Wells, S.A., 2006. CE-QUAL-W2: A Two-dimensional, Laterally Averaged, Hydrodynamic and Water Quality Model, Version 3.5. Instruction Report EL-06-1. US Army Engineering and Research Development Center, Vicksburg, MS. Courant, R., Friedrichs, K., Lewy, H., 1928. Über die partiellen differenzengleichungen der mathematischen. Physik. Math. Ann. 100 (1), 32–74. Downing, J.A., Watson, S.B., McCauley, E., 2001. Predicting cyanobacteria dominance in lakes. Can. J. Fish. Aquat. Sci. 58 (10), 1905–1908. Hamrick, J.M., 1992. A three-dimensional environmental fluid dynamics computer code: theoretical and computational aspects. In: Special Report. Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA, pp. 317. Jensen, J.P., Jeppesen, E., Olrik, K., Kristensen, P., 1994. Impact of nutrients and physical factors on the shift from cyanobacterial to chlorophyte dominance in shallow Danish lakes. Can. J. Fish. Aquat. Sci. 51 (8), 1692–1699. Jian, L., Zhongwu, J., Wenjun, Y., 2014. Numerical modeling of the Xiangxi River algal bloom and sediment-related process in China. Ecol. Inform. 22, 23–35. Lee, I., Hwang, H., Lee, J., Yu, N., Yun, J., Kim, H., 2017. Modeling approach to evaluation of environmental impacts on river water quality: a case study with Galing River, Kuantan, Pahang, Malaysia. Ecol. Model. 353, 167–173. Leon, L.F., Smith, R.E., Hipsey, M.R., Bocaniov, S.A., Higgins, S.N., Hecky, R.E., Antenucci, J.P., Imberger, J.A., Guildford, S.J., 2011. Application of a 3D hydrodynamic–biological model for seasonal and spatial dynamics of water quality and phytoplankton in Lake Erie. J. Great Lakes Res. 37 (1), 41–53. Li, Y., Acharya, K., Yu, Z., 2011. Modeling impacts of Yangtze River water transfer on water ages in Lake Taihu, China. Ecol. Eng. 37 (2), 325–334.
Los, F.J., Villars, M.T., Van der Tol, M.W.M., 2008. A 3-dimensional primary production model (BLOOM/GEM) and its applications to the (southern) North Sea (coupled physical–chemical–ecological model). J. Mar. Syst. 74 (1–2), 259–294. MOE, 2017. Water Information System (WIS). Ministry of Environment, Republic of Korea water.nier.go.kr. MOLIT, 2017. Water Resources Management Information System (WAMIS). Ministry of Land Infrastructure and Transport, Republic of Korea www.wamis.go.kr. National Institute of Environmental Research, 2017. A Manual of Algal Bloom Warning System. Ministry of Environment Korea. O’Neil, J.M., Davis, T.W., Burford, M.A., Gobler, C.J., 2012. The rise of harmful cyanobacteria blooms: the potential roles of eutrophication and climate change. Harmful Algae 14, 313–334. Paerl, H.W., Huisman, J., 2009. Climate change: a catalyst for global expansion of harmful cyanobacterial blooms. Environ. Microbiol. Rep. 1 (1), 27–37. Paerl, H.W., Otten, T.G., 2013. Harmful cyanobacterial blooms: causes, consequences, and controls. Microb. Ecol. 65 (4), 995–1010. Park, K., Jung, H.S., Kim, H.S., Ahn, S.M., 2005. Three-dimensional hydrodynamic-eutrophication model (HEM-3D): application to Kwang-Yang Bay, Korea. Mar. Environ. Res. 60 (2), 171–193. Reynolds, C.S., 2006. Ecology of Phytoplankton (Ecology, Biodiversity and Conservation). Camb. Univ. Press, Cambridge, UK. Seo, D., Song, Y., 2015. Application of three-dimensional hydrodynamics and water quality model of the Youngsan River, Korea. Desalination Water Treat. 54 (13), 3712–3720. Seo, D., Kim, M., Ahn, J.H., 2012. Prediction of chlorophyll-a changes due to weir constructions in the Nakdong River using EFDC-WASP modelling. Environ. Eng. Res. 17 (2), 95–102. Shin, C.M., Na, E.H., Park, J.D., Park, J.H., Lee, S.W., Jeong, J.H., Rhew, D.H., Jeong, D.I., 2008. Application of Parameters and Coefficients of River Water Quality Model for TMDL Plan in Korea. National Institute of Environmental Research, NIER NO. 2008-29-979. Smith, V.H., 1986. Light and nutrient effects on the relative biomass of blue-green algae in lake phytoplankton. Can. J. Fish. Aquat. Sci. 43 (1), 148–153. Tang, C., Li, Y., Acharya, K., 2016. Modeling the effects of external nutrient reductions on algal blooms in hyper-eutrophic Lake Taihu, China. Ecol. Eng. 94, 164–173. Thomann, R.V., Mueller, J.A., 1987. Principles of Surface Water Quality Modeling and Control. Harper & Row Publishers. Trimbee, A.M., Prepas, E.E., 1987. Evaluation of total phosphorus as a predictor of the relative biomass of blue-green algae with emphasis on Alberta lakes. Can. J. Fish. Aquat. Sci. 44 (7), 1337–1342. Watson, S.B., McCauley, E., Downing, J.A., 1997. Patterns in phytoplankton taxonomic composition across temperate lakes of differing nutrient status. Limnol. Oceanogr. 