Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary

Estuarine, Coastal and Shelf Science xxx (2014) 1e9

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss

Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary Maotian Li a, Zhongyuan Chen a, *, Brian Finlayson b, Taoyuan Wei a, Jing Chen a, Xiaodan Wu c, Hao Xu c, Michael Webber b, Jon Barnett b, Mark Wang b a b c

State Key Laboratory of Estuarine and Coastal Research, East China Normal University, Shanghai 200062, China Department of Resource Management and Geography, The University of Melbourne, VIC 3010, Australia Department of Geography, East China Normal University, Shanghai 200062, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 April 2014 Accepted 14 July 2014 Available online xxx

The densely-populated mega-city of Shanghai relies increasingly on freshwater from the Changjiang estuary (70% now). However, this strategy is facing potential threats due to extensive water diversion in the lower Changjiang basin and future sea-level rise. Given this, the present study evaluates the ability of Shanghai to source its water from the estuary, especially in the dry season. Flow <15,000 m3 s
1, which occurs for ca. 50% of dry seasons, represents the threshold for salinity 0.45 psu (chloride 250 mg/L) above which the estuary is unusable for freshwater. Correlating discharge and salinity, maximum salinity and related time duration, and taking the future water diversions and sea-level rise into consideration, we extrapolated salinity events into the future at intervals of 10 years until 2040. We estimate that water diversions of 56.2  109 m3 (1800 m3 s
1), 59.2  109 m3 (1900 m3 s
1) and 61.3  109 m3 (2000 m3 s
1) will occur in 2020, 2030 and 2040, and a rise of sea level of 0.12 m by 2040 (from 2010), equivalent 506 m3 s
1, ca. 19.4% of the total reducing discharge of 2040 into the estuary (ca. 28% projected to the worst case of February of 2040). Based on scenario building, the pattern of salinity distribution would remain >0.45 for 20e65, 75e90 and 120e128 days (in 2020, 2030, and 2040, respectively), for extreme low-flow conditions. These periods exceed the present 68-day maximum freshwater storage in Qingcaosha reservoir, which is meant to secure freshwater for Shanghai in the future. Urgently countermeasures are needed to secure the Shanghai's water in the future. © 2014 Elsevier Ltd. All rights reserved.

Keywords: salinity discharge water diversion sea-level rise the Changjiang estuary water supply

1. Introduction Salinity is a basic measure of water quality in estuarine environments and in particular it determines whether or not estuary water can be used for potable water supply (Huang and Foo, 2002; Hong and Shen, 2012). Large variations in salinity occur in estuaries where riverine freshwater meets and interacts with saltwater to form strata with fresh overlying salt and areas with a range of different salinities produced by mixing (Schroeder et al., 1990; Shen et al., 2003; Gong et al., 2013). Freshwater discharge from upstream, tidal range, physical configuration of the estuary, and in the longer

* Corresponding author. E-mail addresses: [email protected] (M. Li), [email protected], Z.Chen@ sklec.ecnu.edu.cn (Z. Chen), [email protected] (B. Finlayson), weit235@yahoo. cn (T. Wei), [email protected] (J. Chen), [email protected] (X. Wu), [email protected] (H. Xu), [email protected] (M. Webber), jbarn@ unimelb.edu.au (J. Barnett), [email protected] (M. Wang).

term, sea-level rise, interact to determine where, and for what periods of time in the tidal cycle, salinity will limit the usability of the estuarine water (Chen et al., 1991; Caitlin et al., 2009; Wolanski and McLusky, 2011). For rivers with a large seasonal variation in freshwater discharge, as is the case for the Changjiang, the season with lowest flow will be vulnerable to higher salinities, and this may be compounded if management activities serve to reduce the freshwater discharge still further. The Changjiang River mouth is a huge cone-shaped estuary with three major islands located there, trending en-echelon towards the southeast (Fig. 1). The river mouth is divided into the South and North Branches, separated by Chongming Island (CM), and the South Branch is divided into the South and North Channels, separated by Changxing (CX) and Hengsha (HS) Islands. The South Channel is further divided into the South and North passages separated by Jiuduansha (JDS) Island (Fig. 1). In the very long term the river mouth has been shifting southeastward via intensive sediment dynamical processes (Chen et al., 1988; Li et al., 2011), so

http://dx.doi.org/10.1016/j.ecss.2014.07.007 0272-7714/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

2

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

Fig. 1. The lower Changjiang River, showing sites of water diversions; location of freshwater reservoirs, and the sites where hydrological data were collected. Also indicated is the pattern of saltwater intrusion into the estuary via the South and North Branches.

