Seasonal regulation of river discharge by the cascade reservoirs in the Lancang River and its effect on downstream freshwater and estuarine saltwater intrusion

: This study assesses the seasonal regulation of river discharge by hydropower dam-induced cascade reservoirs in the Lancang River and its effect on downstream freshwater and estuarine saltwater intrusion. There are eight main reservoirs in the Lancang River, with a total regulation capacity of 25.67 billion m 3 , which regulates river discharge by conserving water in the flood season and releasing water in the dry season. River discharge during the dry season from 1960 to 2009 accounted for 21% of the annual discharge before the cascade reservoirs were constructed and increased to 33% from 2010 to 2015 after the cascade reservoirs were constructed at the Jinghong hydrological station, which is the lowermost station in the Lancang River. During the 2016 extreme drought in the lower Mekong River basin, the river discharge increased by 550, 367, 1283, 969, and 524 m 3 /s in January, February, March, April, and May, respectively, regulated by the cascade reservoirs at the Jinghong hydrological station. Considering runoff, tides, wind, and continental shelf currents, a high-resolution three-dimensional numerical model was used to simulate the effect of regulation of river discharge by the cascade reservoirs in the Lancang River on the saltwater intrusion in the Mekong River Delta (MRD). The simulation results show that the seasonal regulation of river discharge by the cascade reservoirs in the Lancang River weakens estuarine saltwater intrusion during the dry season, especially in the sand bar areas, which is much more significant in the extreme dry season of 2016. The seasonal regulation of river discharge by the reservoirs in the Lancang River makes the seasonal distribution of downstream river discharge more uniform, favoring down- stream freshwater utilization and alleviating flood disasters and saltwater intrusion in the MRD.


Introduction
Internationally, the Mekong River in China is called the Lancang River, or the upper Mekong River in China and Myanmar, and is called the lower Mekong River in Laos, Thailand, Cambodia, and Vietnam (X. Hou et al. 2021) (Fig. 1). The Mekong River originates from the Tibetan Plateau at elevations mostly greater than 5000 m and flows through six countries along a 4880 km course to the South China Sea, making it the longest river in Southeast Asia and the 12th longest river in the world (Gagliano and McIntire 1968;Guo 1994;Zhong and Wang 2011). The Mekong River, which drains a basin area of 0.795 × 10 6 km 2 (Walling 2008), is ranked 8th in terms of water discharge with an average flow of 15 000 m 3 /s (Mekong River Commission 2003). The river discharge of the Mekong River varies significantly with the monsoon season. From May to October (wet season in Vietnam), 85% (475 billion m 3 ) of the annual river discharge occurs and only 15% (78.8 billion m 3 ) from November to April (dry season in Vietnam) (UNEP 2006;Le et al. 2007). Cascade reservoirs have been built on the Lancang-Mekong River, which made the seasonal distribution of river discharge more uniform by flow manipulation (Mekong River Commission and Ministry of Water Resources of the People's Republic of China 2016). The Lancang River contains the Gushui, Tuoba, Huangdeng, Xiaowan, Manwan, Dachaoshan, Nuozgadu, and Jinghong reservoirs. These cascade reservoirs regulate river discharge by impounding water during the flood season (June to November in China) and implementing water supplements during the dry season (December to May in China), which favors flood control, alleviation of drought, and estuarine saltwater intrusion in the downstream area (Lei and Liu 2008; Mekong River Commission and Ministry of Water Resources of the People's Republic of China 2016; Hou et al. 2021;Yun et al. 2021).
The Mekong River Delta (MRD) is located at the downstream end of the Mekong River basin, which is flat and low-lying, with an area of only 11% of the entire Mekong River basin. The delta plain covers an area of 49 500 km 2 between Phnom Penh in the Cambodian lowlands and the southeast Vietnamese coast (Le et al. 2007). The MRD is an alluvial estuary that consists of many branches that transport a large amount of fresh water to the sea (Fig. 1). The Mekong River, when it enters Vietnam, splits into two branches, the Bassac (known as the Hau River in Vietnam) and the Mekong (known as the Tien River in Vietnam). These two branches form the Mekong Delta. The Hau River is the southern branch of the Mekong River. When the Hau River approaches the sea, it splits into two subbranches: Tran De and Dinh An. The Tien River is the northern branch of the Mekong River system, which separates into two subbranches at My Thuan: Co Chien and My Tho. At a distance of 30 km from the South China Sea, the Co Chien River again splits into two smaller branches, Co Chien and Cung Hau. In the downstream area, the My Tho River separates into four branches: Tieu, Dai, Ba Lai, and Ham Luong. Therefore, the MRD is a multi-channel estuary consisting of eight branches over which fresh water is distributed and discharged into the South China Sea.