42 (3), 487–495. Wong, K.T.M., Lee, J.H.W., Hodgkiss, I.J., 2007. A simple model for forecast of coastal algal blooms. Estuar. Coast. Shelf Sci. 74 (1–2), 175–196. Wu, G., Xu, Z., 2011. Prediction of algal blooming using EFDC model: case study in the Daoxiang Lake. Ecol. Model. 222 (6), 1245–1252. Yin, Z., Seo, D., 2016. Analysis of optimum grid determination of water quality model with 3-D hydrodynamic model using environmental fluid dynamics code (EFDC). Environ. Eng. Res. 21 (2), 171–179.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
Ecological Modelling xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Ecological Modelling journal homepage: www.elsevier.com/locate/ecolmodel
Research Paper
Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model Jaeyoung Kim, Tongeun Lee, Dongil Seo ∗ Department of Environmental Engineering, Chungnam National University, 99, Daehak-ro, Yuseong-gu, Daejeon, 34134, Republic of Korea
a r t i c l e
i n f o
Article history: Received 27 June 2017 Received in revised form 26 October 2017 Accepted 29 October 2017 Available online xxx Keywords: The Han River Algal bloom Toxic cyanobacteria Environmental Fluid Dynamics Code Water quality modeling
a b s t r a c t The lower part of the Han River, which flows through Seoul, Korea, experienced excessive toxic cyanobacterial growth in 2015. Modeling of algal bloom occurrence patterns in the lower part of this river was performed using the Environmental Fluid Dynamics Code (EFDC) to understand algal dynamics and thus better develop management alternatives. For a 71 km long river section, 1175 horizontal 2D grid elements were developed. This grid system was determined adequate, as the maximum values of the Courant–Friedrichs–Lewy condition and orthogonality deviation were 0.5 and 20.1, respectively. Chlorophyll-a (Chl-a) was chosen as the primary indicator for the likelihood of algal blooms. Calibration and verification of EFDC were performed by comparing the model results to three years of field data collected from 2013 to 2015. Calibration accuracy was verified not only for physical variables, including the mean water level and temperature, but also for other water quality variables in various locations of the study area. To improve the prediction accuracy of Chl-a, three dominant groups of algae were considered: diatoms, green algae, and cyanobacteria. The optimum growth temperature ranges were selected based on field data for the study area. It was found necessary to apply different maximum growth rates for algal groups for the upstream and downstream regions of the study area to appropriately reflect field observations. This result indicates that more than three algal groups need to be included to improve Chl-a calibration accuracy for the study area, yet the current EFDC model can consider only up to three phytoplankton groups. Although this problem could be overcome by assigning different maximum growth rates for different regions, it may be necessary to improve EFDC so that it can include more phytoplankton groups. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The lower section of the Han River, which passes through Seoul City, Republic of Korea (hereafter Korea), experienced a relatively severe algal bloom in 2015. The Korean Government developed a four- level algal warning system by measuring Chlorophyll-a (Chla) concentrations and the number of toxic cyanobacteria cells per milliliter. During July and September in 2015, Chl-a concentrations and the number of toxic cyanobacteria were measured at levels greater than 25 mg/m3 and 5000 cells/ml, respectively. A second level algae warning based on the Korean system was declared in several locations of the river. An analysis that clearly understands the causes and effects of algal blooms and water pollution within a water body is often very difficult to achieve due to the many factors involved. For example, Chapra (1997) reported that a sufficient level of nutrients in
∗ Corresponding author. E-mail addresses: [email protected], [email protected] (D. Seo).