at present 95% of the Changjiang discharge is via the South Branch, and the North Branch is being gradually silted up (Chen et al., 1988; Dai et al., 2013). The hydrodynamics and morphology of the Changjiang estuary are such that there are unusual circumstances relating to the intrusion of saltwater (Mao et al., 2000; Xue et al., 2009). Saltwater intrudes into the estuary in the winter low flow period through both the South and North Branches (Fig. 1). The saltwater intrusion in the South Branch is opposed by the main freshwater discharge of the river, and therefore weakens generally around the middle of Changxing Island (Fig. 1). Saltwater can also intrude into the North Branch, passing right around the northwest end of Chongming Island and into the South Branch, meeting the freshwater flow upstream of the saltwater intrusion that comes directly into the South Branch (Zhang et al., 2012). Recent changes to the supply of freshwater to Shanghai have involved building infrastructure in the Changjiang estuary to replace most of the water previously taken from the Huangpu River basin due to severe water pollution (Finlayson et al., 2013) (Fig. 1). The Qingcaosha reservoir has been built adjoining the northeastern shore of Changxing Island (Fig. 1). It has a water surface area of 70 km2 and a storage capacity of 4.38  108 m3. Although this reservoir site is known to be affected by saltwater intrusion in the winter season, the storage holds 68 days supply of water for Shanghai. This will allow the reservoir to continue supplying freshwater from the storage for the duration of a saltwater intrusion event with a return period of 100 years (Le, 2012). Two smallscale intake points with reservoirs, Dongfengxisha and Chenhang, are also located in the Changjiang River mouth (Fig. 1). These together can supply 70% of the freshwater demand of Shanghai city as at 2012. The conditions for which this system was designed are changing and will continue to change in the future. This will be a great concern as new projects are built to divert water from the upper Changjiang into northern China where water shortage has occurred due to over irrigation and urbanization etc. (Zhang et al., 2003; Wang et al., 2008), and particularly, as the sea level is rising and the delta is sinking (Chen et al., 2001). These diversions have reduced the discharge in the winter season, causing more severe saltwater intrusions. The situation is exacerbated during the dry season such as occurred in 2006 (Chen and He, 2009; Zhu et al.,

2010). These factors tend to suggest that the storage capacity of 68-days may not be enough to protect the water supply to Shanghai in the future. In this paper we examine the historical record of discharge in relation to the water diversion projects and sea-level rise to determine the severity of this threat to future water supply to Shanghai, taking account of the saltwater incursions directly into the South Branch and via the North Branch around Chongming Island. We present a scenario of the sustainability of the freshwater sources for Shanghai given its present population of >23 million, and the lack of alternative water sources given the heavy pollution of local water bodies in the Changjiang delta region (Finlayson et al., 2013). 2. Data sources and methodology 2.1. Discharge Daily discharge recorded at Datong station (Fig. 1) was sourced from the hydrology yearbooks of the Changjiang for the periods 1950e1986 and 1997e2011 (The Changjiang Water Conservancy Committee, 1950e1986, 1997e2011). Datong is the furthest downstream hydrological gauging station, but is 600 km upstream of the river mouth. Daily discharges are shown in Fig. 2A. We define the wet season as May 1eOct. 30 and the dry season as Nov. 1eApril 30. We have grouped daily discharge into 4 types, i.e. high flow, normal flow, low flow and extreme low flows. High flow means the maximum discharge recorded on that particular day of the year in the 51 year record; normal flow is the average discharge on the same day of the year over 51 yrs; low flow is 10% more than the extreme low flow; and extreme low flow is the minimum discharge on that date over the 51 yrs (Fig. 2B,C). In this paper we focus attention on the dry season, when strong saltwater intrusions occur. The frequency distribution of daily discharges was used to determine the probability of particular threshold flows that will be cited in the following discussion (Fig. 2D, E). 2.2. Salinity Daily mean salinity readings for 769 days when saltwater intrusions occurred between 1979 and 2011 were collected from the

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

3

Fig. 2. A) Daily discharge at Datong station (1951e2011); B) the range of daily discharge in the wet season; C) the range of daily discharge in the dry season; D) Frequency distribution of daily discharge in the dry season; and E) the cumulative frequency of daily discharge in the dry season.

three key hydrological gauging stations, i.e. Gaoqiao (representing Qingcaosha reservoir, 316 days), Chenhang Reservoir (277 days), and Chongtou (representing Dongfengxisha reservoir, 176 days), all located in the Changjiang River estuary (Fig. 1; data sourced from Mao et al., 1993, 1994, 2001; Wu., 2006; Xu and Yuan, 1994; Zhu et al., 2010; Tang et al., 2011; Li, 2012). Salinity data of the present study recorded as chloride concentration (mg/L) were converted to practical salinity (Shen et al., 2003). The measured daily mean salinity is plotted against the discharge measured at Datong in Fig. 3A1e3 with a 7-day lag to account for the flow time between Datong and the river mouth. For each case, the salinity values were split into 0.27 ranges and the mean value for each range was plotted against discharge at Datong (with a 7-day lag) in Fig. 3B1e3, together with the maximum and minimum value for each range. Least squares regressions have been fitted to each of the minimum, mean and maximum series for each

station (Fig. 3B1e3) and these regressions are all significant (p < 0.001). Fig. 3C1e3 shows the duration of salinity events where the maximum salinity recorded was greater than 0.45. As would be expected in this situation, higher salinities are reached as the duration of the event increases. 2.3. Water diversion The Changjiang Water Conservancy Committee (2007) reports that there were 739 water diversion projects in operation in 2000 along the lower Changjiang spread across three provinces downstream of the Datong gauging station. The total design capacity of water diversion is 788  109 m3 per year. However, the total actual annual water diversion was 38.6  109 m3 in 2000 as reported by The Changjiang Water Conservancy Committee (2007). We have located and verified the capacity of 72 of these diversion sites,