As a typical delta, the MRD is affected by both runoff and tides. The river discharge into the MRD varies seasonally between 2100 m 3 /s in April (dry season) and 40 000 m 3 /s in September (wet season) (Nguyen and Savenije 2006). The tide is dampened due to a relatively strong river discharge, even during the dry season, and as a result, the MRD is a strongly riverine estuary. Inside the tidal region in Vietnam, there are four hydrological stations: Tan Chau and My Thuan in the Tien River, which are located 200 and 95 km from the Dai mouth, respectively, and Chau Doc and Can Tho, which are located 190 and 80 km from the Dinh An mouth, respectively (Fig. 1). Tides have mixed semidiurnal and mesotidal characteristics with a spring tidal range of approximately 3.5 m near the mouth and 2.0 m at Can Tho. Amplitudes of semidiurnal lunar tide M 2 and solar tide S 2 are up to 0.9 and 0.5 m, respectively, while diurnal amplitudes of solar tide K 1 and lunar tide O 1 are up to 0.7 and 0.5 m, respectively, along the MRD (Nguyen 2012). The East Asian monsoon causes strong seasonal climatic variations in the MRD (Xue et al. 2011). In the winter monsoon season from November to early March, winds mainly originate from the northeastern direction, and during the summer monsoon, southwestern winds prevail.
Vast amounts of water flow from the entire basin area through the MRD, causing severe floods in this area annually. Almost all Vietnamese people living in this region suffer from inundations caused by upstream river flows and storm surges. In the dry season, the delta is also affected by saltwater intrusion (Le et al. 2008;van der Scheer 2021). Saltwater intrusion is a common phenomenon in estuaries where fresh water and salt water converge and are mainly controlled by river discharge and tides (Pritchard 1956;Prandle 1985;Zhu et al. 2010;Qiu et al. 2012;Prandle and Lane 2015;Zhu et al. 2020). Saltwater intrusion in the MRD is complex, not only controlled by river discharge, but also by wind, sea level rise due to climate change, topography change induced by land subsidence, and riverbed incision by sand mining. Park et al. (2021) indicated that the 2020 saline water intrusion in the MRD has been recognized as the worst in recent decades and identified the four key drivers of saltwater intrusion as upstream hydropower dams, riverbed mining, land subsidence, and climate change-driven sea-level rise by analyzing the power spectrum of hourly water level series, and preliminarily concluded that riverbed incision might contribute more directly and rapidly to saltwater intrusion than others. Based on the observed extensive hydrological data and bathymetry data, Loc et al. (2021) assessed the recent intensification of saltwater intrusion and drought in the MRD, introduced a novel approach to decouple the compound effect of saltwater intrusion from four main environmental pressures: riverbed incision due to riverbed mining and dam construction, sea level rise, and land subsidence, and concluded that riverbed incision was identified as the main driver of saltwater intrusion. The South China Sea and the Gulf of Thailand are dominated by diurnal tides, and semidiurnal tides dominate the eastern Mekong deltaic coast. The wind monsoon climate can cause damped or amplified tides along the Mekong deltaic coast (Phan et al. 2019). Dang et al. (2019) indicated that under the impacts of hydropower dams, river discharge decreased during the flood season, but increased in the dry season, and the driest months shifted earlier. Furthermore, there will be fewer successive days without freshwater, and the number of hours with freshwater in the dry months will increase. Eslami et al. (2021) assessed the relative effects of various drivers and showed that anthropogenic forces, such as groundwater extraction-induced subsidence and riverbed level incisions due to sediment starvation can increase the salinity-affected areas by 10%-27% compared to the present-day situation, while future sea level rise adds another 6%-19% increase. From a long-term perspective using the sedimentary record, Tamura et al. (2020) proposed that the ongoing shrinkage of the Mekong Delta cannot simply be attributed to the impact of hydropower dam construction and sand extraction; in fact, a high level of mud supply sharply declined in the early 20th century after a vast canal network was built on the delta. Minderhoud et al. (2020) indicated that the MRD is rapidly losing elevation due to accelerating subsidence rates, primarily caused by increasing groundwater extraction, which increases the vulnerability of the delta to flooding, salinization, coastal erosion, and inundation. This will reduce subsidence by limiting groundwater exploitation, thereby limiting future elevation loss. Hackney et al. (2020) suggested that, in