ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 0

Available online at w

journal homepage: www.elsevier .com/locate/watres

Effects of vertical mixing on phytoplankton blooms in XiangxiBay of Three Gorges Reservoir: Implications for management

Liu Liu a, Defu Liu a,*, David M. Johnson b, Zhongqiang Yi a, Yuling Huang a

aEngineering Research Center of Eco-Environment in TGR Region, Ministry of Education, College of Hydraulic & Environmental Engineering,

China Three Gorges University, Yichang 443002, Chinab School of Natural Science & Mathematics, Ferrum College, VA 24088-9000, USA

a r t i c l e i n f o

Article history:

Received 23 August 2011

Received in revised form

5 January 2012

Accepted 23 January 2012

Available online 31 January 2012

Keywords:

Eutrophication

Phytoplankton succession

Mixing depth

Thermal stratification

Density current

Water level fluctuations

* Corresponding author. Tel.: þ86 (0) 7176392E-mail address: dfliu@ctgu.edu.cn (D. Liu

0043-1354/$ e see front matter ª 2012 Elsevdoi:10.1016/j.watres.2012.01.029

a b s t r a c t

Since the initial filling of Three Gorges Reservoir (TGR), serious phytoplankton blooms have

occurred in its tributary bays. Cyanobacteria blooms have been observed in a number of

tributary bays and threaten the drinking water security of residents in the TGR region. To

identify the key factors controlling phytoplankton blooms in tributary bays and propose an

effective management strategy, a one-year water quality study (November 2009 to October

2010) was conducted in Xiangxi Bay (XXB) of TGR. The results show that a rapid decrease in

mixing depth is associated with the spring bloom, fading of the fall bloom occurs with the

rapid increase in mixing depth, and an extremely shallow mixing depth is associated with

cyanobacteria blooms that predominate during the summer. Development of thermal

stratification in XXB is the major cause of seasonal variation in mixing depth and density

current intrusion from TGR is the major cause of short-term variation in mixing depth. The

seasonal thermal stratification of XXB is disrupted by sufficiently large water level fluc-

tuations in TGR. The density current is lifted from mid-depths to the surface and chloro-

phyll a concentrations rapidly decrease in response. Based on these findings, a conceptual

model is proposed for a management strategy to control phytoplankton blooms in tributary

bays via controlled releases from TGR.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction Numerous investigations of phytoplankton blooms in the

Since the initial filling of Three Gorges Reservoir (TGR) in June

2003, intense phytoplankton blooms have been observed in

some tributary backwaters. Both the frequency of phyto-

plankton blooms and the numbers of tributaries involved

increased as the normal high water level shifted upwards

(Ministry of Environmental Protection of China, 2011). Serious

cyanobacteria blooms have been observed in a number of

tributaries (Ministry of Environmental Protection of China,

2011), threatening the drinking water security of residents in

the TGR region.

100.).ier Ltd. All rights reserved

tributary bays of TGR have been conducted. Nutrient levels in

TGR are high and difficult to control because nitrogen and

phosphorus loading of several major tributaries above the

Three Gorges Dam come mainly from non-point source

pollution (Zheng et al., 2009). Minimum values of total

nitrogen and total phosphorusmeasured in TGRwere 0.80 and

0.07mg/L, respectively (Zhou et al., 2011). Intrusions from TGR

maintain high nutrient levels in tributary bays (Luo et al., 2007;

Ji et al., 2010) and nutrients are not the limiting factor for

phytoplankton blooms (Xu et al., 2010a). Xu (Xu et al., 2010b)

identified water temperature as the driving factor of spring

.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 02122

blooms in Xiangxi Bay (XXB). However, intense cyanobacteria

blooms were observed in other tributary bays of TGR even

during the winter (Cao et al., 2009) and evidence that

temperature is the primary driver of spring blooms in tribu-

tary bays is not convincing. Li and Liao (Li and Liao, 2003)

concluded that decreased flow in TGR tributaries is the

primary driver for phytoplankton blooms. Seasonal thermal

stratification is strongly developed in TGR bays but weak in

TGR. These different hydrological conditions explain the lack

of phytoplankton blooms in TGR and the frequent blooms in

its tributary bays (Wang et al., 2011a).

The Critical Depth Model proposed by Sverdrup (Sverdrup,

1953) emphasizes that a decreasedmixing depth promotes the

onset of spring blooms in the ocean and this concept has been

applied to freshwaters. Seasonal development and disap-

pearance of temperature stratification in lakes result in

seasonal deepening of mixing depth and stimulate seasonal

onset and succession of phytoplankton blooms (Berger et al.,

2007; Fonseca and Bicudo, 2008). The ratio of euphotic depth

(Zeu) to mixing depth (Zmix) is often used as a measure of

illumination in the mixed layer (Jensen et al., 1994). Cushing

(Cushing, 1989) has succeeded in predicting seasonal succes-

sion of phytoplankton in Lake Ontario using Zeu/Zmix. A

phytoplankton bloom will generally not occur if Zeu/Zmix is

lower than the “critical value” (Scheffer et al., 1997).