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

4

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

Fig. 3. A1eA3) Daily mean salinity at recording stations in the estuary in relation to daily discharge recorded at Datong with a 7-day lag for travel time to the estuary; B1eB3) correlation of daily mean salinity to discharge (7-day lag from Datong), showing sectional minima, means and maxima (explanation in text); C1eC3) duration in relation to maximal salinity level during saltwater intrusion events in the estuary.

shown in Fig. 1. The monthly distribution of water diversions for the year 2000 are shown in Fig. 4A. The annual water diversions in the lower Changjiang below Datong station for the period 2005e2010 were obtained from the Water Resource Bulletins of Anhui and Jiangsu provinces, and Shanghai municipality (The Water Resources Department of Anhui Province, 2005e2011; The Water Resources Department of Jiangsu Province, 2005e2011; The Water Resources Department of Shanghai, 2005e2011). We have assumed that water diversions will increase in the future following the pattern of increases in the past. As shown in Fig. 4B we have constructed a regression of annual water diversions against years (with the year 2000 being assigned the value of 1) and used this to predict the possible diversions out to 2040. Using the diversion levels for 2000, we apportioned the water diversion of the dry season for 2020e2040, based on normal flow conditions. Three scenarios were made for the normal flow, low flow and extremely low flow types. The future water diversions for the low flow and extreme low flow type were based on the assumption of 10% and 15% more than that of the normal flow type (Zhang et al., 2003) (Fig. 4C,D,E). 2.4. Tide level data The monthly average high tide levels at Wusong station (Fig. 1) for the period 1967e1984, were collected from the hydrology yearbook of the lower Changjiang (Changjiang Water Conservancy Committee, 1967e1984) and are plotted in Fig. 5A together with discharge. Fig. 5B shows the relationship between high tide level at Wusong and

discharge at Datong, separated into the wet and dry season flows. The tidal levels in the dry season were separated to 11 height categories at intervals of 5 cm from 2.81 m to 3.38 m, and correlated with the dry season discharges (Fig. 5C). This shows that discharge from upstream can influence the high tide level by up to 0.6 m. The mean annual high tide level at Wusong station for the period 1912e1994 was sourced from Chen (1990) and Zheng and Yu, 1996. The Wusong Datum level was used in this study. The tidal levels were calibrated with the Sheshan Datum to remove the effect of land subsidence. The difference in high tide level between the annual and dry season tidal levels is 0.31 m over the period 1967e1985. This value was subtracted from the annual average values dataset to provide tide level fluctuations in the dry seasons of 1912e1994 (Fig. 5D). We used the SAROS cycle of approximately 19 years to smooth the dry season tidal levels (Fig. 5E), and then applied the regression equation of high tide level against year (1934 ¼ 1) to predict the future high tide water levels in 2020, 2030 and 2040 (Fig. 5E). Combining the equation that describes the effect of upstream discharge on tidal level (Hd ¼ 0.00002Qdþ2.6729, Fig. 5C) and the equation that describes the change of the high tide levels through time (Hd ¼ 0.003n þ 2.770, Fig. 5E) allows us to calculate the future monthly discharge into the estuary, which correlates with different water level positions (Fig. 4F). The high tide level of 3.06 m in the dry season was simulated by the authors for 2010 as tidal level data for that year are not available (Fig. 4F). This correlates to an average discharge 16,703 m3 s
1 for the dry season of 2010 as recorded at Datong station.

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

5

6000 to 10,000 m3 s
1 (Fig. 2D). At discharges above 10,000 m3 s
1, the frequency rapidly drops to ca. 0.2% as discharge rises to ca. 30,000 m3 s
1, and further to nearly 0.1% as discharge increases to ca. 50,000 m3 s
1. The cumulative frequency plot (Fig. 2E) shows that 50% of flows are less than or equal to 15,000 m3 s
1 and 20% are less than or equal to 10,000 m3 s
1. 3.2. Correlations among discharge, salinity and time duration There is a negative correlation between discharge and salinity during the dry season. Salinity at the three stations decreases from 6.14 to nearly zero, associated with an increase in discharge from 6000 to 20,000 m3 s
1 (Fig. 3A,B). Salinity is frequently greater than 1.81 when discharge falls below the threshold of 15,000 m3 s
1. Also of note is that the salinity corresponding to 15,000e20,000 m3 s
1 at the three stations showed a spatial difference, i.e. from upstream to downstream, from 0.18 to 0.36 at Gaoqiao, to 0.36e0.90 at Chenhang station, and to 0.18e2.71 at Chongtou station (Fig. 3)A1eA3. Fig. 3B shows that there is a negative relation between mean salinity and discharge at all three stations and this is the case also for minimum and maximum salinities. As shown in Fig. 3C, higher salinities persist for significantly longer durations than lower salinities. 3.3. Influence of water diversion projects on discharge into estuary

Fig. 4. A) monthly actual water diversions in the Changjiang estuary in 2000; B) water diversions in 2000 and 2005e2010, and predicted water diversions in 2020, 2030, and 2040; and C,D,E) prediction of monthly water diversion in the dry season for normal flow, low flow and extreme low flow types, 2020, 2030, and 2040.