the water body for phytoplankton to feed on, an appropriate water temperature for algal growth, and an adequate amount of light for photosynthesis are all essential conditions for an algal bloom to occur. Reynolds (2006) reported the growth of algae is affected by competition between different algal groups, predation in the food chain, the degree of pollution, and the flow characteristics of the river are important factors. In the case of the lower Han River, many factors can affect algal blooms. It would be difficult to conclude that only one or a few factors would be responsible. Various methods may be applied to control algal blooms in surface waters. Physical methods, such as dilution or mechanical mixing may be used, but require enormous amounts of energy. Chemical methods such as algaecides or coagulants can be used, but are usually applied to relatively small and stagnant water bodies and are not used for rivers. Biological methods, such as the introduction of predators or constructed wetlands also can be used. All of these methods only deal with surficial symptoms of algal blooms and only last a limited period of time. Therefore, it is necessary to identify feasible water quality management alternatives that can be applied efficiently (Anderson, 2009).
https://doi.org/10.1016/j.ecolmodel.2017.10.015 0304-3800/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10 2
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
Water quality modeling is a powerful tool for the comprehensive interpretation of complex water quality interactions (Ambrose et al., 2009). Many different models have been applied to water bodies to analyze which countermeasures may be effective for mitigating algal blooms (Chen and Mynett, 2006; Wong et al., 2007; Los et al., 2008; Leon et al., 2011; Seo et al., 2012; Camacho et al., 2014; Jian et al., 2014). The Environmental Fluid Dynamics Code (EFDC) (Hamrick, 1992) is a versatile surface water model that has been used widely for various water body types such as rivers, lakes, wetlands, dams, estuaries, and coasts for environmental assessment and management. Park et al. (2005) applied the EFDC water quality model in Gwangyang Bay, Korea. Li et al. (2011) applied the concept of water age using EFDC to understand the effect of the migration of dissolved substances on the water quality of Lake Taihu, China. Wu and Xu (2011) and Tang et al. (2016) used EFDC to predict the occurrence of algal blooms in Daoxiang and Taihu lakes in China, respectively. Seo and Song (2015) conducted a three-dimensional hydrodynamic and water quality modeling analysis using EFDC in the Youngsan River, the fourth largest river in Korea. Yin and Seo (2016) analyzed an optimal grid selection for EFDC water quality modeling of the Ara Navigation Channel in Korea. Lee et al. (2017) conducted a study to evaluate the impact of installing of a new sewage treatment plant on the water quality of the Galing River, Malaysia. Although many studies have been carried out, an accu-
rate prediction of the occurrence of algal blooms has been difficult to achieve. Since algae are not simple substances but living organisms, their population dynamics are difficult to explain with simple mass balance equations or chemical processes. Sufficient data is rarely available to estimate boundary conditions, define the parameters of water quality variables, or solve the various algal growth related problems required to perform the modeling. This study aims to improve the prediction accuracy for harmful algal bloom in the lower Han River, Seoul, Korea so that the model can be used in decision making processes in developing management alternatives. The growth-related parameters of algae groups were selected from the literature and from field observations in the study area, and then further adjusted in the calibration of EFDC. To improve model prediction accuracy for algal blooms in this study, different maximum growth rates were applied for algal groups in regions of the river influenced by different water quality conditions. 2. Materials and methods 2.1. Study area The Han River is the second largest river in Korea. It is 483 km long and its basin area is 34,428 km2 . This study focuses on a lower 71 km section of the river that passes through Seoul from the Pal-
Fig. 1. Monitoring stations and important in-stream structures in the study area.
Fig. 2. Fraction of algae in upstream and downstream.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
dang Dam to the Jeonryu water level station as shown in Fig. 1. There are five water level stations, fourteen water quality monitoring stations, and nine algae monitoring stations within the study site. Two fixed weirs, the Jamsil Weir (hereafter Weir 1) and Singok Weir (hereafter Weir 2) have been installed to maintain the water level and prevent salt water intrusion into the river, as Weir 2 is affected by tidal movement from the western coast of Korea. Between the two weirs, four major wastewater treatment plants in Seoul are operational and discharge effluent directly into the Han River.