Mixing depth of oceans and lakes generally depends on

thermal stratification and wind-induced circulation (Farmer

and Carmack, 1981; Kling, 1988). Yi (Yi et al., 2009) reported

significantwater temperaturestratification inXXBduringspring

and summer but the density current in XXB (Ji et al., 2010a)may

modify the natural condition because the layer receiving the

density current intrusion is well mixed. When an overflow

occurs, there is a sudden increase in mixing depth; plunging

depth of the density current may be an important determinant

of mixing depth in XXB. Research (Ji et al., 2010b) indicates that

the continuing increase ofwater level in TGRduring the late fall

impoundment period influences plunging depth of the density

current. Nonetheless, discharge restrictions are created by the

need forwaterdownstreamfromTGRduring thedry seasonand

the need for a pre-flood season drawdown of TGR. Thus,

increases in the long-term water level are limited and leave

short-term manipulation of water level as the only acceptable

hydrological option for controlling phytoplankton blooms. In

this study, the feasibility of using temporary water level fluctu-

ations in TGR to change plunging depth and Zeu/Zmix to exert

control on phytoplankton blooms in XXB is investigated.

Because of an incomplete understanding of phytoplankton

bloom mechanisms in the tributary bays of TGR, a practical

management strategy has not been developed. Resolving

phytoplankton bloom problems in tributary bays of TGR by

controlling nutrient loads can only be a long-term goal

because the TGR watershed is very large (1 million km2)

(Huang and Li, 2006). A practical measure for short-time (and

emergency) control of phytoplankton blooms is needed. The

goal of this research was to gain a fuller understanding of the

phytoplankton bloom mechanisms by studying the vertical

mixing regimes of XXB and the response of phytoplankton

blooms in XXB to transitory water level fluctuations in TGR. A

management strategy for phytoplankton blooms in tributary

bays is recommended based on the results.

2. Materials and methods

2.1. Study area description and sampling methods

Xiangxi River, the largest tributary in the lower reach of TGR,

has a watershed area of 3095 km2, length of 94 km and annual

average flow 47.4 m3/s (Huang and Li, 2006). It originates in

Shennongjia National Park located in northwest Hubei Prov-

ince, and flows southward through Xingshan and Zigui

Counties and into TGR at Xiangxi Town (32 km from Three

Gorges Dam). After initial impoundment of TGR in June 2003 to

water level of 135 m, a deep riverine bay formed in XXB as the

lower 26 km were submerged by backwater; the backwater

reach extended to 40 km when the reservoir was filling to

normal water level (175 m) in October 2010. The sampling site

is located at the middle reach of XXB and 19 km from the

confluence of XXB and TGR (Fig. 1) because it receives small

direct impacts from TGR and upstream inflow. Depth of this

site varies between 19 and 49 m in response to water level

fluctuations of TGR throughout a year. Daily sampling was

conducted from November 2009 to October 2010.

Water level and discharge data were provided by China

Three Gorges Corporation. The TGR operational cycle can be

divided into four stages according to water level and discharge

(Fig. 2): dry season (November to April), drawdown (May to

early June), flood season (June to August) and impoundment

(September to October). Wind speed and direction data were

collectedusinga continuouswindspeedanddirectionrecorder

(ZDR-1F, China) fixed on the platform located at the middle

reachofXXB.The instrumentwasoperatedat a 1minsampling

interval. Secchi Depth (SD) was measured using a Secchi Disk

and Photosynthesis Active Radiation (PAR, 400e700 nm) was

measured by an underwater illuminometer (IL1400A, China).

Surfacewater samples (350mL)were collectedand transported

to the laboratory for chlorophyll a concentration (Chl. a) anal-

ysis using the national standard method (Wang et al., 2002).

Additional surface water samples (1 L) were collected for

phytoplankton identification and enumeration (Olympus

microscope, Japan; ShinesoAlgacount,China) and thedatawas

expressed as phytoplankton density (cells/L).

Simultaneous profiles of velocity and water quality were

measured from a boat equipped with an Acoustic Doppler

Vector velocimeter (ADV; Nortek, Norway) and a Hydrolab DS

5X multi-probe sonde (Hach, USA). Parameters profiled

included depth, water temperature and turbidity. Vertical

resolution of the measurements was 1 m. Using the method

proposed by Ma (Ma et al., 2011), flow velocities in east (Ve),

north (Vn) and vertical (Vu) directions were measured at

depths of 0.5 and 2 m and then to the bottom at 2 m intervals.

2.2. Data analysis

Zeu is generally calculated as the corresponding depth where

underwater PAR is 1% of the surface (Kirk, 1994). A linear

regression formula between Zeu and SD was obtained using

the underwater PAR in 2010 (Zeu ¼ 1.8021SD þ 1.8182,

R2 ¼ 0.8404) to calculate Zeu. The depth corresponding to

surface temperature gradient � 0.2 �C/m (Lawrence et al.,

2000) was computed as Zmix in this study.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 0 2123

Layer depths of density currents were computed using

a set of moments with the upper limit of the underflow taken

as the zero velocity line for the integration (Garcıa et al., 2007).

The same method was applied to calculate layer depth and

depth of the interface of overflows and interflows.

Squared buoyancy frequency (N2) was computed using the

formula:

N2 ¼ ðg=r0Þðdr=dzÞ (1)

Where g is the gravitational acceleration ( g ¼ 9.8092597m/s2),

r0 is the water density at 3.98 �C and z is water depth using

water surface as origin of coordinates and vertical downwards

as positive direction.

Richardson number (Ri) was computed using the formula:

Ri ¼ N2=ðdU=dzÞ2 (2)

Where N2 is squared buoyancy frequency, U is synthesized by

Ve and Vn.

3. Results

3.1. Seasonal variations of euphotic depth and mixingdepth

Fig. 3 displays water depth (H), Zeu, Zmix, and Zeu/Zmix in

XXB during the study period. Zeu was larger than 5 m from

November 2009 to March 2010 and October 2010 but less

Fig. 1 e Map of study area and sampling site (areas submerged

and the areas submerged at an impoundment elevation of 175

interpretation of the references to colour in this figure legend, t

than 5 m in other months. The maximum value was

8.20 m in October and the minimum 2.87 m in August.