3. Results 3.1. Discharge effects The daily discharge recorded at Datong Station (Fig. 2A) averages 25,500 m3 s
1 and shows no trend through time. The 4 types of discharge described above, i.e. high flow, normal flow, low flow and extreme low flow, are all >15,000 m3 s
1 during the wet season (Fig. 2B). This contrasts with the dry season when the extreme low flow discharge is always below 15,000 m3 s
1, and the low flow and normal flow discharges are below 15,000 m3 s
1 for 70% and 50% of the dry season respectively (Fig. 2C). The frequency distribution of daily discharge in the dry season shows a rapid rise from 0.1 to 1.7% over the range of discharge from

Of the 739 water diversion projects operating in 2000, 581 projects are located along the Changjiang river bank in Jiangsu Province, and the rest are spread along the river bank within Anhui Province and the Shanghai municipality. The designed capacity for water diversion of these projects in 2000 totals 788  109 m3 (24,979 m3 s
1), but the actual monthly water diversion ranges from 2.5 to 6.0  109 m3 (800e2200 m3 s
1) (Fig. 4A), and the yearly total is 38.6  109 m3 (1250 m3 s
1) (Fig. 4B). The actual water diversions over the period 2000 to 2010 increased gradually from 38.6e50.9  109 m3 (Fig. 4B). Using the regression shown in Fig. 4B, the predicted annual water diversions for 2020, 2030 and 2040 are 56.2  109 m3 (1800 m3 s
1), 59.2  109 m3 (1900 m3 s
1) and 61.3  109 m3 (2000 m3 s
1), respectively. The future monthly water diversions in the dry season, apportioned according to the pattern in 2000 (normal flow; Fig. 4A), will be more than that of 2000, shown as below. Normal flow type: Water to be more diverted in 2020, 2030 and 2040 is 0.6e1.8  109 m3 (200e600 m3 s
1), 0.75e2.1  109 m3 (250e700 m3 s
1), and 0.9e2.4  109 m3 (300e800 m3 s
1), respectively. Low flow type: Water to be more diverted in 2020, 2030 and 2040 is 0.9e2.1  109 m3 (300e700 m3 s
1), 1.05e2.4  109 m3 (350e800 m3 s
1) and 1.2e2.7  109 m3 (400e900 m3 s
1), respectively. Extreme low flow type: Water to be more diverted in 2020, 2030 and 2040 is 1.2e2.4  109 m3 (400e800 m3 s
1), 13.5e2.7  109 m3 (450e900 m3 s
1);and 1.5e3  109 m3 (500e1000 m3 s
1), respectively (Fig. 4C,D,E). 3.4. Sea level rise Our data (1967e1985) has shown an in-phase fluctuation between discharge and high tide level at the monthly scale (Fig. 5A). The minimum and maximum discharges ranging from 10,000e60,000 m3 s
1at Datong Station correspond to the high tide levels of 2.7e3.8 m recorded at Wusong station (Fig. 5A,B). The correlation between water level and discharge in the dry season shows that a water level difference of 0.6 m can be found between

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

6

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

Fig. 5. A) In-phase monthly fluctuations of high tide levels at Wusong station and discharge record at Datong (1967e1984); B) relation between monthly high tide water level and discharge (1967e1984); C) regression of high tide water level on discharge in the dry season (note: 0.6 m water level difference between the neap and spring tides); D) sea-level rise as recorded by the high tide levels at Wusong station (1912e1994); E) 19-yr smoothing of dry season water level (SAROS cycle) and predicted water levels in 2020, 2030 and 2040; and F) apparent discharge differences between future water levels of the dry season (details given in text).

the minimum and maximum levels, depending on discharge (Fig. 5B,C). The long-term record (1912e1995) of high tide levels in the dry season clearly shows a sea-level rise of ca. 0.7 m (Fig. 5D). A linear regression equation (Fig. 5E) based on the 19-point SAROS smoothing can be used to obtain the future high tide water levels in the dry season, i.e. 3.10 m, 3.14 m and 3.17 m for 2020, 2030 and 2040, respectively (Fig. 5D,E). On the basis of the relationship in Fig. 5C, the apparent freshwater discharges at these times would be 16,871, 17,040, and 17,208 m3 s
1 respectively (Fig. 5F). These apparent discharges differ to those of 2010 by 169 m3 s
1, 337 m3 s
1, and 506 m3 s
1 (Fig. 5F). Finally, these differences can be subtracted from the monthly-averaged discharge of the dry season of the past 51 yrs, in order to estimate the salinity and time duration in future scenarios (Fig. 3B,C; Fig. 6). 4. Discussion 4.1. Freshwater sources e dry season shortage The Changjiang is now the main source for the supply of freshwater to Shanghai and is likely to become more important in