the Mekong River, excessive sand mining induces bank instability, potentially damaging housing, infrastructure, and threatening lives, it is imperative to limit sediment extraction rates to levels of a sustainable balance. Jordan et al. (2019) pointed out that a high magnitude of sand mining activities induced an imbalance between sediment supply and sand extraction can be substantial within the MRD, resulting in incision of river channels and bank erosion. The Mekong Delta is experiencing a significant decrease in the shoreline progradation rate, and accretion and erosion alternately occur along the delta coast. The shoreline experienced a shift from growing to shrinking around 2005, with a gradual increase in dams over the entire river basin (Y.Q. . The future of sedimentary processes on the MRD is ominous under conditions altered by human actions and climate change (Nittrouer et al. 2017). Within the distributary channels, McLachlan et al. (2017) predicted changes in the estuarine regime due to continued upstream dam construction and sea-level rise. Although there are many competing scenarios, the most likely is that more sediment will become trapped within the distributary channels than at present, limiting the source signal to the coastal ocean. Allison et al. (2017) suggested that both thalweg deepening and alteration of the high and low freshwater discharge from the catchment (i.e., dam control of water release) will likely impact estuarine circulation and sediment trapping, subsequently altering sand release to the shelf as well as changing the spatial extent and timing of mud mantling over sandy channel beds. Large river discharges can effectively restrain the estuarine saltwater intrusion. How the seasonal regulation of river discharge by the cascade reservoirs in the Lancang River affects downstream water resources and saltwater intrusion in the MRD remains unclear. Quantitative evaluation of these issues is crucial for the effective management of downstream freshwater resources, flood disasters, and saltwater intrusion in the MRD.
In this study, we investigated the effect of the seasonal regulation of river discharge by cascade reservoirs in the Lancang River on downstream river discharge and estuarine saltwater intrusion based on observed data and a three-dimensional (3D) numerical model. The configuration of the numerical model is described, which includes river discharge, tide, morphology, wind, and mixing. The changed river discharge at the Jinghong hydrological station before and after the construction of the cascade reservoirs in normal and extremely dry hydrological years was determined and adopted to simulate their effects on saltwater intrusion in the MRD.  (Table 1) and their locations are shown in Fig. 1. The three largest reservoirs are Nuozhadu, Xiaowan, and Gushui, with regulation capacities of 11.34, 9.9, and 2.46 billion m 3 , respectively, and each reservoir was completed in 2015, 2010, and 2015, respectively. The regulation capacity of the other five reservoirs is less than 1.0 km 3 . The total regulation capacity of the eight reservoirs is 25.67 billion m 3 . Except for the smaller reservoirs Dachaoshan and Manwan, which were built in 2003 and 2007, respectively, the others were completed after 2009. Therefore, we consider 2009 to be the study timeline. Before 2009, the river discharge of the Lancang River was under natural conditions, and after 2009, the river discharge was affected by the cascade reservoirs.

Materials and methods
The monthly mean river discharge before and after the construction of the cascade reservoirs was obtained from the report Mekong River Commission and Ministry of Water Resources of the People's Republic of China (2016), and their seasonal regulation of river discharge and the effect on downstream freshwater resources will be analyzed below.

Numerical model of estuarine saltwater intrusion
The ECOM-si model developed by Blumberg (1994) and later improved by Chen et al. (2001), Zhu (2003), and Wu and Zhu (2010) was used in this study to study hydrodynamics and substance transport. The independent variables were distributed on an Arakawa C-grid. The Mellor-Yamada 2.5 order turbulence closure module (Mellor and Yamada 1982) with stability parameters from Kantha and Clayson (1994) was included. The model used a sigma coordinate system in the vertical direction and a curvilinear non-orthogonal grid in the horizontal direction (Chen et al. 2004). A wet/dry scheme was included to describe the intertidal area with a critical depth of 0.2 m (Zheng et al. 2003). The HSIMT-TVD (high-order spatial interpolation at the middle temporal level coupled with a total variation diminishing scheme limiter) of the advection scheme was developed to significantly reduce the numerical diffusions with third-order accuracy .