The seasonal cycling of Zmix in XXB was extreme with

mixing depth nearly to the bottom in January, decreasing

rapidly in April to a minimum of 1.08 m in August, and then

increasing again in the fall. The seasonal variation of Zeu/Zmix

was also large. During the winter dry season (November 2009

to March 2010), Zeu/Zmix was small (0.14e0.21) and began to

increase rapidly in April, peaking in May (1.77) and again in

August (2.66). Because seasonal variation of Zeu (2.87e8.20 m)

was much less than that of Zmix (1.94e39.04), the seasonal

variation of Zeu/Zmix was driven primarily by seasonal varia-

tion of Zmix.

3.2. Factors controlling euphotic depth and mixing depth

3.2.1. Factors influencing euphotic depthCorrelation of Zeu with turbidity and Chl. a is shown in Fig. 4.

The nonlinear correlation coefficient is higher than the linear,

and Chl. a (phytoplankton biomass) has the larger influence

on Zeu.

3.2.2. Influence of temperature stratification and wind onmixing depthFig. 5a indicates that temperature stratification in XXB began

in April, was strongest in August and began to diminish in

September. As shown in Fig. 5b, the velocity profiles of XXB in

winter (February) were extremely disordered and the flow

velocity very low (�0.06e0.02 m/s). In August the velocity

at an impoundment elevation of 145 m are tinted light blue

m are tinted green; isohypse resolution is 5 m). (For

he reader is referred to the web version of this article.)

Fig. 2 e TGR operational cycle: water level and inflow

discharge curve.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 02124

profile was stratified and indicates an interflow. The middle

layer (3e18 m) flowed into the bay while water flowed out of

the bay from the surface and bottom with a maximum

velocity of 0.13 m/s at the bottom.

Fig. 5c displays N2 and Ri profiles during weak temperature

stratification in winter (February 1, 2010) and strong temper-

ature stratification in summer (August 2, 2010). In February

2010, N2 of the water column ranged only from �5 � 10�5 to

2 � 10�4, displaying an increase from 30 m to the river bed. In

August 2010,N2 of thewater column ranged from 0 to 4� 10�3,

with large gradients in the surface and bottom layers and near

0 in the middle layer (12e22 m). In winter, temperature

stratification was weak and vertical stability low and, in

summer, temperature stratification was strong and vertical

stability high. In August, the low value of N2 in the middle

layer implies vertical stability was very low. This was a result

of the density current plunging into to middle layer and

consistent with the temperature profile in Fig. 5a and the

velocity profile in Fig. 5b.

In February 2010, the Ri profile (Fig. 5c) indicates low,

uniform values, until near the bottom. Below 27 m, there was

a sharp increase in Ri, exceeding 1 below 31m. In August, Ri for

theentirewater column inXXBranged from0.59 to188.21.With

the exception of the minimum value (0.59) occurring at the

bottom,Riwereall larger than1,withamaximumvalueat15m.

Both the Ri profile and N2 profile suggest that vertical

mixing in the upper 30 m in XXB was strong in winter

(February) and restrained in the summer (August) and that

turbulence was significantly lower in summer than in winter.

The peak value of Ri (15 m) and the minimum N2 values

0.0

0.5

1.0

1.5

2.0

2.5

3.0 0

10

20

30

40

50

Zeu

/Zm

ix

Dep

th (

m)

Time (YY/MM/DD)

Zeu Zmix H Zeu/Zmix

Fig. 3 e Seasonal variations of depth (H), Zeu, Zmix and Zeu/

Zmix in XXB during the study period.

(12e22 m) indicate that the density current plunged into XXB

as a laminar flow.

3.2.3. Influence of density current on mixing depthSeasonal variation in plunging depth of the density current

was found to occur because of seasonal change of tempera-

ture stratification in XXB (Ji et al., 2010a). Within the density

current, physico-chemical properties are uniform and density

current plunges alter the physical and chemical properties of

the corresponding layer of XXB. Fig. 6 indicates the effect of

TGR intrusion on temperature stratification and Zmix in XXB

over 10 days (August 22e31, 2010). There were continuing

density currents in XXB during this period (flood season).

From August 22e24 the density currents were thin interflows,

plunging close to the surface (4.44e5.93 m). The density

currents changed to overflows August 25e29 with increased

thickness (19.14e22.50 m) and then changed back to an

interflow again on August 30 as thickness and plunging depth

decreased (9.52e12.10 m).

Fig. 6b shows that temperature stratification in XXB was

affected by the intrusions. When plunging depth of density

current changed, the temperature profile at the corresponding

depth was altered. For example, the interflow produced two

inflections in the temperature profile from August 22e24, but,

from August 25e29, water temperature at the surface was

approximately uniform as a result of the intrusion rising from

the middle layer to the surface.

Fig. 6c shows the upper and lower interfaces of the density

current and their depths relative to Zeu and Zmix. When the

interflow occurred from August 22e24, Zmix was shallow and

decreasing (5.60/2.23 m). Following the transition of the

intrusion from the middle layer to the surface on August 25,

there was a large increase in Zmix and it approached the lower

interface and increased with density current thickness from

August 25e29. On August 30, the intrusion plunged to the

middle layer again, but Zmix still reached to the lower interface

of the density current because of incomplete development of

surface temperature stratification. Zmix then declined sharply

and returned to above the upper interface of interflow until

the temperature stratification was fully developed on August

31. Variation of Zeu was small from August 22e31, implying

that the influence of intrusions is slight (2.72e3.26 m during

interflow, 2.72e3.17 m during overflow). The results indicate

that plunging depth of the density current exerts a strong

influence on Zmix but little influence on Zeu. When the intru-

sion rises to the surface, there is a large increase in Zmix and,

when the intrusion sinks to an interflow, there is a lag time in

the resulting decline in Zmix as thermal stratification is re-

established.