2.2. Chl-a and cyanobacteria in the Han River during the period between 2013–2015 Fig. 2 shows the algae fractions observed in the study site between 2013 and 2015, both in the upstream region between Paldang Dam and Weir 1 and in the downstream region between Weir 1 and Weir 2. Diatoms were shown to be dominant during most of each year studied. Green algae and cyanobacteria growth only became significant during each summer. In 2015, cyanobacteria became dominant in the summer, and severe algal blooms occurred between August and September. Some of the most common toxinproducing cyanobacteria include the N-fixing genera: Anabaena, Aphanizomenon, Cylindrospermopsis, Lyngbya, Nodularia, Oscillatoria, and Trichodesmium; and non-N-fixing genera: Microcystis and
3
Planktothrix (Paerl and Otten, 2013). The Ministry of Environment developed the Korean Algae Warning System in 1997, which was updated in 2015. In the new Korean Algae Warning System, the cell numbers of Anabaena, Microcystis, Aphanizomenon, and Oscillatoria are used to trigger different warning levels from 2016 onwards (National Institute of Environmental Research, 2017). Nitrogen to phosphorus (N/P) ratios of twenty or greater generally indicate a phosphorus limited system for algal growth while N/P ratios of five or less indicate nitrogen limited systems (Thomann and Mueller, 1987). As bodies of freshwater become enriched with nutrients, especially phosphorus, there is often a shift in the phytoplankton community towards a dominance by cyanobacteria (Smith, 1986; Trimbee and Prepas, 1987; Watson et al., 1997; Paerl and Huisman, 2009; O’Neil et al., 2012). In 2015, the N/P ratio in the Han River was 70 upstream of Weir 1 and 25 downstream of Weir 1. The nutrient limiting algal growth in the system was identified as phosphorus (Fig. 3). Fig. 4 shows the number of cyanobacteria cells and total phosphorus (TP) concentrations upstream of Weir 1 in 2015. In this region, phosphorus concentrations ranged from 0.02 to 0.08 mg/L, and the maximum toxic cyanobacteria number was about 10,000 cells/ml, while the maximum non-toxic cyanobacteria number was about 39,000 cells/ml. On the other hand, the concentration of phosphorus in the downstream region, between Weir 1 and Weir 2, ranged from 0.05 to 0.43 mg/L. In this area, toxic cyanobacteria were dominant and
Fig. 3. N/P ratio of the lower Han River.
Fig. 4. Toxic and non-toxic cyanobacteria in upstream and downstream areas.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
4
Fig. 5. Conceptual diagram of the EFDC model.
its maximum count was about 40,000 cells/ml. Many researchers have reported that summer phytoplankton numbers are dominated by cyanobacteria at phosphorus concentration between approximately 0.1 and 1.0 mg/L (Trimbee and Prepas, 1987; Jensen et al., 1994; Watson et al., 1997; Downing et al., 2001; O’Neil et al., 2012). It seems that toxic cyanobacteria grow better in the high phosphorus concentrations in the summer period than non-toxic cyanobacteria. It is suggested that the growth pattern of each algal group can be different and may vary depending on the region of study. 2.3. Model development and boundary conditions The EFDC three-dimensional hydrodynamic and water quality model was used to simulate water quality and algal blooms in the study area. The model includes water quality, sediment transport, and toxicity modules as shown in Fig. 5. The EFDC computational grid was developed by the curvilinear orthogonal coordinate method. 1175 horizontal grid points were selected by a trial and error method. In the selected grid system, the Courant–Friedrichs–Lewy (CFL) condition (Courant et al., 1928) was estimated to be less than 0.50 and the average value of orthogonality deviation was 3.14. Calibration and verification of the model to observed data were performed for the three years period from 2013 to 2015. As shown in Fig. 1, the boundary conditions included thirteen tributaries along with the headwater from the discharge of the Paldang Dam, five water intake systems, five wastewater treatment plants, and open boundary conditions that reflect the tidal movement along the west coast. The necessary field data for this study were obtained from Korean governmental database management systems. The
Table 1 Summary of the kinetic coefficients employed in the current model application. Parameter
Value
Optimal temperature for algal growth (◦ C) Basal metabolism rate for algae [day−1 ] Predation rate for algae [day−1 ] Half-saturation constant for nitrogen uptake of algae [mg L−1 ] Half-saturation constant for phosphorus uptake of algae [mg L−1 ] Half-saturation constant for silica uptake of diatoms [mg L−1 ] Algae settling rate [m day−1 ] Decay rate of organic carbon [day−1 ] Decay rate of organic phosphorus [day−1 ] Decay rate of organic nitrogen [day−1 ] Settling velocity of particulate organic matter [m day−1 ] Benthic flux rate of phosphate [g m−2 day−1 ] Benthic flux rate of ammonia nitrogen [g m−2 day−1 ] Benthic flux rate of nitrate [g m−2 day−1 ]
25, 13–15, 25a 0.06 0.1 0.01 0.001 0.05 0.215 0.005, 0.075, 0.1b 0.005, 0.075, 0.1b 0.005, 0.075, 0.015b 2.19 0.003 0.15 0.03
a
For cyanobacteria, diatoms, and green algae, respectively. b For refractory particulate, labile particulate, and dissolved organic matter, respectively.