the future. Demand is rising as the population (presently >23 million) grows, the local sources are being heavily polluted, and there are threats developing in the Changjiang system in the dry season related to water diversions and saltwater intrusions (Chen et al., 2001; Gong et al., 2013). Undoubtedly, there is sufficient freshwater available in the Changjiang basin as the annual precipitation averages 1042 mm across the whole basin (Finlayson et al., 2013), and the annual discharge flowing into the estuary averages 900  109 m3 (Chen et al., 1988), with low interannual variability and long term stability (Fig. 2A). However, the seasonal differences in freshwater discharge are quite substantial (Fig. 2B,C). Obviously, no saltwater intrusions occur in the wet season, but in the dry season, when discharge falls below 15,000 m3 s
1, they occur frequently (Mao et al., 2000; Shen et al., 2003). Therefore, the dry season is the focus in this discussion, concentrating on the three flow types: i.e. normal flow, low flow and extreme low flow (Fig. 2C), with discharge of <15,000 m3 s
1 for 50%, 70% and 100% of the time, respectively (Fig. 2C). The high flow type in the dry season is not discussed here, since flow remains above 15,000 m3 s
1, with no significant saltwater intrusions occurring (Fig. 2C). Clearly, flows in the range 8000e15,000 m3 s
1 are critical, as this range includes the highest frequencies of flows in the dry

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

7

Fig. 6. Future predictions of the dry season low flow and extremely low flow types. A1-A2) predicted discharge for 2020, 2030, and 2040; B1eB2) predicted salinity for 2020, 2030, and 2040. Salinity deviations in B2 (Gaoqiao station, representing Qingcaosha reservoir) range from 20 to 30%, 10e12% and 4e5%, were calculated from the formulae presented in Fig. 3B1; and C1eC2) predicted duration of saltwater intrusion, 2020, 2030 and 2040. The duration ranging from 20 to 65, 75e90, and 120e128 days at Gaoqiao station was calculated by the formula in Fig. 3C1, by substitution of predicted discharge and salinity presented in Fig. 6A2 and B2 (Gaoqiao station).

season (Fig. 2D). The cumulative frequency plot (Fig. 2E) indicates that discharge below 15,000 m3 s
1 occurs 50% of the time in the dry season. This range of discharges is therefore important in relation to the upstream water diversion projects and future sea level rise. 4.2. Salinity distribution in relation to freshwater availability The spatial differences of salinity at Gaoqiao, Chenhang and Chongtou stations reveal significant features in the pattern of saltwater intrusion (Fig. 3A). The higher salinities at discharges less than 15,000 m3 s
1 at Gaoqiao and Chongtou station than at Chenhang indicates that the saltwater intrusion occurs not only directly from the South Branch but also from the North Branch around Chongming Island (Figs. 1 and 3A) (Shen et al., 2003; Xue., 2009). In the range of discharge 15,000e20,000 m3 s
1, the salinity gradually decreases from the inner river mouth at Chongtou seaward down the South Branch to Chenhang, and Gaoqiao (Fig. 1). This demonstrates the importance of the saltwater intrusions via the North Branch in the overall salinity levels in the main river channel, the South Branch, where the Shanghai Municipality sources some 70% of its freshwater supplies (Mao et al., 2000; Shen et al., 2003; Xue et al., 2009). A salinity level of 0.45 is the limit for freshwater intake to the Shanghai supply system (Shen et al., 2003). It needs also to be noted

that Shanghai only has the capacity to store up to 68 days of freshwater supply. The spatial differences of salinity distribution show that at Gaoqiao (Qinchaosha reservoir), while discharge is > 15,000 m3 s
1, the salinity is below the threshold value of 0.45, while at Chongtou (Dongfengxisha reservoir) and Chenhang station, significant occurrences of water with salinity >0.45 occur (Fig. 3A1eA3). This is a direct consequence of the saltwater intrusions that come via the North Branch. Salinity levels >0.45 occur at all three stations while the discharge is < 15,000 m3 s
1 (Fig. 3A1eA3) and this probably accounts for ca. 90 days during the dry season (Fig. 2E). The 90 days (50% probability; Fig. 2E) with salinity >0.45 were summed from discrete events of saltwater intrusions occurring at the river mouth station. However, our analyses show not only the negative correlation of salinity with discharge (Fig. 3B), but also the fact that the higher the salinity, the longer the time duration (Fig. 3C). In particular, the results derived for the 3 levels (minimum, sectional mean and maximum; Fig. 3B) would allow high salinities to persist longer than the 68 day storage limit (see below for further discussion). 4.3. Water diversion e present and future case Some 739 water diversion projects have operated along the lower Changjiang River below Datong station over recent decades.

