Application of a 2D hydrodynamic model for the assessment of the influencing factor in Xiangxi Bay of the Three Gorges Reservoir

: A two-dimensional (2D), laterally averaged, hydrodynamic and water quality model, CE-QUAL-W2, was used to simulate the hydrodynamics and water temperatures of Xiangxi Bay (XXB) of the Three Gorges Reservoir (TGR). Algal blooms occurred seasonally in XXB which is influenced by density currents. The CE-QUAL-W2 model was calibrated using data collected in XXB from January to December 2010. The model was then used to examine the environmental factors that affect the density currents. The model predicted density currents such as overflows, interflows and underflows of the XXB. The results suggest that more than 25%, 58.33% and 16.67% of the density currents that occurred in the XXB were overflow, interflow and underflow, respectively. Field observations were conducted including travel distance and thickness of intrusion layer for 24 days in 2010. The average travel distance of overflow, interflow and underflow in XXB was 15.50 km, 8.78 km, and 9 km based on observations and 16 km, 9.40 km, 9.60 km based on the model results. Therefore, numerical simulations of density currents can be predicted to better understand the hydrodynamics and aquatic ecosystems in XXB of TGR.


Introduction
The Three Gorges Reservoir (TGR) is located at the end of the upper Yangtze River (YR) in central China. It has many tributaries, including Xiangxi Bay (XXB), which is the largest tributary in the lower reach of the TGR and 32 km from Three Gorges Dam ( Figure. 1a). The Three Gorges Project (TGP) has created substantial social and economic benefits, such as power generation, flood control, river navigation, drought prevention and freshwater impoundment in central China [1,2,3,4]. However, the aquatic environment of the area has significantly changed since impoundment (2003) [1,2,3,5]. Algal blooms have occurred frequently in many tributaries, including XXB [1,2,3]. Density currents influence algal blooms because these currents transport nutrients and trace elements [1,3,7,13]. Additionally, density currents are influenced by the thermal stratification of the reservoir [2,8,9,10,15]. Analyses of long-term data records suggest that algal blooms in the tributaries of TGR are sensitive to changes in thermal stratification, hydrodynamics and nutrient loads [1,4,7]. Based on long-term field observations, Ji et al. (2010) noted that density currents occur frequently in XXB. Je et al. (2010) also noted that the water in the TGR mainstream flowed backward into XXB at different plunging depths due to the water temperature differences between the TGR mainstream and XXB. Density currents are also affected by upstream inflows into XXB and intrusion flows from the TGR mainstream and these currents affect the algal blooms and aquatic ecosystems in XXB [1,2,3,4]. Density currents exist below the ambient water level and form different flows, such as overflows, interflows and underflows [8,11,12]. Numerous studies have investigated density currents around the world, including research on density currents [6,10,11,12]. Numerical models have become widely accepted tools for predicting and diagnosing hydrodynamic problems [17,18,19,20,21].
The objective of this paper is to investigate density currents in XXB of the TGR using a two-dimensional, laterally averaged, hydrodynamic model based on CE-QUAL-W2. This paper focus on the characteristics of density currents, such as overflows, interflows, underflows, the thickness of the intrusion layer, plunge point locations and the travel distance of the density currents in XXB.

Study area
XXB is located (from 110°25′E to 111°06″E and from 30°57′N to 31°34′N) in Hubei Province, China. The bay has a watershed area of 3095 km 2 and an annual average flow of 47.4 m 3 /s [1,2,3,4]. Eleven field stations were considered for the model developed from upstream to downstream. The sampling points were XX09, XX06 and XX01, corresponding to computational grid sections 9, 25 and 58, respectively ( Figure. 1b). A computational grid of XXB was developed based on the bathymetric and geometric data.
The XXB was represented by 64 longitudinal segments, each 500 m in length, and 109 vertical layers, each 1 m thick [1,4]. CE-QUAL-W2 has several coefficients that can be adjusted in the calibration process [17,18,19,20,21]. Model performance was evaluated by the absolute mean error (AME) and the root mean square error (RMSE) statistics.  (Table 1). In the early spring, the water density of XXB gradually decreased and thermal stratification occurred. Overflow density currents enter from the mainstream to XXB in a clockwise direction, as shown in Figure. 3 (a) and Figure 3 (b). In mid-spring, the temperature gradually increases and thermal stratification pattern should reverse to the anti-clockwise direction as shown in Figure. 3 (c). In the late spring, the temperature should be uniform but upstream inflow will be colder than surface waters, as shown in Figure. (Table 1). In summer, the water temperature increased significantly and the density was very low in XXB. Interflow patterns reversed to an anti-clockwise direction on June 9, July 30 and August 25 but not on August 1 as shown in Figure. (Table 1). In XXB, underflow density currents occurred in the late summer and autumn when tributaries were still warm as shows in Figure.

Density current effects on thermal stratification in XXB
Thermal stratification is influenced by the density currents in XXB. The upstream reaches exhibited much stronger stratification than those downstream [1,2,3,4]. In XXB, the CE-QUAL-W2 model effectively captures the seasonal changes in water temperature and stratification in the reservoir [1,4,7,18]. The difference between observed and simulated temperature is generally below 4°C in XXB. The thermocline depth is accurately simulated during the simulation period. The model performance was evaluated according to the absolute mean error (AME) and the root mean square error (RMSE) values. The upstream AME was 1.

Density current control technique for XXB
Though Xinagxi Bay of the TGR is located in a subtropical continental monsoon climate area algal blooms occurred seasonally [1,2,3]. Upstream inflow is separated into two types: over the surface as overflow and on the bottom as underflow. After installation the curtain weir inflow propagation path will be interrupted as shown in Figure. 1(c). Inflow exhibited plunge flow before the curtain weir location and later travel as interflow under curtain with different heights such as 3 m, 5 m and 7 m below the water surface [4,7,14,15,18]. The dilution rate of the surface layer increases with the curtain weir height and enhanced the submergence interflow [4,7,13,14,16,17,19,20,21]. The submergence level of interflow should be interrupted because the curtain weirs have different heights and locations including upstream, midstream and downstream, corresponding to XX09, XX06 and XX01, respectively Figure. 1(c). We assumed that after installation the curtain weir density current would be interrupted and would, in turn, alter the thermal stratification in XXB (Figure. 1c). Overflow, interflow and underflow should be interrupted and control algal blooms due to the effect of curtain weirs applied at different reservoirs [4,7,15,16,18]. Our further research will be on modeling curtain weirs and its effects on controlling algal blooms in a subtropical reservoir of China.

Conclusions
Density currents provide important information regarding transport dynamics and stratified flows. These currents are the main environmental factors that affect XXB. The CE-QUAL-W2 model performed well in simulating the unusual thermal stratification, overflows, interflows, underflows, travel distances, and thicknesses of the intrusion layers in XXB.
The model predicted that, more than 25%, 58.33% and 16.67% of density currents in XXB were overflows, interflows and underflows, respectively. The average travel distance of overflow, interflow and underflow in XXB was 15.50 km, 8.78 km, and 9 km based on observations and 16 km, 9.40 km, 9.60 km based on the model results. The model theoretically confirmed the long-term existence of density currents and indicated that density currents have a major hydrodynamic influence in XXB. Density currents carry nutrients, solids and other substances that affect reservoir water quality. These materials may enter XXB through overflow, interflow and underflow from the mainstream. Curtain weir numerical simulations should be performed to control density currents in XXB.
Additionally, due to the application of curtain weirs algal blooms should be controlled in XXB of the TGR.

Funding
This research received no external funding.