Chagan Lake serves as an irrigation storage reservoir for the Qianguo Irrigation Area and an important ecological barrier in western Jilin. The coupled TUFLOW-FV and Aquatic Ecodynamic (AED2) models were used to simulate the hydrodynamic and water quality of Chagan Lake, and propose the water diversion scheme that could improve the water quality to reach Grade III and maintain the ecological water level. The simulation results showed a satisfactory agreement with observations. The total carrying loads of NH3-N, total nitrogen (TN) and total phosphorus (TP) for Chagan Lake were 1,147.6, 3,686.2 and 100.8 t from May to October. The range of the minimum amounts of water diversion to keep the water quality as Grade III and maintain the maximum ecological water level of 131.5 m for TN, TP were separately [32.60, 49.84, 57.19, 63.70, 70.91], [117.25, 135.26, 168.17, 190.65, 218.32] million m3 and the corresponding reduction amounts of farmland drainage for TN, TP were separately [4.03, 0, 0, 0, 0], [73.08, 61.88, 50.23, 40.94, 31.98] million m3 under the rainfall guarantee rates of 10%, 20%, 50%, 75%, 90%, respectively. The simulation results provide a scientific basis for the water quality improvement and ecological water supplement required for the interconnected river–lake system network (IRLSN) in Western Jilin Province.
Of all the natural resources required for economic development, water is the most essential, especially in arid and semi-arid areas (Meng et al. 2009). Due to rapid economic development and urbanization, water scarcity and pollution have significantly compromised the availability and integrity of water resources. An absence of effective management practices and regulations has exacerbated this problem.
Chagan Lake, a typical shallow soda/saline lake in the semi-arid region of southwestern Songnen Plain, northeast China, serves as an important ecological barrier in western Jilin Province and as the most important fishery base in Jilin Province. The lake's ecological environment has worsened over time, decreasing in water area and exhibiting an expansion of alkali spots and saline–alkali land from 1988 to 2004. Shen & Zhang (2009) concluded that enclosure development and construction in Chagan Lake and the implementation of water conservancy projects had resulted in serious ecological degradation. Several studies have also identified excessive total nitrogen (TN), total phosphorus (TP) and CODMn (permanganate index) pollution levels in the lake (Dai & Tian 2011). Chagan Lake is connected to the Songhua River by irrigation channels, through which the lake receives a large amount of water from the Second Songhua River and from Qianguo Irrigation Area agricultural drainage, which is also sourced from the river (Zhu et al. 2012). The Second Songhua River features contamination problems, and the farmland drainage of the Qianguo Irrigation Area contains high concentrations of salt, alkali, nitrogen and phosphorus, which threaten the ecological health and security of the lake. The Songyuan irrigation district, which will draw water from the Second Songhua River to exploit large saline wastelands in Qianguo, Qian'an, and Daan, will ameliorate moderate- and low-productivity land and restore degraded grassland to form the national commodity grain base of the comprehensive agricultural irrigation project. Chagan Lake serves as one of main centers of the irrigation district drainage areas. Thus, with the development of the saline–alkali land, large volumes of farmland drainage with high concentrations of TN, TP, and salt are projected to flow into Chagan Lake and will inevitably affect the water quality and ecological security of Chagan Lake. Therefore, effective water quality simulation and prediction for Chagan Lake following the development of the Songyuan irrigation district must be undertaken to ensure water safety for the purposes of sustainable development and public health.
Water diversion has been a successful strategy for improving water quality, as has been undertaken for many Chinese lakes such as Taihu, Chaohu, and Dianchi. The restoration practice suggests that diversion is a suitable method for managing degraded water quality conditions in lakes (Yang & Liu 2010; Chen et al. 2011; Li et al. 2011).