Water Information System (MOE, 2017) was used for water quality data and the Water Resources Management Information System (MOLIT, 2017) was used for water quantity data. Table 1 shows selected calibrated parameters used in this study. The parameters were initially chosen based on previous research results (Bowie et al., 1985; Cole and Wells, 2006; Shin et al., 2008) and then adjusted further to improve the calibration accuracy in the cur-
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
5
Table 2 Scenarios used in this study. Maximum Growth rate [day−1 ]
Cyanobacteria
(a) Overall value (b) Different value for each algal group (c) Different value for per region and algal group
2.5 2.5 Upper 3.5
Diatoms
Green Algae
1.0 Upper 1.5
Lower 2.5
3.0 Upper 4.0
Lower 1.1
Lower
rent study. The optimal temperatures for algal growth for the three different groups in Table 1 were selected based on field observations of the study area, as shown in Fig. 2. Although this study did not simulate sediment transport, nutrients released from bottom sediments were considered as shown in Table 1.
Table 3 AME and RMSE of the water level for selected locations of the Han River.
2.4. Development of scenarios for algae prediction
regarding algal growth not considered in the model. To increase the accuracy of the algal growth simulation in this study, different maximum growth rates for different algal groups were assigned for each region. The spatial variability reflects various factors affecting algal growth.
It is difficult to simulate the occurrence of algal blooms accurately for complex natural water systems. Many conditions affect the growth and loss of algae in a water body and there is often a lack of appropriate data. In EFDC, the overall dynamics and growth of algae in a cell are determined by Eqs. (1) and (2). EFDC can simulate up to three groups of algae. Because the number of simulated algae groups is limited, EFDC cannot reflect various factors
Station
WL1
WL2
WL3
WL4
AME [m] RMSE [m]
0.14 0.29
0.29 0.41
0.19 0.36
0.17 0.34
∂Bx ∂ WBx = (Px − BMx − PRx )Bx + (WSx · Bx ) + V ∂t ∂z
(1)
Bx = algal biomass of algal group x (g C m−3 )
Fig. 6. Calibration results of water level and water temperature for selected locations of the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
6
Table 4 AME and RMSE for temperature in selected locations of the Han River. Station
Q2
Q3
Q4
Q5
Q6
Q7
Q8
Q9
Q10
Q11
Q12
Q13
AME [◦ C] RMSE [◦ C]
0.63 0.85
0.69 0.86
0.83 1.04
0.99 1.16
1.07 1.33
1.10 1.36
1.10 1.37
1.32 1.71
1.55 1.94
1.48 1.85
1.37 1.66
1.68 2.07
t = time (day) Px = production rate of algal group x (day−1 ) BMx = basal metabolism rate of algal group x (day−1 ) PRx = predation rate of algal group x (day−1 ) WSx = positive settling velocity of algal group x (m day−1 ) WBx = external loads of algal group x (g C day−1 ) V = cell volume (m3 ) Px = PMx ·f 1 (N)·f 2 (I)·f 3 (T)
perature and light intensity were chosen from observations Fig. 2 for each group of algae and these were shown in Table 1. 3. Results and discussion 3.1. Hydrodynamic model calibration (2)
PMx = maximum growth rate under optimal conditions for algal group x (day−1 ) f1 (N) = effect of suboptimal nutrient concentration (0–1) f2 (I) = effect of suboptimal light intensity (0–1) f3 (T) = effect of suboptimal temperature (0–1) Three different calibrations were performed and compared to field observations to determine their accuracy. Table 2 shows the three calibration scenarios: a) the same maximum growth rate for all algae groups, b) different maximum growth rate for three algal groups and c) different growth rates for different regions (upper and lower stream areas from Weir 1 in Fig. 1). The optimum tem-
Calibration and verification of the hydrodynamic model were carried out using water level and water temperature data for the Han River collected for three years from 2013 to 2015. For water level, data from four monitoring stations in the study site were used for calibration. For water temperature, calibration was carried out for twelve water quality monitoring stations. Fig. 6 shows the calibration results for all water level stations and some selected stations for water temperature. Tables 3 and 4 show the results of the analysis using both absolute mean error (AME) and root mean square error (RMSE) to evaluate the accuracy of the simulated results. The mean values for the water level were 0.20 m and 0.35 m respectively. In the case of temperature, the mean values were 1.15 ◦ C and 1.43 ◦ C respectively.