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

8

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

Our investigations have verified that the design capacity of the water diversions, reaching 788  109 m3, is already equal to 88% of the mean annual discharge at Datong station. It is clearly unimaginable that all these diversions could operate at capacity at the same time. In fact, the monthly water diversion in 2000 ranged from 700 to 1500 m3 s
1 (1.8e3.9  109 m3) (FebruaryeApril) to 2200 m3 s
1 (5.7  109 m3) (May) (Fig. 4A). Also, the annual water diversion in 2000 was 1250 m3 s
1 (38.6  109 m3), which had increased continuously to 1650 m3 s
1 (50.9  109 m3) in 2010 (Fig. 4B). Assuming this increase continues at its historic rate, then, it will reach 1800 m3 s
1 (56.2  109 m3), 1900 m3 s
1 (59.2  109 m3), and 2000 m3 s
1 (61.3  109 m3) in 2020, 2030 and 2040 respectively (Fig. 4B). What we need to reiterate here is that water return to the river in the dry season is quite low (The Changjiang Water Conservancy Committee, 2007; Zhang et al., 2012), and thus negligible to our simulation results given below. Water diversion that occurs between May and October (wet season) will not create problems in relation to saltwater intrusions because of the high discharge. However, the quantity of water diverted in the dry season is a potential threat to freshwater supplies in the river mouth, especially considering there is a 50% probability that the discharge will be in the range of 8000e15,000 m3 s
1 (Fig. 2E). According to our predictions, there could be a further water diversion of ca. 800e2200 m3 s
1; 1000e2800 m3 s
1 and 1200e3200 m3 s
1 in the dry season from 2020 to 2040, for the normal flow, low flow and extreme flow types, respectively (Fig. 4C,D,E). These values will be subtracted from the monthly average discharge in the dry season in the discussion of future scenarios below (Fig. 2C; Fig. 6). 4.4. Sea level rise e equivalent to increase in salinity and decrease in discharge We have shown that there is an in-phase fluctuation between high-tide level and discharge at the monthly scale (Fig. 5A,B). While this fluctuation at the river mouth does not represent sea level fluctuation, the long-term record of water level can help track the trend in sea level rise. The average high tide level of the dry season from 1912 to 1995 shows a sea level rise of 0.7 m (Fig. 5D). The simulated water levels of the dry season in 2020, 2030, and 2040 that will reach 3.10 m, 3.14 m and 3.17 m are averaged on the basis of the SAROS cycle (~19 yrs) (Fig. 5D, E), meaning a future sea level rise of 0.12 m (Fig. 5E). This prediction based on a linear regression is more conservative than that if an exponential curve had been fitted to the SAROS smoothed data in Fig. 5D. The values of sea level rise fit with the long-term observation by the State Oceanic Administration, China (http://www.soa.gov.cn). The simulated water levels need to be considered in relation to the tidal level fluctuation of 0.6 m as recorded at Wusong gauging station (Fig. 5C, E). Increasing sea levels in the future will drive up salinity in the river mouth area, equivalent to a decrease in discharge. This relationship has been clearly established in Fig. 3B. The discharge differences between 2010 and 2020, 2010e2030 and 2010e2040 worked out by using the simulated water levels (Fig. 5F) need to be subtracted from the monthly average discharges of the past 51 yrs. The results will be used to discuss the effect of reduced discharge, in terms of increase in salinity and related time duration in the future scenarios for 2020, 2030 and 2040. 5. Future scenarios The ultimate purpose of this study is to understand the nature of the salinity distribution and its persistence over time in the

Changjiang River mouth. This is of particular concern given that only 68 days supply of freshwater can be stored in Qingcaosha reservoirs in the Changjiang estuary. Therefore, our scenario study only applies to Gaoqiao station (Qingcaosha reservoir). The two low flow types (‘low flow’ and ‘extreme low flow’) are considered here in relation to future discharge, salinity and related time duration. Fig. 6A shows the monthly average discharge for the low flow and extremely low flow months (November to April) for the past 51 yrs, and those for 2020, 2030 and 2040 after subtraction of the amount lost to future water diversions and the effects of sea level rise, which will increase the strength and duration of salinity for any given discharge level (Fig. 4C,D,E; Fig. 5F). The results show that: 1) during low flow, dry season discharge will be reduced by 2100e2600 m3 s
1 (1930e2094 m3 s
1 due to water diversion [averaged for 6 months of the dry season]; 169e506 m3 s
1 due to sea level rise) in 2020e2040, and 2) during extreme low flow it will be reduced by 2200e2700 m3 s
1 (2030e2194 m3 s
1 due to water diversion [averaged for 6 months of the dry season]; 169e506 m3 s
1 due to sea level rise) for the same duration (Fig. 6A1). In summary, the contribution of increasing salinity at Qingcaosha reservoir due to future sea level rise could be ca. 8.0e19.4%. Our projection was also made for the worst case of February 2020e2040 (extreme low flow), showing a total reduction in discharge of 1357e1805 m3 s
1, of which 1188e1299 m3 s
1 (87.6e72%) is due to water diversion, and 169e506 m3 s
1 (12.4e28.0%) is due to sea level rise (Fig. 6A2). These results are based on the assumption that future discharge at Datong will remain at 7000e11,000 m3 s
1 and 6500e9000 m3 s
1 for low flow and extreme low flow, respectively, in the dry seasons of 2020e2040 (Fig. 6A1eA2). Fig. 6B,C shows the salinity and time duration simulation. It should be noted that the numbers of days of duration of saltwater intrusion were derived from the salinity levels (Fig. 3B,C). Taking Qingcaosha reservoir (represented by Gaoqiao station; Figs. 1 and 6) as the most important case, we can calculate the duration of high salinity using the minimum and maximum salinity values shown in Fig. 3B1. In the worst case with the lowest discharge in February, taking deviations ranging from 20 to 30%, 10e12% and 4e5% (Fig. 6B2; the higher the salinity, the less the deviation, see Fig. 3B), we estimate the longest time that the salinity would exceed 0.45 would be 20e65, 75e90, and 120e128 days for 2020, 2030, and 2040, respectively (Fig. 6C2). While the findings represent a useful contribution to our general understating of the implications of the dry season on water resources in the Yangtze Estuary system, there are a few limitations that need to be addressed in further research. These include 1) a continuous on-site measurement in the dry season is needed at the reservoir site, in order to establish a direct correlation between discharge and salinity, and 2) a great effort needs to be paid to seek a correlation between historical salinity and maximal water level in order to better assess the impact of future sea-level rise on salinity in the river mouth area. 6. Conclusions This hydrological study focuses on the future availability of freshwater in the Changjiang estuary, on the basis of an examination of the available data: discharge, salinity, water diversion projects and sea level rise. Owing to the pressure from increasing population and heavy pollution of surface water available locally, the city of Shanghai, with >23 million people, has become increasingly dependent on the Changjiang estuary. The new Qingcaosha reservoir in the river mouth area was designed to store 68 days supply of freshwater as protection against the longest duration of saltwater intrusions in the records.