The computational domain of the model covers the Mekong River Estuary and its adjacent sea (Fig. 2a), from 105.15°E to 107.65°E and from 8.35°N to 10.84°N. The western river boundary was set at the Tan Chau and Chau Doc hydrological stations. The model grid is composed of 330 × 455 curvilinear cells horizontally and five uniform σ levels vertically. In the horizontal direction, the grid was carefully adjusted to fit the coastline, was sufficiently refined in the area of interest with a minimal resolution of 40 m inward to the river mouths, and gradually increased to 2 km at the southwest open boundary. The integrated time step was set to 30 s. The water depth was derived from the National Oceanic and Atmospheric Administration (https://maps.ngdc.noaa.gov/viewers/wcs-client/) off the river mouths and digitized from the maps within the river mouths from the published literature. Interpolating the above data to the grids, the distribution of the model water depth shows that it is less than 10 m south of the river mouths and relatively deeper on the east side of the river mouths with a maximum value of ∼20 m. Bars exist at the river mouths where the water depth is less than 5 m (Fig. 2b).
The monthly mean river discharge measured at the Tan Chau and Chau Doc hydrological stations before 2009 was derived as the river boundary condition (https:// www.mrcmekong.org/). The river discharge at Tan Chau station is much larger than that at Chau Doc station, and the total river discharge reaches a minimum of 1256 m 3 /s in April and a maximum of 12 402 m 3 /s in September (Table 2).
Along the open sea boundary, the water elevation was determined by the tidal level and residual water level. The tidal level was calculated using the tidal harmonic constants of 16 astronomical tidal constituents (M 2 , S 2 , N 2 , K 2 , K 1 , O 1 , P 1 , Q 1 , MU 2 , NU 2 , T 2 , L 2 , 2N 2 , J 1 , M 1 , and OO 1 ), derived from the NaoTide dataset (https://www.miz.nao.ac.jp/staffs/nao99/). The residual water level reflects the continental shelf circulation and was derived from Simple  Ocean Data Assimilation (SODA), which is the global monthly mean data with a spatial resolution of 0.5°× 0.5°(http://iridl.ldeo.columbia.edu/SOURCES/.CARTON-GIESE/.SODA/). The sea temperature was set to a constant temperature of 30°C. The salinity at the open sea boundaries and initial salinity were also obtained from the global monthly average data of SODA. The sea surface boundary condition of the momentum equations is given in the form of wind stress, which has a quadratic relationship with the wind speed. The monthly mean wind was derived from the European Center for Medium-Range Weather Forecasts (ECMWF) database with a spatial resolution of 0.5°× 0.5°and a temporal resolution of 6 h (http://apps.ecmwf.int/datasets/data/interim-full-daily/levtype=sfc/). The wind is a northeasterly wind with a speed of approximately 8 m/s in February and an easterly wind with a speed of approximately 4.5 m/s in April (Fig. 3).
Saltwater intrusion in the MRD is complex and dam-induced change of river discharge is not the only factor, but there are also other important environmental pressures that influence saltwater intrusion. The purpose of this study is to focus on and analyze the effect of seasonal regulation of river discharge by the cascade reservoirs in the Lancang River on estuarine saltwater intrusion with a 3D numerical model, in which the dynamic factors of morphology, river discharge, tide, wind, and mixing are considered; other environmental pressures such as climate change, land subsidence, riverbed incision by sand mining, coastal erosion, and sea level rise are not considered, given the purpose of the study, lack of available data, and limitation of paper length.