Fig. 7. Water quality calibration results for scenarios (a) and (b) in the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
7
Fig. 7. (Continued)
3.2. Algae and water quality calibration 3.2.1. Scenarios (a) and (b); overall and different growth rate for each algal group The water quality calibration was performed at 12 monitoring stations. Calibrated water quality variables include total organic carbon (TOC), TP, phosphorous-phosphate (PO4 -P), total nitrogen (TN), nitrogen-ammonia (NH3 -N), nitrogen-nitrate (NO3 -N), dis-
solved oxygen (DO) and Chl-a. Chl-a was chosen as a primary indicator for overall algal bloom for all the scenario in this study as this is only available in the national water quality monitoring database system (MOE, 2017). Fig. 7 shows the calibration results for scenarios (a) and (b) in selected locations and for four water quality variables. It seems that the calibration results for TP and TN are relatively more accurate than other water quality variables in this study. In the case of TOC and Chl-a, calibration results for
Table 5 AME analysis of the water quality calibration results. AME
Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Average
Scenario (a)
Scenario (b) 3
Scenario (c) 3
TOC [mg/L]
Chl-a [mg/m ]
TOC [mg/L]
Chl-a [mg/m ]
TOC [mg/L]
Chl-a [mg/m3 ]
0.37 0.37 0.40 0.67 0.39 0.48 0.47 0.62 0.60 0.66 1.00 0.61 0.55
3.11 5.63 6.36 7.57 7.71 7.23 7.65 10.73 13.11 14.00 17.34 16.50 9.75
0.36 0.35 0.36 0.75 0.32 0.44 0.46 0.70 0.35 0.46 1.32 0.60 0.54
2.92 5.25 6.10 6.49 6.12 4.71 4.70 7.90 10.70 12.28 16.54 16.19 8.33
0.37 0.36 0.37 0.68 0.33 0.38 0.39 0.60 0.30 0.34 1.10 0.45 0.47
3.03 5.12 5.73 6.43 5.91 4.96 4.96 6.94 10.40 11.86 16.85 15.69 8.16
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model
ARTICLE IN PRESS
ECOMOD-8294; No. of Pages 10
J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
8 Table 6 RMSE analysis of the water quality calibration results. RMSE
Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Average
Scenario (a)
Scenario (b)
Scenario (c)
TOC [mg/L]
Chl-a [mg/m3 ]
TOC [mg/L]
Chl-a [mg/m3 ]
TOC [mg/L]
Chl-a [mg/m3 ]
0.67 0.63 0.63 1.15 0.57 0.63 0.62 0.91 0.73 0.78 1.35 0.71 0.78
3.98 8.01 10.40 11.47 10.24 8.95 9.63 14.01 16.89 17.43 22.67 20.91 12.88
0.66 0.62 0.60 1.20 0.50 0.61 0.59 0.96 0.46 0.57 1.59 0.69 0.75
3.89 7.88 10.43 11.19 9.24 6.47 6.71 12.60 16.34 17.92 23.60 23.80 12.51
0.66 0.62 0.60 1.11 0.51 0.54 0.54 0.83 0.40 0.44 1.40 0.54 0.68
4.06 7.20 9.54 10.35 8.50 6.01 6.18 10.93 15.57 16.94 22.38 22.65 11.69
upstream areas were more accurate than downstream areas. Scenario (b) showed better reproducibility for TOC and Chl-a results, although greater errors result in the downstream area especially for high Chl-a concentrations. As illustrated in Fig. 4, the growth of the three algal groups in the study area was different in each region. These results suggest that it is necessary to consider growth characteristics separately for each group to achieve better accuracy in the modeling of Chl-a.
3.2.2. Scenario (c): different maximum growth rate per each region and algal group Fig. 8 shows the simulation results for scenario (c). Calibration results for NH4 -N, NO3 -N, PO4 -P, and DO were added to show the complete range of water quality variables considered in this study. The calibration accuracy for TOC and Chl-a showed greater improvement than that for TN and TP. Tables 5 and 6 show a statistical analysis of TOC and Chl-a calibration results. The two tables show that the prediction errors become smaller as the growth
Fig. 8. Water quality calibration results for scenario (c) in the Han River.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
9
Fig. 8. (Continued)
rates of the algal groups become more specific. This result suggests that it may be necessary to include more algal groups or genera to improve the prediction accuracy as shown in Fig. 4. Chapra et al. (2017) also suggested using four phytoplankton functional groups by dividing cyanobacteria group to N-fixers and non-Nfixers to improve prediction accuracy of cyanobacterial harmful algal blooms in his QUALIDAD model. Though it was possible to improve the calibration accuracy by subdividing the growth rates for different conditions, calibration of Chl-a in the downstream area, for example in Q13, did not seem to reflect the field data, especially for more extreme conditions. This is interpreted as the limit of data that can be reflected by the model, as well as the effects of other factors on the growth of algae that EFDC does not consider. 4. Conclusion This study was carried out to simulate the changes in water quality associated with algal blooms in the Han River using the EFDC model. For the simulation period from 2013 to 2015, calibration of the model was performed with respect to water level, water temperature, and water quality variables in various locations of the study area. The aim of this study was to improve the reproducibility of algal growth as reflected by Chl-a using growth-rate-related parameters of multiple algal groups applied to the Han River. The dominant groups of algae in the study area were diatoms, green algae, and cyanobacteria. We observed that non-toxic