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007

M. Li et al. / Estuarine, Coastal and Shelf Science xxx (2014) 1e9

As our analysis shows, the situation is first complicated by the passage of saltwater intrusions through the North Branch which exacerbates the problem by introducing saltwater higher in the estuary than if intrusions were confined to the South Branch alone, which carries the main freshwater discharge. The situation is further complicated by the diversion of flow from the Changjiang downstream of the Datong gauging station, so that the Datong flow record can not be relied on for the prediction of saltwater intrusion events since significant, though not well known or recorded, volumes of water are lost from the river downstream of Datong. This means that the estuary is more susceptible to saltwater intrusions in the low flow season than the gauged flows at Datong would indicate. Finally, sea level rise, driven by both global warming and the sinking of the Changjiang delta, increases the impact of saltwater intrusions in the dry season. The future security of freshwater supply to Shanghai is dependent on, first, a better understanding of the hydrodynamics of the estuary; second, the development of a strategy to combat this problem; and thirdly, the political will to take the necessary actions. Acknowledgements This research is supported by the Australian Research Council (Grant No: P110103381), the China National and Natural Science Foundation (Grant No. 41271520), the Creative Research Groups of the China Natural Science Foundation (Grant No. 40721004), and The Chongming Tourism Administration (Grant No. 48102860). References Le, Q., 2012. Research and analysis of the water supply potentiality of Qingcaosha reservoir during the low water period of Yangtze River. Water Wastewater Eng. 38 (9), 123e127 (in Chinese). Li, L., 2012. Spatial-temporal Dynamic Characteristics of Saltwater Intrusion in the Changjiang Estuary. The East China Normal University, pp. 100e161. Ph.D thesis (in Chinese). Caitlin, M.C., Benjamin, S.H., Mike, W.B., Carrie, V.K., 2009. Understanding and managing human threats to the coastal marine environment. Ann. N. Y. Acad. Sci. 1162, 39e62. Chen, X.Q., 1990. Sea level changes from 1922 to 1987 in the Changjiang River mouth and its significance. Acta Geogr. Sin. 45 (4), 387e398. Chen, J.Y., He, Q., 2009. Study of Water Security of Shanghai, a Special Reference to the Drought Year 2006. Ocean Press, Beijing, China, pp. 15e82 (in Chinese). Chen, J.Y., Shen, H.T., Yun, C.X., 1988. Processes of Dynamics and Geomophology of the Changjing Estuary. Shanghai Scientific and Technical Publisher, Shanghai, China, pp. 1e62 (in Chinese). Chen, Z.Y., Huang, Y.H., Zhou, T.H., Tang, E.X., Yu, Y.F., Tian, H., 1991. A preliminary study on mean sea level of the Changjiang River estuary. Oceanol. Limnol. Sin. 22 (4), 315e320. Chen, X.Q., Zong, Y.Q., Zhang, E.F., Xu, J.G., Li, S.J., 2001. Human impacts on the Changjiang Yangtze River basin, China, with special reference to the impacts on the dry season water discharges into the sea. Geomorphology 41, 111e123. Dai, Z.J., Liu, J.T., Fu, G., Xie, H.L., 2013. A thirteen-year record of bathymetric changes in the North Passage, Changjiang (Yangtze) estuary. Geomorphology 187, 101e107. Li, M.T., Chen, Z., Yin, D.W., Chen, J., Wang, Z.H., Sun, Q.L., 2011. Morphodynamic characteristics of the dextral diversion of the Yangtze River mouth, China: tidal and the Coriolis force controls. Earth Surf. Process. Landf. 36, 641e650.