With the rapid development of computer technologies and mathematical techniques, especially numerical methods, numerous water quality models have been developed (Hipsey et al. 2015), and are increasingly useful tools to simulate and predict the levels, distributions, and risks of chemical pollutants in a given water body; they have brought about a virtual revolution in the approach to solving problems in the areas of water resources and environmental engineering. More than 100 surface water quality models have been developed up to now (Janssen et al. 2015). Cao & Zhang (2006) classified models based on water body types, model-establishing methods, water quality coefficient, water quality components, model property, spatial dimension, and reaction kinetics. The early modeling effort can be traced back to the 1920s when the Ohio River Commission performed a comprehensive survey to investigate the pollution and self-recovery mechanism of the Ohio River. In this investigation, one of the most significant water quality models, the Streeter–Phelps biochemical oxygen demand/dissolved oxygen (BOD/DO) model, was created (Streeter & Phelps 1925). Then the one-, two- and three-dimensional (3D) models including the N and P cycling system, phytoplankton and zooplankton system and focusing on the relationships between biological growth rates and nutrients, sunlight and temperature, and phytoplankton and the growth rate of zooplankton, were developed, and the commonly used platforms QUAL2E/2 K (Brown & Barnwell 1987), MIKE11/21 (DHI 2000, 2005), CE-QUAL-W2 (Cole & Buchak 1995), Delft3D (Delft Hydraulics 2003), WASP (Tim et al. 2001), EFDC (Hamrick 1996), and CAEDYM (Hipsey & Hamilton 2008). Specifically for lakes, Mooij et al. (2010) presented a broad variety of modeling approaches, such as static models (steady state and regression models), complex dynamic models (CAEDYM, CE-QUAL-W2, Delft 3D-ECO, LakeMab, MyLake, PCLake, PROTECH, SALMO), structurally dynamic models and minimal dynamic models, and suggested that multiple modeling approaches can help develop an integrative view on the functioning of lake ecosystems. So due to the constraints of current models and the complexity of a lake's water environment, the development of coupled hydrodynamic–water quality models based on the transport and transformation processes of pollutants and flow is an important research aspect of water resources management and water environment protection.
The aims of this research have been to establish a coupled hydrodynamic–water quality model of Chagan Lake using a novel finite-volume model approach to assess the impacts of farmland drainage on the lake's water quality, and to calculate the pollutant-carrying capacity of NH3-N, TN and TP that would meet the target water quality as Grade III. The minimum amount of water diversion from the Second Songhua River in the Hadashan diversion project channel is then estimated under different rainfall frequencies to keep the water quality as Grade III and the ecological water level. The results are of significance to the improvement of the Chagan Lake water environment, and provide a scientific basis for reasonable water resources allocation for the ‘Interconnected River–Lake System Network in Western Jilin Province’ project.
MATERIALS AND METHODS
Chagan Lake, located in the southwest area of the Songnen Plain, Northeast China (124°03′28″ – 124°30′59″ E, 45°05′42″- 45°25′50″ N; Figure 1), is the tenth largest lake in China. The lake covers a mean surface area of 372 km2, has a mean depth of 1.52 m, and features a full storage capacity of 5.98 × 108 m3. The regional climate is categorized as a mid-temperate zone. The annual mean air temperature for the area is 4.5 °C, and the local freezing period lasts from October to the following May, with the extreme lowest temperature of −36.1 °C recorded in the winter of 1953. Chagan Lake is located in the semi-arid area of western Jilin Province. Hence, annual runoff levels are low, and average runoff depths of 5–10 mm have been recorded for several years. The lake is exploited by fishery activities, which yield more than 6,000 t of fish per year. The commercial species migrate to the lake for feeding and are caught during the winter fish-hunting festival each year (Zhu et al. 2012).
The Second Songhua, Huolin, Tao'er and Nenjiang Rivers are tributaries of Chagan Lake, and natural precipitation and groundwater serve as auxiliary water supplies for the lake. Farmland drainage from the Qianguo irrigation district serves as the main source of water for the lakes in the region (Duan et al. 2008). Farmland drainage from the Qianguo irrigation district includes high concentrations of salt, alkali, nitrogen and phosphorus, which threaten the ecological health and security of the aquatic environments.
Now ‘the Interconnected River–Lake System Network in Jilin Province’ is being implemented, which mainly refers to making full use of crossing water resources in flood periods, such as filling some lakes that have ecological, economic value, to protect and restore some important and valuable wetland, such as Xianghai, Momoge, and Chagan Lake Nature Reserve Wetland. But the amount of water diversion has not been determined in this project, only the suitable connecting method was proposed.