Fig. 6d shows how the Chl. a level in XXB changes in

response to water level daily fluctuation (WLDF) in TGR during

late August. The water level of TGR increased each day from

August 22e27 and the WLDF peaked at 3.8 m/d on August 25.

The WLDF then decreased, becoming negative on August 28.

The large increase in WLDF produced a rapid decrease in Zeu/

Zmix and had a dramatic impact on the phytoplankton

community; as WLDF peaked, Chl. a levels crashed. The rapid

decline of Chl. a is attributed to disruption of the normal

temperature stratification caused by the uplift of plunging

depth (Fig. 6a, b, c).

Fig. 4 e Correlation of Zeu with turbidity and Chl. a.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 0 2125

3.3. Effects of vertical mixing on seasonal variation ofChl. a and phytoplankton succession

Fig. 7a shows the seasonal variation of Chl. a from April to

August, indicating peaks in May and August and low levels

from November to the following February (0.15e1.88 mg/m3).

Values of Zeu/Zmix and Chl. a were both low fromNovember to

Fig. 5 e Temperature stratification, flow velocity, and vertical st

indicate flow into the bay and the negative values indicate flow

February and intense phytoplankton blooms were associated

with rapid increases of Zeu/Zmix during spring and summer.

Fig. 7b displays the seasonal succession of phytoplankton

and the succession was associated with changes in the

vertical mixing regime. Diatoms and green algae predomi-

nated in spring and cyanobacteria in summer with a brief

dinoflagellate bloom in June produced by a sudden increase of

ability in XXB. For the flow velocity profiles, positive values

out of the bay.

Fig. 6 e Influence of density currents on temperature stratification and Zmix and influence of water level daily fluctuation

(WLDF) in TGR on Chl. a in XXB. a. White arrows indicate flow into XXB and black arrows indicate flow out of XXB; b.

variations of water temperature profile in response to changes of plunging depth of density current; c. Hupper indicates upper

interface of density current, Hlower indicates the lower; d. positive value of WLDF indicates rise of water level and negative

values a fall of water level.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 02126

0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8

0

10

20

30

40

50

60

Zeu

/Zm

ix

Chl

. a (

mg/

m3 )

Chl.a Zeu/Zmix

a

0%

20%

40%

60%

80%

100%

Time (YY/MM/DD)

Dinoflagellate Diatom Green algae Cyanobacteria The othersb

Fig. 7 e Influence of Zeu/Zmix on Chl. a and seasonal

succession in phytoplankton composition. a. Seasonal

variations of Chl. a and Zeu/Zmix; b. seasonal succession of

phytoplankton composition.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 0 2127

Zmix. Variations of phytoplankton biomass and succession

were both associated with seasonal deepening of Zmix and

a predictive model for phytoplankton blooms can be

developed.

4. Discussion

Light extinction by water varies with wavelength. Shorter

wavelengths are absorbed near the surface and longer wave-

lengths penetrate to deeper water (Wetzel, 2001). The atten-

uation of light by lake water results from two factors:

scattering by suspended particles and absorption by phyto-

plankton. Scattering only changes direction of light propaga-

tion so the attenuation of light underwater is primarily due to

absorption by phytoplankton (Scheffer, 2004). The results in

XXB confirm previous studies; in turbid waters, the under-

water light attenuation coefficient was found to be closely

correlated with turbidity (Oliver et al., 2000). However, it

should be noticed (Fig. 3) that the turbidity of XXB water was

generally less than 10 NTU with a low average (5.54 NTU) and

the average Chl. a was relatively high (26.55 mg/m3). This

means phytoplankton growth is free from light limitation

caused by high turbidity in XXB.

A surface layer of water is created by seasonal temperature

stratification and it is thenmixed by wind that produces shear

forces on the surface. However, wind-induced mixing of the

surface layer increases Zmix and the increase depends onwind

speed and fetch (Tucker and Green, 1977). During this inves-

tigation the maximum wind speed was 7.2 m/s, the mean

wind speed was 2.5 m/s, and seasonal variation was not

significant. The influence of winds on Zmix of XXB is small

because XXB is narrow and winding with fetches always less

than 2 km. N2 reflects the change of density gradient along

vertical and Ri represents the relative buoyancy and inertial

force often used as measures of vertical stability and

turbulence of water (Holford and Linden, 1999). Vertical

turbulence is restrained when Ri > 0 (Omstedt and Murthy,

1994). When Ri > 1, vertical turbulence is completely

restrained and flow is laminar (Miles, 1986; Stull, 1988) and

when Ri < 0.25, laminar flow changes to turbulent flow

(Stewart, 2008). The results displayed in Fig. 5 indicate that

seasonal development of temperature stratification is the

main cause for the variations in vertical stability and turbu-

lence and intrusions can modify vertical structure of normal

stratification.

Thermal stratification and the accompanying decrease in

Zmix are known to cause the seasonal onset of phytoplankton

blooms (Sverdrup, 1953) and the seasonal succession of

phytoplankton (Lindenschmidt and Chorus, 1998). The find-

ings in XXB support the conclusions of previous studies.