9

Finlayson, B.L., Barnett, B., Wei, T.Y., Webber, M., Li, M.T., Wang, W.Y., Chen, J., Xu, H., Chen, Z., 2013. The drivers of risk to water security in Shanghai. Reg. Environ. Change 13, 329e340. Gong, W.P., Shen, J., Jia, L.W., 2013. Salt intrusion during the dry season in the Huangmaohai Estuary, Pearl River Delta. J. Mar. Syst. 111e112, 235e252. Hong, B., Shen, J., 2012. Responses of estuarine salinity and transport processes to potential future sea-level rise in the Chesapeake Bay. Estuar. Coast. Shelf Sci. 104e105, 33e45. Huang, W.R., Foo, S., 2002. Neural network modeling of salinity variation in Apalachicola River. Water Resour. 36 (1), 356e362. Mao, Z.C., 1994. The salt intrusion character of the South Channel in the Changjiang mouth. Shanghai Water Conserv. 4, 22e32 (in Chinese). Mao, Z.C., Shen, H.T., Yao, Y.D., 1993. Analysis of sources of saltwater intrusion along south bank of south branch of the Yangtze River estuary. Mar. Sci. Bull. 12 (1), 17e26 (in Chinese). Mao, Z.C., Shen, H.T., Xu, P.L., 2000. The Pattern of saltwater intruding into the Changjiang Estuary and the utilization of freshwater resources. Acta Geogr. Sin. 55 (2), 243e250 (in Chinese). Mao, Z.C., Shen, H.T., Xiao, C.,Y., 2001. Saltwater intrusion patterns in the Qing Caosha area Changjiang River estuary. Oceanol. Limnol. Sin. 32 (1), 58e66 (in Chinese). Schroeder, W.W., Dinnel, S.P., Wiseman, W.J., 1990. Estuaries and Coasts. Salinity stratification in a river-dominated estuary. Estuaries 13 (2), 145e154. Shen, H.T., Mao, Z.C., Zhu, J.R., 2003. Saltwater Intrusion in the Changjiang Estuary. China Ocean Press, Beijing, pp. 20e175 (in Chinese). Tang, J.H., Xu, J.Y., Zhao, S.W., Liu, W.Y., 2011. Research on saltwater intrusion of the south branch on the Changjiang Estuary based on measured data. Resour. Environ. Yangtze Basin 20 (6), 677e685 (in Chinese). The Changjiang Water Conservancy Committee, 1950e1986. Hydrology Year Book of the Changjiang, China, pp. 10e31620 (in Chinese, unpublished). The Changjiang Water Conservancy Committee, 1997e2011. Hydrology Year Book of the Changjiang, China, pp. 10e31620 (in Chinese, unpublished). The Changjiang Water Conservancy Committee, 1967e1984. Hydrology Year Book of the Lower Changjiang, China, pp. 10e10830 (in Chinese, unpublished). The Changjiang Water Conservancy Committee, 2007. The Report of Water Resource Assigning Plan in the Lower Changjiang below Datong in Dry Season, pp. 1e156 (in Chinese, unpublished). The Water Resources Department of Anhui Province, 2005e2011c. The Water Resource Bulletin of Anhui Province, pp. 1e200 (in Chinese). http://www.ahsl. gov.cn. The Water Resources Department of Jiangsu Province, 2005e2011a. The Water Resource Bulletin of Jiangsu Province, pp. 1e210 (in Chinese). http://szyc. jswater.gov.cn. The Water Resources Department of Shanghai, 2005e2011b. The Water Resource Bulletin of Shanghai, pp. 1e186 (in Chinese). http://www.shanghaiwater.gov.cn. Wang, Y.G., Huang, H.M., Li, X., 2008. Critical discharge at Datong for controlling operation of South-to-North water transfer project in dry seasons. Water Sci. Eng. 1 (2), 47e58. Wolanski, E., McLusky, D., 2011. Treatise on Estuarine and Coastal Science. Elsevier, Amsterdam, p. 12. Wu, H., 2006. Saltwater Intrusion in the Changjiang Estuary. East China Normal University, pp. 1e205. Ph.D thesis (in Chinese). Xu, J.Y., Yuan, J.Z., 1994. The research on salt-water intrusion into the south branch of the Yangtze estuary. Hydrology 5, 1e5 (in Chinese). Xue, P., Chen, C., Ding, P., Beardsley, R.C., Lin, H., Ge, J., Kong, Y., 2009. Saltwater intrusion into the Changjiang River: a model-guided mechanism study. J. Geophys. Res. 114, C02006. http://dx.doi.org/10.1029/2008JC004831. Zhang, E.F., Chen, X.Q., Wang, X.L., 2003. Water discharge changes of the Yangtze River downstream from Datong during the dry season. J. Geogr. Sci. 13, 355e362 (in Chinese). Zhang, E.F., Savenije, H.H.G., Chen, S.L., Chen, J.Y., 2012. Water abstraction along the lower Yangtze River, China, and its impact on water discharge into the estuary. Phys. Chem. Earth 47-48, 76e85. Zheng, D.W., Yu, N.H., 1996. Research on the long-range forecast for the trend of sea level rising in the area of Shanghai. Ann. Shanghai Observ. Acad. Sin. 17, 36e42 (in Chinese). Zhu, J.R., Wu, H., Li, L., Wang, B., 2010. Saltwater intrusion in the Changjiang Estuary in the extremely drought hydrological year 2006. J. East China Norm. Univ. Nat. Sci. 4, 1e4 (in Chinese).

Please cite this article in press as: Li, M., et al., Water diversion and sea-level rise: Potential threats to freshwater supplies in the Changjiang River estuary, Estuarine, Coastal and Shelf Science (2014), http://dx.doi.org/10.1016/j.ecss.2014.07.007