cyanobacteria were more significant upstream, where the phosphorus concentration was lower. On the other hand, toxic cyanobacteria were more significant downstream, which had a higher concentration of phosphorus. The results of this study suggest that the degree of algal growth differs depending on the environment and region. Therefore, to reflect this phenomenon better, the maximum growth rates of algal groups were classified for different sections and entered into the model accordingly. We analyzed three scenarios to compare the prediction accuracy of Chl-a. The application of different maximum growth rates for different algal groups was found to be more reproducible than was applying the same growth rate. Moreover, accuracy was the highest when the maximum growth rate for algal groups was used separately for upstream and downstream areas of Weir 1 of the study area. Although the overall model accuracy was improved compared to the original model, the error was larger downstream of Weir 1. For extremely high concentrations of Chl-a, algal growth was not accurately reflected by the model. This implies that other factors not considered in EFDC affect the growth of algae. This result suggests that the current algal simulation method in EFDC may need to be modified to improve the prediction accuracy, especially for large harmful algal blooms. The predicted results for water quality and the occurrence of algal blooms in the Han River made using these assumptions are considered significant, and continued research to confirm these results is needed.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015
G Model ECOMOD-8294; No. of Pages 10 10
ARTICLE IN PRESS J. Kim et al. / Ecological Modelling xxx (2017) xxx–xxx
Acknowledgement This study was supported by Water Circulation Policy Division of Seoul Metropolitan Government. References Ambrose Jr., R.B., Wool, T.A., Barnwell Jr., T.O., 2009. Development of Water Quality Modeling in the United States. Environ. Eng. Res. 14, 200–210. Anderson, D.M., 2009. Approaches to monitoring, control, and management of harmful algal blooms (HABs). Ocean Coast. Manag. 52 (7), 342–347. Bowie, G.L., Mills, W.B., Porcella, D.B., Campbell, C.C., Pagenkopf, J.R., Rupp, G.L., Johnson, K.M., Chan, R.W.H., Gherini, S.A., Chamberlin, C.E., 1985. Rates, Constants and Kinetics Formulations in Surface Water Quality Modeling, second ed. Environment Research Laboratory, Environmental Protection Authority, EPA/600/385/040. Camacho, R.A., Martin, J.L., Watson, B., Paul, M.J., Zheng, L., Stribling, J.B., 2014. Modeling the factors controlling phytoplankton in the St. Louis Bay estuary, Mississippi and evaluating estuarine responses to nutrient load modifications. J. Environ. Eng. 141 (3), 04014067. Chapra, S.C., Boehlert, B., Fant, C., Bierman Jr., V.J., Henderson, J., Mills, D., Mas, D.M.L., Rennels, L., Jantarasami, L., Martinich, J., Strzepek, K.M., Paerl, H.W., 2017. Climate change impacts on harmful algal blooms in U.S. freshwaters: a screening-level assessment. Environ. Sci. Technol. 51 (16), 8933–8943. Chapra, S.C., 1997. Surface Water-Quality Modeling. McGraw-Hill Companies, Inc. Chen, Q., Mynett, A.E., 2006. Modeling algal blooms in the Dutch coastal waters by integrated numerical and fuzzy cellular automata approaches. Ecol. Model. 199 (1), 73–81. Cole, T.M., Wells, S.A., 2006. CE-QUAL-W2: A Two-dimensional, Laterally Averaged, Hydrodynamic and Water Quality Model, Version 3.5. Instruction Report EL-06-1. US Army Engineering and Research Development Center, Vicksburg, MS. Courant, R., Friedrichs, K., Lewy, H., 1928. Über die partiellen differenzengleichungen der mathematischen. Physik. Math. Ann. 100 (1), 32–74. Downing, J.A., Watson, S.B., McCauley, E., 2001. Predicting cyanobacteria dominance in lakes. Can. J. Fish. Aquat. Sci. 58 (10), 1905–1908. Hamrick, J.M., 1992. A three-dimensional environmental fluid dynamics computer code: theoretical and computational aspects. In: Special Report. Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA, pp. 317. Jensen, J.P., Jeppesen, E., Olrik, K., Kristensen, P., 1994. Impact of nutrients and physical factors on the shift from cyanobacterial to chlorophyte dominance in shallow Danish lakes. Can. J. Fish. Aquat. Sci. 51 (8), 1692–1699. Jian, L., Zhongwu, J., Wenjun, Y., 2014. Numerical modeling of the Xiangxi River algal bloom and sediment-related process in China. Ecol. Inform. 22, 23–35. Lee, I., Hwang, H., Lee, J., Yu, N., Yun, J., Kim, H., 2017. Modeling approach to evaluation of environmental impacts on river water quality: a case study with Galing River, Kuantan, Pahang, Malaysia. Ecol. Model. 353, 167–173. Leon, L.F., Smith, R.E., Hipsey, M.R., Bocaniov, S.A., Higgins, S.N., Hecky, R.E., Antenucci, J.P., Imberger, J.A., Guildford, S.J., 2011. Application of a 3D hydrodynamic–biological model for seasonal and spatial dynamics of water quality and phytoplankton in Lake Erie. J. Great Lakes Res. 37 (1), 41–53. Li, Y., Acharya, K., Yu, Z., 2011. Modeling impacts of Yangtze River water transfer on water ages in Lake Taihu, China. Ecol. Eng. 37 (2), 325–334.