Model description and setup
TUFLOW-FV (BMT WBM 2013) was applied in this study to simulate the hydrodynamics, mixing and transport of the Chagan Lake. This model is a 3D flexible-mesh (finite volume) hydrodynamic model that can be used for modeling a diverse array of inland and coastal water bodies. The finite volume method solves the conservative integral form of the non-linear shallow water equations (i.e. assuming that pressure varies hydrostatically with depth), including viscous flux terms and source terms for Coriolis force, bottom-friction and various surface and volume stresses. The scheme is also capable of simulating the advection and dispersion of multiple scalar constituents (e.g. salinity, temperature) within the model domain. Surface momentum exchange and heat dynamics are solved internally within the model from available meteorological boundary condition data. In the current application, turbulent mixing of momentum and scalars has been calculated using the Smagorinsky scheme in a horizontal plane and through coupling with the General Ocean Turbulence Model (GOTM) (Umlauf & Burchard 2003) for vertical mixing.
Generation of TUFLOW FV mesh files is a critical aspect of the modeling process. The SMS (Surface-water Modeling System) package supplied by Aquaveo (www.aquaveo.com/sms) was used to establish the mesh of Chagan Lake based on the delineation of the boundary of the domain and surveys of riverbed elevation (Figure 2(a)), using the paving mesh approach. The lake was divided into 1911 mesh elements of mixed triangles and quadrilaterals using 1707 nodes (Figure 2(b)). The generated mesh matched the convergence condition according to the mesh quality check. The spatial distribution of water and wetland plants (Phragmites spp.) was obtained according to the TM image (Figure 2(c)).
Water quality model
The hydrodynamic model was coupled with the open-source Aquatic Ecodynamic (AED2) modules (https://github.com/AquaticEcoDynamics) that could simulate the C, N, P, O cycles including inorganic nutrient, organic matter, phytoplankton components, and zooplankton components. This paper highlights the nutrient (NH3-N, TN, TP) predictions. Further details can be seen in the Aquatic Ecodynamics (AED2) Model Library Science Manual (Hipsey et al. 2013).
Input data and boundary conditions
The domain was forced by inflows, outflows and meteorological information. Meteorological data (solar radiation, wind, air temperature, humidity, rain and cloud cover) required for modeling at the daily scale were obtained from Songyuan meteorological station. Inflows and outflow were controlled by weirs. Inflows were set at Gaojia Bridge, Beidadu Weir (Figure 1). Shijiazi Weir was the outflow boundary, and was defined as WL-Q (water level-flow) boundary. The function of Chuantou Weir was to maintain the water level of Xinmiao Lake at 131.5 m to protect the area of Phragmites. Since the normal water level of Chagan Lake was in the range of 129.5 m and 131.0 m, an initial file was needed to define the different initial water level and concentration of water quality indicators for each cell.
Observed data used for the calibration and the validation process included water level (m), NH3-N (mg/L), TN (mg/L), TP (mg/L) from May 1 to Oct. 1, 2010 and 2011, at the inlet of Chagan Lake (downstream of Chuantou Weir).
RESULTS AND DISCUSSION
Mode calibration and validation
The values of parameters after a manual calibration process are shown in Table 1. The fitting results of simulated and observed values in calibration are shown in Figure 3. Statistical measures such as coefficient of determination (R2), root mean square error (RMSE) and mean absolute percentage error (MAPE) were applied to these results to quantify the accuracy of the model, which are as listed in Table 2.
The calibration and validation results of the coupled TUFLOW-FV and AED2 model were in accordance with the monitoring values, with a few exceptions. Water level got a satisfactory simulation result with MAPE less than 5% both in the calibration and validation periods. The MAPE values of NH3-N, TN, TP were around 10% in the calibration and validation periods. The simulated results could reproduce the temporal variation trend but they did not fluctuate as drastically as the observed values throughout the year. The possible reasons for the deviation were the nonpoint-source pollution and also the interactions between lake and groundwater that were not considered in this model. An additional consideration is that the Changan Lake is a Nature Reserve Wetland and fishery, and that the wetland plants (Phragmites) and fish may have an impact on the water quality could be the focus of further study. Above all, the simulation results were acceptable for realizing water environmental management targets, considering the limited data available for this region.
Carrying capacity estimation for NH3-N, TN and TP of Chagan Lake
Water environmental carrying capacity refers to the capacity of the system to allow the maximum pollutant load into a certain water body under given hydrological conditions and water quality goals.