Because the tide is just the oscillation of water under the influence of the attractive gravitational forces of the Moon and Sun (Simm et al. 1996), the differences in astronomical  tides during the dry season in different years are small. To ensure that the tidal conditions were identical in the numerical simulations before and after the construction of the cascade reservoirs, we selected the same simulation time. All model simulations cover the period from 1 January to 31 May 2015, which is the dry season. Because the model needs 1-2 months to adjust the hydrodynamics and salinity to stabilize, the model results from March to May are output and analyzed. Two numerical experiments were performed to quantitatively investigate the effects of the seasonal regulation of river discharge by cascade reservoirs in the Lancang River in a normal hydrological year on saltwater intrusion in the MRD. In numerical experiment 1 (Exp. 1), the river discharge flowing into the MRD from 1 January to 31 May 2009 without the cascade reservoirs in the Lancang River was specified as the open river boundaries at the Tan Chau and Chao Doc hydrological stations (Table 2). In experiment 2 (Exp. 2), seasonal regulation of river discharge by the cascade reservoirs in the Lancang River was considered, and the river discharge from January to May in Exp. 1 was increased by 288, 236, 576, 550, and 341 m 3 /s (Table 3), respectively. The distance between the Jinghong hydrological station and the Tan Chau and Chau Doc hydrological stations was approximately 2600 km. Because river discharge can be altered by other factors such as changes in rainfall patterns, water diversion, and drainage along the river, the river discharge by the cascade reservoirs in the Lancang River is no longer the one at the Tan Chau and Chau Doc hydrological stations. It is difficult to calculate the change value caused by rainfall patterns, water diversion, and drainage along the river due to the lack of available data. In this study, we hypothesized that the change in river discharge by the cascade reservoirs in the Lancang River is directly superimposed on the monthly mean river discharge without the cascade reservoirs at the Tan Chau and Chau Doc hydrological stations. This method was also adopted to study the influence of seasonal runoff regulation by the Three Gorges Reservoir on saltwater intrusion in the Changjiang River Estuary (Qiu et al. 2013).

Effects of the seasonal regulation of river discharge by cascade reservoirs in the Lancang River on downstream river discharge
The water discharge of the cascade reservoirs occurs from December to May of the following year (dry season), and water storage occurs from June to November (flood season). The monthly mean river discharge at the Jinghong hydrological station is shown in Table 3, which indicates that the minimum river discharge of 550 m 3 /s occurred in March, and the maximum river discharge of 3840 m 3 /s occurred in August from 1960 to 2009 (before the cascade reservoirs), while the minimum river discharge of 812 m 3 /s occurred in February, and the maximum river discharge of 2885 m 3 /s still occurred in August from 2010 to 2015 (after the cascade reservoirs). Because of the regulation of the cascade reservoirs, the river discharge from December to May of the following year during water discharge increased, Table 3. Monthly mean river discharge at Jinghong hydrological station before and after construction of the cascade reservoirs in the Lancang River and their difference (m 3 /s), which is also used for the upstream boundary condition of the model. and from June to November during water storage, it decreased (Table 3). The discharge in the dry season during 1960-2009 accounted for 21% of the annual discharge before the reservoirs and increased to 33% during 2010-2015 after cascade reservoirs were constructed at the Jinghong hydrological station. The seasonal regulation of river discharge by the reservoirs in the Lancang River greatly increased the downstream river discharge in the dry season and significantly decreased the downstream river discharge during the flood season, which helped to increase freshwater resources and reduce flood disasters in the lower reaches of the downstream Mekong River basin (Si and Liu 2008;Hou et al. 2021;Yun et al. 2021).

Effects of the seasonal regulation of river discharge by cascade reservoirs in the Lancang River on saltwater intrusion in the MRD
In Exp 1, driven by tide, runoff, wind, and continental shelf currents, the vertically averaged total current in late April is shown in Fig. 4. During spring, the current at the maximum flood flows upstream inside the river mouths. Away from the river month, the current flows toward the coast, while it flows southward along the coast between the Dinh and Cung Hau River mouths and between the Co Chien and Ham Luong River mouths, and the current flows northwestward off the Dai and Tieu River mouths and northeastward along the coast west of 106°E. The current at the maximum ebb was the opposite. During the neap tide, the current pattern is similar to that during the spring tide, and only the current speed decreases. Previous studies indicated that the tidal regime off the MRD is affected by two sources: regular semidiurnal tides dominated by the main semidiurnal tidal constituent M 2 from the South China Sea and irregular diurnal tides dominated by the main diurnal tidal constituent K 1 from the Gulf of Thailand (Fang et al. 1999;Zu et al. 2008;Le et al. 2008). The effect of the former is stronger than that of the latter, as simulated by our numerical model.