Los, F.J., Villars, M.T., Van der Tol, M.W.M., 2008. A 3-dimensional primary production model (BLOOM/GEM) and its applications to the (southern) North Sea (coupled physical–chemical–ecological model). J. Mar. Syst. 74 (1–2), 259–294. MOE, 2017. Water Information System (WIS). Ministry of Environment, Republic of Korea water.nier.go.kr. MOLIT, 2017. Water Resources Management Information System (WAMIS). Ministry of Land Infrastructure and Transport, Republic of Korea www.wamis.go.kr. National Institute of Environmental Research, 2017. A Manual of Algal Bloom Warning System. Ministry of Environment Korea. O’Neil, J.M., Davis, T.W., Burford, M.A., Gobler, C.J., 2012. The rise of harmful cyanobacteria blooms: the potential roles of eutrophication and climate change. Harmful Algae 14, 313–334. Paerl, H.W., Huisman, J., 2009. Climate change: a catalyst for global expansion of harmful cyanobacterial blooms. Environ. Microbiol. Rep. 1 (1), 27–37. Paerl, H.W., Otten, T.G., 2013. Harmful cyanobacterial blooms: causes, consequences, and controls. Microb. Ecol. 65 (4), 995–1010. Park, K., Jung, H.S., Kim, H.S., Ahn, S.M., 2005. Three-dimensional hydrodynamic-eutrophication model (HEM-3D): application to Kwang-Yang Bay, Korea. Mar. Environ. Res. 60 (2), 171–193. Reynolds, C.S., 2006. Ecology of Phytoplankton (Ecology, Biodiversity and Conservation). Camb. Univ. Press, Cambridge, UK. Seo, D., Song, Y., 2015. Application of three-dimensional hydrodynamics and water quality model of the Youngsan River, Korea. Desalination Water Treat. 54 (13), 3712–3720. Seo, D., Kim, M., Ahn, J.H., 2012. Prediction of chlorophyll-a changes due to weir constructions in the Nakdong River using EFDC-WASP modelling. Environ. Eng. Res. 17 (2), 95–102. Shin, C.M., Na, E.H., Park, J.D., Park, J.H., Lee, S.W., Jeong, J.H., Rhew, D.H., Jeong, D.I., 2008. Application of Parameters and Coefficients of River Water Quality Model for TMDL Plan in Korea. National Institute of Environmental Research, NIER NO. 2008-29-979. Smith, V.H., 1986. Light and nutrient effects on the relative biomass of blue-green algae in lake phytoplankton. Can. J. Fish. Aquat. Sci. 43 (1), 148–153. Tang, C., Li, Y., Acharya, K., 2016. Modeling the effects of external nutrient reductions on algal blooms in hyper-eutrophic Lake Taihu, China. Ecol. Eng. 94, 164–173. Thomann, R.V., Mueller, J.A., 1987. Principles of Surface Water Quality Modeling and Control. Harper & Row Publishers. Trimbee, A.M., Prepas, E.E., 1987. Evaluation of total phosphorus as a predictor of the relative biomass of blue-green algae with emphasis on Alberta lakes. Can. J. Fish. Aquat. Sci. 44 (7), 1337–1342. Watson, S.B., McCauley, E., Downing, J.A., 1997. Patterns in phytoplankton taxonomic composition across temperate lakes of differing nutrient status. Limnol. Oceanogr. 42 (3), 487–495. Wong, K.T.M., Lee, J.H.W., Hodgkiss, I.J., 2007. A simple model for forecast of coastal algal blooms. Estuar. Coast. Shelf Sci. 74 (1–2), 175–196. Wu, G., Xu, Z., 2011. Prediction of algal blooming using EFDC model: case study in the Daoxiang Lake. Ecol. Model. 222 (6), 1245–1252. Yin, Z., Seo, D., 2016. Analysis of optimum grid determination of water quality model with 3-D hydrodynamic model using environmental fluid dynamics code (EFDC). Environ. Eng. Res. 21 (2), 171–179.
Please cite this article in press as: Kim, J., et al., Algal bloom prediction of the lower Han River, Korea using the EFDC hydrodynamic and water quality model. Ecol. Model. (2017), https://doi.org/10.1016/j.ecolmodel.2017.10.015