The Chagan Lake was considered as a National Wetland Nature Reserve and tourism area, and the main water quality pollutant indexes were TN, NH3-N, TP, so the water quality target should be considered as Grade III (NH3-N <1 mg/L, TN <1 mg/L, TP <0.05 mg/L) referring to the Standard of Surface Water Environment Quality (GB3838-2002) (SEPA 2002). Then the carrying capacity of NH3-N, TN and TP were performed by the coupled TUFLOW-FV-AED2 model using the trial-and-error method (Dong et al. 2014) to obtain the maximum acceptable NH3-N, TN and TP loadings from May to September in the benchmark year 2011.
Table 3 shows the results of the maximum carrying capacity and reduction rate of pollutant loads into Chagan Lake. The total carrying load of NH3-N refers to the maximum input load of NH3-N that meets the protection target of both NH3-N and TN. The carrying capacity of TP was estimated after the changed from 0.4 to 0.01, because the reduction of input concentration had little effect on the variation of TP in the inlet of Chagan Lake (Figure 4). So if the TP of Chagan Lake could reach the Grade III water quality standard, not only the load from farmland drainage but also the desorption rate of P from the sediment needed to be decreased (Figure 4).
The total carrying load of NH3-N, TN and TP ( = 0.01) from farmland drainage into Chagan Lake from May to September could reach 1,147.6 t, 3,686.2 t and 100.8 t, as shown in Table 3. The total carrying loads of NH3-N, TN and TP ( = 0.01) all reached their maximum in July, which were 420.56 t, 1,322.01 t and 30.35 t, whereas May was characterized by the lowest carrying capacity. In May, the pollutants transported into the lake could be mixed to a higher extent due to the low water volume, and poor dynamic conditions.
The reduction of the amounts of pollutants to reach the carrying capacity were calculated and are listed in Table 3. Positive values indicate that the pollution load exceeds the environmental capacity and needs to be reduced, and negative values indicate that the environmental capacity remains in surplus and can accommodate a greater pollution load. Based on the data in Table 3, the reduction rates of TN and TP pollution load were almost more than 50%. The peak of reduction rate for NH3-N, TN and TP was in June (81.5%), September (77.9%), September (67.2%), respectively. The pollution load of NH3-N only showed a surplus in August among all months. The reduction of the amount of TP load did not include the decreased load when the parameter changed from 0.4 to 0.01.
Ecological water level calculation
Ecological water requirements are the required hydrological regime for maintaining valuable features and functions of aquatic ecosystems (Tharme 2003). The suitable ecological water level of lakes and wetlands was defined by the minimum water level necessary for a natural lake or wetland system to maintain and restore its normal ecological system integrity and functions (Liu et al. 2012). The methods to determine the ecological water level for lake wetlands included the historical lake level method, lake morphology analysis, lake surface area method, water balance analysis, water quality modeling, habitat analysis and species-environment models (Xu et al. 2004; Beca 2008; Cui et al. 2010; Shang 2013). This study used the historical lake level method, lake morphology analysis, and habitat analysis methods to define the suitable ecological water level of Chagan Lake.
For the historical lake level method, the temporal variation of maximum water level in Chagan Lake indicated that the annual mean maximum water level was 130.3 m, as shown in Figure 5. In 1998, the biggest flood made the water level reach a peak which was above 132.0 m. For lake morphology analysis, Figure 6 shows the relationship between water level, water surface area and water storage of Chagan Lake. Both the increase rate of water surface area and volume reached their peaks when the water level was 131.5 m. For habitat analysis, according to the ecological function of Chagan Lake, the macrophyte (Phragmites) and fish were the indicator species in Chagan Lake. The Phragmites were mainly distributed in Xindian Lake (Figures 1 and 2). The average elevation at the bottom of Xindian Lake was 130.6 m. According to studies, the minimum, suitable, and maximum water levels for Phragmites growth were 15 cm, 35–45 cm, and 1.01 m, respectively (Deng et al. 2012). So the elevation required for Phragmites growth was 130.75–131.61 m. The lake is used for fishery production that yields more than 6,000 tons of fish per year. Commercial species that migrated to the lake for feeding are caught during the winter fish-hunting festival every year. In addition, the region is a tourist area known for its salubrious climate and the famous winter fish-hunting festival. The average number of river-frozen days for Chagan Lake was 140–150 days from early November to early April, with an average of 1 m ice thickness. The required water level for activity space for fish under the ice in winter was about 0.5 m (Yan 2010). The fishery was located in the north-central area of Chagan Lake (Figure 1), and the average elevation was 129.2 m. So the elevation that was suitable for fish farming was 130.7 m. Based on the comprehensive analysis of the above results, the ecological water level elevation for Chagan Lake was 130.7–131.5 m.