Vertical mixing is suppressed when the water column is

strongly stratified and nutrients are depleted in the euphotic

zone (Sommer, 1985). When Zmix is larger than Zeu, algae are

carried out ofZeu bymixing and light limitations occur. Intense

phytoplankton blooms are frequently observed when Zeu/

Zmix¼ 1 if nutrients arenot limiting.A critical value forZeu/Zmix

was defined in the development of critical depth theory. Light

limitation will occur if Zeu/Zmix is lower than the critical value.

By integrating the equation of daily photosynthesis and

respiration rate of phytoplankton along mixing depth, Talling

(Talling, 1957) obtained the value of 0.30. A range of 0.20e0.35

for critical valuewas reportedbyOliver (Oliver et al., 2000), that

accounts for seasonal variation in length of day and sun angle.

There is significant difference of duration of daylight between

winter and summer becauseXXB is located in subtropical zone

and the critical value range of 0.20e0.35 is appropriate for XXB

(0.35 from late fall to early spring and 0.20 from late spring to

early fall). The corresponding values for Chl. a and Zeu/Zmix

displayed in Fig. 7a confirm the validity of using this range for

the critical value in XXB. The results support Sverdrup’s Crit-

ical Depth Theory (Sverdrup, 1953). Controlled releases from

TGR could potentially be used to disrupt cyanobacteria blooms

because cyanobacteria are particularly sensitive to variations

in Zmix (Huisman et al., 2004).

Previous studies (Kuang et al., 2004, 2005) on phyto-

plankton composition showed that diatoms and green algae

were the dominant phytoplankton species (approximately

80% for total) before and after the initial filling in June 2003 to

water level of 135 m. The results in XXB show this trend is

persistent. Diatoms and green algae are still the most

common phytoplankton species in XXB; Cyclotella sp., Melosira

sp. and Synedra sp. are the most frequently observed diatoms

and Chlorella sp. and Chlamydomonas sp. are the most

frequently observed green algae. After the initial filling, Peri-

diniopsis sp. and Cyclotella sp. were the predominant species

and produce serious spring blooms while no single species

was observed to predominate before filling (Kuang et al., 2005).

The spring blooms predominated by Peridiniopsis sp. and

Cyclotella sp.were observed in tributary bays of TGR every year

since filling (Cai and Hu, 2006) and the diel vertical migration

of Peridiniopsis sp. has been the subject of several investiga-

tions (Xu et al., 2010c; Yang et al., 2010a). Dinoflagellates were

limited primarily to a single species (Peridiniopsis sp.) in this

study, however, they were replaced by the green algae in

spring blooms. Serious cyanobacteria blooms (Microcystis

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 02128

aeruginosa) were observed in XXB for the first time in summer

2008 (Wang et al., 2009) after the second impoundment to

water level of 156 m in October 2006. Then cyanobacteria

blooms (Anabaena sp.) were found in XXB in summer 2010 as

a part of seasonal phytoplankton succession. The phyto-

plankton communities in TGR tributary bays are evolving year

by year since the recent completion of TGR (October 26, 2010).

However, the seasonal cycle of phytoplankton and yearly

cyanobacteria blooms in summer indicate the ecosystem of

the bays is stabilizing. To better understand developing

ecosystem of this large new reservoir, phytoplankton data

should be collected continuously in the future.

Significant differences in vertical mobility among phyto-

plankton species leads to predictable phytoplankton succes-

sion with seasonal change in vertical mixing conditions. Non-

buoyant algae, such as diatoms, predominate when strong

mixing and turbulence occur (Happey, 1970). Algae with

flagella or pseudo-vacuoles predominate when water is ther-

mally stratified (Baker and Brook, 1971; Reynolds, 1972). Cya-

nobacteria regulate buoyancy via pseudo-vacuoles and

descend to absorb nutrients and ascend to absorb PAR

(Kromkamp and Walsby, 1990). When Zmix is small, diatoms

sink below the mixed layer and are replaced by cyanobacteria

(Huisman et al., 2004). This is illustrated by the fact that

Anabaena sp. predominated in August when Zeu/Zmix

increased to its peak annual value (2.66) in XXB.

A large number of studies suggested that eutrophication of

lakes and reservoirs should be addressed by controlling and

reducing inputs of phosphorus and nitrogen (Schindler et al.,

2008; Conley et al., 2009). As noted earlier, this is feasible

only in the long term in TGR. Alternative approaches, such as

the use of artificial mixing to control cyanobacteria blooms by

increasing Zmix has been proposed (Visser et al., 1996; Imteaz

and Asaeda, 2000) based on knowledge of the response of

phytoplankton growth to Zeu/Zmix. However, the techniques

used in these investigations are expensive to apply on a large

scale. The effects of water level fluctuation on water quality in

reservoirs have been widely investigated. Based on these

results, water level management has been proposed as a tool

to control phytoplankton blooms (Coops and Hosper, 2002;

Naselli-Flores and Barone, 2005; Paillisson and Marion, 2010).

Zhou (Zhou, 2008) was the first to propose that water level

fluctuations in TGR, produced by reservoir operation, could be

used to control phytoplankton blooms in its bays.Wang (Wang

et al., 2011b) found that Zmix in XXB is sensitive to large water

level fluctuations of TGR. Previous studies during post-flood

season (fall) indicate that a continuous rise of water level in

TGR can effectively reduce the degree and the area of phyto-

plankton blooms (Ji et al., 2010b; Yang et al., 2010b). However,

the hydrodynamic conditions during impoundment are

exceptional and cannot be easily extended to the other periods.