After the tidal current was filtered out, the residual current flowed seaward in the branches and then turned right and flowed along the coast under the influence of the Coriolis force (Fig. 5). Away from the river mouths, the residual current flows westward and is weaker in the surface layer and flows eastward in the bottom layer during the spring tide; during neap tide, the residual current flows northwestward and then turns anticlockwise and flows southwestward with a relatively higher speed in the surface layer and flows landward with a lower speed forced by the easterly wind in late April. The residual current pattern during neap tide is consistent with the theory of classic estuarine two-layer circulation (Fang et al. 1999;Zu et al. 2008;Le et al. 2008). Fang et al. (1998) indicated that the coastal current off the MRD has a southwestern direction forced by the northwest East Asian monsoon during winter. The simulated coastal current was weaker because the wind was easterly in late April rather than northeasterly.
The distribution of tidally averaged salinity shows that the strongest saltwater intrusion occurs in the Dai and Tieu branches, which is the northernmost branch of the Mekong; the second strongest saltwater intrusion occurs in the Tran De and Dinh An branches, which are the southernmost branches of the Mekong; and the weakest saltwater intrusion occurs in the Ham Luong branch (Fig. 6). The surface salinity is almost the same as the bottom salinity, meaning that vertical mixing is strong in the shallow water induced by wind stirring at the water surface and tidal current friction at the bottom. Saltwater intrusion in the spring tide is stronger than that in the neap tide. Salinity fronts exist at river mouths. The river plume extends southwestward along the coast, which is consistent with the flow of the coastal residual current.
Eight model output sites were selected in the branches of the MRD to show the surface salinity variation from 15 March to 15 May (Fig. 7, left panel). At site A1, which is located in the upper reaches of Tran De, the salinity is far less than 0.45 (salinity standard for drinking water), meaning that there is no saltwater intrusion there. At site A2, which is located in the upper reaches of Dinh An, the salinity is less than 0.45 most of the time, the maximum reaches 0.45, and saltwater intrusion is weak overall but stronger than that at site A1. At site A3, which is located in the upper reaches of Cung Hau, the salinity is greater than 0.45 in tidal flood current during spring tide, and the maximum reaches 1.6, indicating that saltwater intrusion is stronger there. At site A4, which is located in the upper reaches of Co Chien, the salinity is higher than that at site A3, meaning that the saltwater intrusion in the north branch is stronger than that in the south branch because the Coriolis force affects the tidal flood current, which brings saline water into the river. At site A5, which Fig. 7. Temporal variation in surface salinity from 15 March to 15 May at model output sites A1, A2, A3, A4, A5, A6, A7, and A8 (position is labeled in Fig. 1) in a normal hydrological year (left panel) and in an extremely dry hydrological year 2016 (right panel). Black line: without river discharge regulation by the reservoirs; red line: with river discharge regulation by the reservoirs; and green dashed line: salinity 0.45, which is the salinity standard of drinking water.
is located in the middle reaches of Ham Luong, the salinity is low and far less than 0.45, and saltwater intrusion is weak there. At site A6, which is located in the upper reaches of Ba Lai, the saltwater intrusion is weak before March and becomes stronger after mid-April to mid-May. At site A7, which is located in the upper-middle reaches of Dai, the salinity is far greater than 0.45, the maximum reaches 11, and the saltwater intrusion is very strong there. At site A8, which is located in the upper-middle reaches of Tieu, the salinity is between 1.6 and 3.1, and the saltwater intrusion is stronger overall but weaker than that at site 7.
The cascade reservoirs in the Lancang River discharge more water during the dry season, which is saved during the flood season, increasing the river discharge and weakening the saltwater intrusion (Fig. 8). An obvious decrease in salinity occurs in the sand bar area with a value greater than 1.0, where there is a salinity front that is sensitive to changes in river discharge. In the Dai and Tien branches, the salinity decreased to 3.0. Figure 7 once again clearly shows that the salinity decreases significantly at the eight sites, meaning that the cascade reservoirs in the Lancang River discharge more water and weaken estuarine saltwater intrusion in the dry season.