Replenishment mechanisms for Chagan Lake on the control of eco-water level and water quality for different rainfall scenarios
According to the ‘Interconnected River–Lake System Network in Jilin Province’ plan, to improve the water quality of Chagan Lake, a channel was constructed to connect and utilize the Second Songhua River as another water supply resource of the Chagan Lake, especially in flood periods. So with the constraint of pollutant concentration and ecological water level, the suitable ratio of water diversion and farmland drainage amounts was estimated by the simulation of TUFLOW-FV and AED2 models to be under 10%, 25%, 50%, 75% and 90% of rainfall guaranteed rates.
The constraint condition of the water quality target was set as Grade III (TN <1 mg/L, TP <0.05 mg/L), and the water level was set as 130.7 m and 131.5 m. The NH3-N was not considered because if the TN can meet the water quality and quantity standard, then the NH3-N can also reach it. The supplemental diversion amounts and the reduction of farmland drainage amounts under five precipitation probabilities are listed in Table 4.
The amount of water diversion and farmland drainage required to dilute the TP of Chagan Lake to meet Grade III were simulated and calculated after changing the parameter to 0.01. The supplemental amounts of water diversion required to reach the Grade III standard and maintain the ecological water level at 130.7 m for TN, TP ( = 0.01) were separately [19.21, 46.19, 51.67, 59.13, 67.55], [114.13, 124.58, 131.81, 146.73, 168.45] million m3 under the rainfall guarantee rates of 10%, 20%, 50%, 75%, 90%, and [32.60, 49.84, 57.19, 63.70, 70.91], [117.25, 135.26, 168.17, 190.65, 218.32] million m3 for TN, TP ( = 0.01) to meet the Grade III standard and water level of 131.5 m. The reduction amounts of farmland drainage needed to reach the water quality of Grade III and water level of 130.7 m for TN, TP ( = 0.01) were separately [96.10, 65.05, 42.63, 13.00, 4.00], [87.16, 82.48, 76.93, 73.61, 69.61] million m3 under the rainfall guarantee rates of 10%, 20%, 50%, 75%, 90%, and [4.03, 0, 0, 0, 0], [73.08, 61.88, 50.23, 40.94, 31.98] million m3 for the water level of 131.5 m.
The biggest increase in diversion was in September for TN and TP ( = 0.01). The water diversion amounts for TP ( = 0.01) were far more than for TN, which indicated that the TP was the major pollutant and the phosphorus pollution in the lake is very serious.
The coupled TUFLOW-FV and AED2 models were demonstrated to be a suitable tool to simulate the process of hydrodynamic and water quality in Chagan Lake. With the constraint of the Grade III water quality standard, the carrying capacities of NH3-N, TN and TP ( = 0.01) load for Chagan Lake were 1,147.6, 3,686.2, and 100.8 t from May to October. In order to meet the Grade III standard and maintain the ecological water level from 130.7 to 131.5 m, the range of supplemental amounts of water diversion for TN, TP ( = 0.01) were separately [19.21–32.60, 46.19–49.84, 51.67–57.19, 59.13–63.70, 67.55–70.91], [114.13–117.25, 124.58–135.26, 131.81–168.17, 146.73–190.65, 168.45–218.32] million m3 under the rainfall guarantee rates of 10%, 20%, 50%, 75%, 90%. The reduction amounts of farmland drainage needed to reach the water quality of Grade III and maintain the ecological water level from 130.7 to 131.5 m for TN, TP ( = 0.01) were separately [96.1–4.03, 65.05–0, 42.63–0, 13.00–0, 4.00–0], [87.16–73.08, 82.48–61.88, 76.93–50.23, 73.61–40.94, 69.61–31.98] million m3 under the rainfall guarantee rates of 10%, 20%, 50%, 75%, 90%.
This work was supported by the National Natural Science Foundation of China (Grant No. 41301086, No. 41371108), Scientific Research Project of Public Welfare Industry of the Ministry of Water Resources, China (No. 201401014), and the Science and Technology Development Plan of Jilin Province (No. 20160520087JH). The authors acknowledge AED – Aquatic Ecodynamics Research Group for providing and teaching the TUFLOW-FV and AED2 models, and BMT-WBM for provision of the TUFLOW-FV model during Lei Zhang's visiting position at the University of Western Australia.