The three-day rise in water level lifted the density current

from the middle layer to the surface and the overflow thick-

ness increased due to the continuously rising water level and

the intrusion remained at the surface for two days after water

level began to fall. The response of phytoplankton biomass in

XXB towater level fluctuations of TGR indicates good potential

for controlling phytoplankton blooms in tributary bays by

appropriate control of TGR discharge. However, the large

water level fluctuation in August resulted from retention of

a flood peak and it is unrealistic to apply a WLDF of 3.8 m/d as

an operational criterion. The response relationship between

water level fluctuations in TGR and phytoplankton blooms in

its tributary bays was discussed based on available field data,

but a quantitative relationship has not yet been developed. To

implement a hydrodynamic approach for controlling phyto-

plankton blooms, an eco-hydrodynamic model coupling

temperature, silt and nutrients must be developed and the

existing dataset must be extended in order to develop the

quantitative response of temperature stratification (and Zmix)

to water level fluctuation. Management rules for controlling

phytoplankton blooms in tributaries through water level

management in TGR (minimumWLDF, duration of water level

rise, etc.) can then be integrated into an optimal operational

strategy that includes navigation, power generation, flood

control and water supply.

5. Conclusions

Results from one-year intensive field monitoring show that

thermal stratification is themain cause of seasonal variation of

Zmix. However, plunging depth of the density current can cause

transitory changes of Zmix. Phytoplankton growth and succes-

sion in XXB are driven primarily by variation of Zmix. Short-

term impoundment of TGR can lift the plunging depth of

density currents and interrupt temperature stratification in

XXB. This temporarily increases Zmix and causes a dramatic

decrease in the Chl. a level. Thus, water level fluctuations in

TGR provide a possible means for controlling phytoplankton

blooms in its tributary bays, particularly the cyanobacteria

blooms that threaten regional drinking water supplies.

Management rules for controlling phytoplankton blooms in

tributary bays using water level management in TGR would

necessarily be integrated into anoperational strategy including

navigation, power generation, flood control and water supply.

Acknowledgements

The authors greatly appreciate the financial support

provided by National 11th Five-Year Science & Technology

Research Plan of China (No. 2008BAB29B09), National Critical

Projects in Control and Management of Water Pollution

(No. 2008ZX07104-004) and NSFC (No. 51009080, No. 51009081,

No. 51179095 &No. 50925932).We thank Yang Zhengjian and Ji

Daobin for their good suggestions to this study. We also thank

all the members of Engineering Research Center of Eco-

Environment in TGR Region, Ministry of Education, CTGU for

participating in the field monitoring.

r e f e r e n c e s

Baker, A.L., Brook, A.J., 1971. Optical density profiles as an aid tothe study of microstratified phytoplankton populations inlakes. Archiv Fur Hydrobiologie 69 (2), 214e233.

Berger, S.A., Diehl, S., Stibor, H., Trommer, G., Ruhenstroth, M.,Wild, A., Weigert, A., Jager, C.G., Striebel, M., 2007. Water

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 0 2129

temperature and mixing depth affect timing and magnitude ofevents during spring succession of the plankton. Oecologia150 (4), 643e654.

Cai, Q.H., Hu, Z.Y., 2006. Studies on eutrophication problem andcontrol strategy in the Three Gorges Reservoir. ActaHydrobiologica Sinica 30 (1), 7e11 (in Chinese).

Cao, C.J., Zheng, B.H., Zhang, J.L., Huang, M.S., Chen, Z.L., 2009.Systematic investigation into winter and spring algal bloomsin Daning River of three Gorges reservoir. EnvironmentalScience 30 (12), 3471e3480 (in Chinese).

Conley, D.J., Paerl, H.W., Howarth, R.W., Boesch, D.F.,Seitzinger, S.P., Havens, K.E., Lancelot, C., Likens, G.E., 2009.Controlling eutrophication: nitrogen and phosphorus. Science(Washington) 323 (5917), 1014e1015.

Coops, H., Hosper, S.H., 2002. Water-level management as a toolfor the restoration of shallow lakes in the netherlands. LakeReservoir Management 18 (4), 293e298.

Cushing, D.H., 1989. A difference in structure between ecosystemsin strongly stratified waters and in those that are only weaklystratified. Journal of Plankton Research 11 (1), 1e13.

Farmer, D.M., Carmack, E., 1981. Wind mixing and restratificationin a lake near the temperature of maximum density. Journal ofPhysical Oceanography 11 (11), 1516e1533.

Fonseca, B.M., Bicudo, C.E.M., 2008. Phytoplankton seasonalvariation in a shallow stratified eutrophic reservoir (Gar asPond, Brazil). Hydrobiologia 600 (1), 267e282.

Garcıa, C.M., Oberg, K., Garcıa, M.H., 2007. ADCPmeasurements ofgravity currents in the Chicago River, Illinois. Journal ofHydraulic Engineering 133 (12), 1356e1366.

Happey, C.M., 1970. Some physico-chemical investigations ofstratification inAbbot’s Pool, Somerset: thedistribution of somedissolved substances. The Journal of Ecology 58 (3), 621e634.

Holford, J.M., Linden, P., 1999. Turbulentmixing in a stratified fluid.Dynamics of Atmospheres and Oceans 30 (2e4), 173e198.

Huang, Z.L., Li, Y.L., 2006. Water Quality Prediction and WaterEnvironmental Carrying Capacity Calculation for Three GorgesReservoir. China Waterpower Press, Beijing (in Chinese).