Discussion
Cascade reservoirs regulate river discharge every year after construction. For drought years, more river discharge can be released to alleviate drought in the lower Mekong River. For example, there was an extreme drought downstream of the Mekong River in 2016. The river discharge at the Tan Chau and Chau Doc hydrological stations during the drought of 2016 was approximately two-thirds that of the normal hydrological year. Emergency water supplementation from the Jinghong Reservoir was implemented, increasing the water level and river discharge along the Mekong mainstream.
In this section, we discuss the effect of cascade reservoirs in the Lancang River on downstream freshwater and estuarine saltwater intrusion under the extreme drought of 2016. Comparing the monthly mean river discharge at the Jinghong hydrological station from 1960 to 2009, the Jinghong Reservoir had increased river discharge by 550, 367, 1283, 969, and 524 m 3 /s in January, February, March, April, and May, respectively, which was much larger than that in normal hydrological years, which provided more freshwater and greatly alleviated the drought downstream (Mekong River Commission and Ministry of Water Resources of the People's Republic of China 2016). Figure 9 shows the surface and bottom salinity distributions during spring tide and neap tide in late April 2016, indicating that saltwater intrusion became much stronger than that in normal hydrological years due to low river discharge (Fig. 6). Extremely low river discharge can cause severe saltwater intrusion in the MRD and seriously affect local rice production and residents' lives (Le et al. 2008;van der Scheer 2021). The distribution of tidally averaged salinity shows that the regulation of river discharge by the cascade reservoirs in the Lancang River significantly weakened saltwater intrusion in the MRD (Fig. 10). Compared with the salinity change in a normal hydrological year (Fig. 8), the degree and range of salinity decrease became significantly larger and reached more than 3.0 in the sand bar area. The variation in salinity (Fig. 7, right panel) similarly indicates that the saltwater intrusion became much stronger even in the Tran De and Dinh An branches, and the salinity was greatly reduced by the cascade reservoirs in each branch of the Lancang River. Some previous studies have argued that the reservoirs in the Lancang River enhance saltwater intrusion in the MRD (Duong et al. 2018) because they thought the reservoirs in the Lancang River reduce the river discharge in the dry season, while the opposite is true.
Our results show that the seasonal regulation of river discharge by the reservoirs in the Lancang River reduces downstream river discharge during flood seasons, increases river discharge in dry seasons, favors downstream freshwater utilization, and alleviates flood disasters in flood seasons and saltwater intrusion in dry seasons in the MRD, especially during drought years. Based on the research results of this study, we propose to establish a Mekong River international commission of six South Asian countries, China, Myanmar, Laos, Thailand, Cambodia, and Vietnam to make full use of water resources and mitigate flood disaster and estuarine saltwater intrusion by human activities, that is, the seasonal regulation of river discharge by the cascade reservoirs in the Lancang River and the lower Mekong River in the future.

Conclusions
Considering runoff, tides, wind, and continental shelf currents, a high-resolution 3D numerical model was used to simulate the hydrodynamic processes and saltwater intrusion in the MRD. The residual current flows seaward in the branches and then turns right and southwestward flows along the coast under the influence of the Coriolis force. The strongest saltwater intrusion occurs in the Dai and Tieu branches, which is the northernmost branch of the Mekong; the second strongest saltwater intrusion occurs in the Tran De and Dinh An branches, which are the southernmost branches of the Mekong; and the weakest saltwater intrusion occurs in the Ham Luong branch. The cascade reservoirs in the Lancang River discharge more water and weaken the estuarine saltwater intrusion during the dry season. An obvious decrease in salinity occurs at the river mouth where there is a salinity front, with a value greater than 1.0. In the Dai and Tien branches, the salinity decreased to 3.0. Saltwater intrusion became much stronger under low river discharge in 2016. The regulation of river discharge by the cascade reservoirs in the Lancang River in the dry season of 2016 weakened the saltwater intrusion in the MRD more distinctly. Compared with the salinity change in normal hydrological years, the degree and range of salinity decrease became significantly larger and reached more than 3.0 in the sand bar area.
The seasonal regulation of river discharge by the reservoirs in the Lancang River favors downstream freshwater utilization and alleviates flood disasters and saltwater intrusion in the MRD, especially during drought years. Accordingly, we propose to establish a Mekong River international commission of the six South Asian countries to make full use of water resources and mitigate flood disaster and estuarine saltwater intrusion with the seasonal regulation of river discharge by the cascade reservoirs in the Lancang River and lower Mekong River in the future.