Huisman, J., Sharples, J., Stroom, J.M., Visser, P.M.,Kardinaal, W.E.A., Verspagen, J.M.H., Sommeijer, B., 2004.Changes in turbulent mixing shift competition for lightbetween phytoplankton species. Ecology 85 (11), 2960e2970.

Imteaz, M.A., Asaeda, T., 2000. Artificial mixing of lake water bybubble plume and effects of bubbling operations on algalbloom. Water Research 34 (6), 1919e1929.

Jensen, J.P., Jeppesen, E., Olrik, K., Kristensen, P., 1994. Impact ofnutrients and physical factors on the shift fromCyanobacterial to Chlorophyte dominance in shallow Danishlakes. Canadian Journal of Fisheries and Aquatic Sciences 51(8), 1692e1699.

Ji, D.B., Liu, D.F., Yang, Z.J., Xiao, S.B., 2010a. Hydrodynamiccharacteristics of Xiangxi Bay in Three Gorges reservoir.Science China: Physica, Mechanica Astronomica 40 (1),101e112 (in Chinese).

Ji, D.B., Liu, D.F., Yang, Z.J., Yu, W., 2010b. Adverse slope densityflow and its ecological effect on the algae bloom in Xiangxi Bayof Three Gorges Reservoir during the reservoir impounding atthe end of flood season. Shuili Xuebao (Journal of HydraulicEngineering) 41 (6), 691e696 (in Chinese).

Ji, X.P., Liu, D.F., Huang, Y.L., Ji, D.B., Yi, Z.Q., 2010. Nutrientdynamics of Xiangxi Bay and converse impact of the mainstream during drainage period of the Three Gorges Reservoir.Chinese Journal of Environmental Engineering 4 (12),2687e2693 (in Chinese).

Kirk, J.T.O., 1994. Light and Photosynthesis in AquaticEcosystems. Cambridge University Press.

Kling, G.W., 1988. Comparative transparency, depth of mixing,and stability of stratification in lakes of Cameroon, WestAfrica. Limnology Oceanography 33 (1), 27e40.

Kromkamp, J., Walsby, A.E., 1990. A computer model of buoyancyand vertical migration in cyanobacteria. Journal of PlanktonResearch 12 (1), 161e183.

Kuang, Q.J., Hu, Z.Y., Zhou, G.J., Ye, L., Cai, Q.H., 2004.Investigation on phytoplankton in Xiangxi River watershedand the evaluation of its water quality. Journal of WuhanBotanical Research 22 (6), 507e513 (in Chinese).

Kuang, Q.J., Bi, Y.H., Zhou, G.J., Cai, Q.H., Hu, Z.Y., 2005. Study onthe phytoplankton in the Three Gorges Reservoir before andafter sluice and the protection of water quality. ActaHydrobiologica Sinica 29 (4), 353e358 (in Chinese).

Lawrence, I., Bormans, M., Oliver, R., Ransom, G., Sherman, B.,Ford, P., Wasson, B., 2000. Physical and nutrient factorscontrolling algal succession and biomass in BurrinjuckReservoir. Technical Report, National EutrophicationManagement Program, Land Water Resources Australia.

Li, J.X., Liao, W.G., 2003. Study of the primary evoked factor ofeutrophication in the Three Gorges Reservoir. Science andTechnology Review 2, 49e52 (in Chinese).

Lindenschmidt, K.E., Chorus, I., 1998. The effect of water columnmixing on phytoplankton succession, diversity and similarity.Journal of Plankton Research 20 (10), 1927e1951.

Luo, Z.X., Zhu, B., Zheng, B.H., Zhang, Y., 2007. Nitrogen andphosphorus loadings in branch backwater reaches and thereverse effects in the main stream in Three GorgesReservoir. China Environmental Science 27 (2), 208e212(in Chinese).

Ma, J., Liu, D.F., Ji, D.B., Yang, Z.J., Yi, Z.Q., 2011. Study on flowmeasurement methods under low flow velocity conditions oftributaries in Three Gorges Reservoir. Journal of Yangtze RiverScientific Research Institute 28 (6), 30e34 (in Chinese).

Miles, J., 1986. Richardson’s criterion for the stability of stratifiedshear flow. Physics Fluids, doi:10.1063/1.865812.

Ministry of Environmental Protection of China, 2011. Bulletin onthe Ecological and Environmental Monitoring Results of theThree Gorges Project (1997e2010). The bulletin was publishedin May each year since 1997 by Ministry of EnvironmentalProtection of China, Beijing. Available at: http://english.mep.gov.cn/down_load/Documents/.

Naselli-Flores, L., Barone, R., 2005. Water-level fluctuations inMediterranean reservoirs: setting a dewatering threshold asa management tool to improve water quality. Hydrobiologia548 (1), 85e99.

Oliver, R.L., Hart, B.T., Olley, J., Grace, M., Rees, C., Caitcheon, G.,2000. The Darling River: Algal Growth and the Cycling andSources of Nutrients. Murray-Darling Basin Commission.Project M386 1999.

Omstedt, A., Murthy, C.R., 1994. On currents and vertical mixingin Lake Ontario during summer stratification. NordicHydrology 25 (3), 213e232.

Paillisson, J.M., Marion, L., 2010. Water level fluctuations formanaging excessive plant biomass in shallow lakes. EcologicalEngineering 37 (2), 241e247.

Reynolds, C.S., 1972. Growth, gas vacuolation and buoyancy ina natural population of a planktonic blue-green alga.Freshwater Biology. 2 (2), 87e106.

Scheffer, M., Rinaldi, S., Gragnani, A., Mur, L.R., van Nes, E.H.,1997. On the dominance of filamentous cyanobacteria inshallow, turbid lakes. Ecology 78 (1), 272e282.

Scheffer, M., 2004. Ecology of Shallow Lakes. Springer.Schindler, D.W., Hecky, R.E., Findlay, D.L., Stainton, M.P.,

Parker, B.R., Paterson, M.J., Beaty, K.G., Lyng, M.,Kasian, S.E.M., 2008. Eutrophication of lakes cannot becontrolled by reducing nitrogen input: results of a 37-yearwhole-ecosystem experiment. Proceedings of the NationalAcademy of Sciences 105 (32), 11254e11258.

Sommer, U., 1985. Seasonal succession of phytoplankton in LakeConstance. Bioscience 35 (6), 351e357.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 1 2 1e2 1 3 02130

Stewart, R.H., 2008. Introduction to Physical Oceanography. TexasA & M University.

Stull, R.B., 1988. An Introduction to Boundary Layer Meteorology.Springer.

Sverdrup, H.U., 1953. On conditions for the vernal blooming ofphytoplankton. ICES Journal of Marine Science 18 (3), 287e295.

Talling, J., 1957. The phytoplankton population as a compoundphotosynthetic system. New Phytologist 56 (2), 133e149.

Tucker, W.A., Green, A.W., 1977. A time-dependent model of thelake-averaged, vertical temperature distribution of lakes.Limnology Oceanography 22 (4), 687e699.

Visser, P., Ibelings, B., Van Der Veer, B., Koedood, J., Mur, R., 1996.Artificial mixing prevents nuisance blooms of thecyanobacterium Microcystis in Lake Nieuwe Meer, theNetherlands. Freshwater Biology 36 (2), 435e450.

Wang, L., Cai, Q.H., Zhang, M., Tan, L., Xu, Y.Y., Kong, L.H., 2009.Spatiotemporal dynamics and related affecting factors ofsummer algal blooms in Xiangxi Bay of Three GorgesReservoir. Chinese Journal of Applied Ecology 20 (8),1940e1946 (in Chinese).

Wang, L., Cai, Q.H., Zhang, M., Tan, L., Kong, L.H., 2011a.Longitudinal patterns of phytoplankton distribution ina tributary bay under reservoir operation. QuaternaryInternational 244 (2), 280e288.

Wang, L., Cai, Q.H., Xu, Y.Y., Kong, L.H., Tan, L., Zhang, M., 2011b.Weekly dynamics of phytoplankton functional groups underhigh water level fluctuations in a subtropical reservoir-bay.Aquatic Ecology 45 (2), 197e212.

Wang, X.F., Wei, F.S., Qi, W.Q., 2002. Water and WastewaterMonitoring and Analysis Methods, fourth ed. ChinaEnvironmental Science Press, Beijing (in Chinese).

Wetzel, R.G., 2001. Limnology: Lake and River Ecosystems, thirded. Academic Press, San Diego, CA.

Xu, Y.Y., Cai, Q.H., Han, X.Q., Shao, M.L., Liu, R.Q., 2010a. Factorsregulating trophic status in a large subtropical reservoir,

China. Environmental Monitoring and Assessment 169 (1e4),237e248.

Xu, Y.Y., Cai, Q.H., Ye, L., Shao, M.L., 2010b. Asynchrony of springphytoplankton response to temperature driver withina spatial heterogeneity bay of Three Gorges Reservoir, China.Limnologica Ecology and Management of Inland Waters 41 (3),174e180.

Xu, Y.Y., Cai, Q.H., Wang, L., Kong, L.H., Li, D.F., 2010c. Dielvertical migration of Peridiniopsis niei, Liu et al., a new speciesof dinoflagellates in an eutrophic bay of Three-GorgeReservoir, China. Aquatic Ecology 44 (2), 387e395.

Yang, Z.J., Liu, D.F., Yi, Z.Q., Ma, J., Yang, X., Ji, D.B., 2010a. Diurnalvertical migration of Peridiniopsis sp. in Xiangxi Bay of ThreeGorges reservoir. Environmental Science Research 23 (1),26e32 (in Chinese).

Yang, Z.J., Liu, D.F., Ji, D.B., Xiao, S.B., 2010b. Influence of theimpounding process of the Three Gorges Reservoir up to waterlevel 172.5 m on water eutrophication in the Xiangxi Bay.Science China Technological Sciences 53 (4), 1e12.

Yi, Z.Q., Liu, D.F., Yang, Z.J., Ma, J., Ji, D.B., 2009. Watertemperature structure and impact of which on the bloom inspring in Xianxi Bay of Three Gorges Reservoir. Journal ofHydroecology 2 (5), 6e11 (in Chinese).

Zheng, B.H., Wang, L.J., Gong, B., 2009. Load of non-point sourcepollutants from upstream rivers into Three Gorges Reservoir.Environmental Science Research 22 (2), 125e131 (in Chinese).

Zhou, G.J., Zhao, X.M., Bi, Y.H., Liang, Y.B., Hu, J.L., Yang, M.,Mei, Y., Zhu, K.X., Zhang, L., Hu, Z.Y., 2011. Phytoplanktonvariation and its relationship with the environment in XiangxiBay in spring after damming of the Three Gorges, China.Environmental Monitoring and Assessment 176 (1e4),125e141.

Zhou, J.J., 2008. Improvement of eco-environmental conditions ofThree Gorges Reservoir by optimal operations. Science andTechnology Review 26 (7), 64e71 (in Chinese).