ADVANCES IN ATMOSPHERIC SCIENCES, VOL. 29, NO.

6, 2012, 1360–1373

Simulation and Exploration of the Mechanisms Underlying

the Spatiotemporal Distribution of Surface
Mixed Layer Depth in a Large Shallow Lake
ZHAO Qiaohua∗1 (赵巧华), SUN Jihua2 (孙绩华), and ZHU Guangwei3 (朱广伟)
1
College of Remote Sensing, Nanjing University of Information Science and Technology,

Key Laboratory of Meteorological Disaster of Ministry of Education, Nanjing 210044

2
Yunnan Institute of Metrological Science, Kunming 650034
3
State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography and Limnology,

Chinese Academy of Sciences, Nanjing 210008

(Received 12 December 2011; revised 30 January 2012)

ABSTRACT
The aquatic eco-environment is significantly affected by temporal and spatial variation of the mixed
layer depth (MLD) in large shallow lakes. In the present study, we simulated the three-dimensional water
temperature of Taihu Lake with an unstructured grid with a finite-volume coastal ocean model (FVCOM)
using wind speed, wind direction, short-wave radiation and other meteorological data measured during
13–18 August 2008. The simulated results were consistent with the measurements. The temporal and
spatial distribution of the MLD and the possible relevant mechanisms were analyzed on the basis of the
water temperature profile data of Taihu Lake. The results indicated that diurnal stratification might be
established through the combined effect of the hydrodynamic conditions induced by wind and the heat
exchange between air and water. Compared with the net heat flux, the changes of the MLD were delayed
approximately two hours. Furthermore, there were significant spatial differences of the MLD in Taihu Lake
due to the combined impact of thermal and hydrodynamic forces. Briefly, diurnal stratification formed
relatively easily in Gonghu Bay, Zhushan Bay, Xukou Bay and East Taihu Bay, and the surface mixed layer
was thin. The center of the lake region had the deepest surface mixed layer due to the strong mixing process.
In addition, Meiliang Bay showed a medium depth of the surface mixed layer. Our analysis indicated that
the spatial difference in the hydrodynamic action was probably the major cause for the spatial variation of
the MLD in Taihu Lake.

Key words: mixed layer depth, temporal and spatial distribution, Taihu Lake, thermal stratification
Citation: Zhao, Q. H., J. H. Sun, and G. W. Zhu, 2012: Simulation and exploration of the mechanisms
underlying the spatiotemporal distribution of surface mixed layer depth in a large shallow Lake. Adv. Atmos.
Sci., 29(6), 1360–1373, doi: 10.1007/s00376-012-1262-1.

1. Introduction demand of algal photosynthesis. Hence, the dynamic

action of the surface mixed layer (ML), which is de-
The abundance, composition and vertical distri- fined as a quasi-homogeneous region where there is
bution of phytoplankton in the euphotic zone can be little variation in temperature or density with depth
affected by various physical, chemical, and biological (Kara et al., 2000), is particularly important for phy-
processes (Doyon et al., 2000). Phytoplankton agglom- toplankton, even the whole aquatic ecosystem. The
erate in shallow water due to the vertical attenuation establishment and destruction of stratification, as well
of photosynthetically active radiation (PAR) and the as changes in its depth, can affect the consumption and

∗ Corresponding author: ZHAO Qiaohua, [email protected]

© China National Committee for International Association of Meteorology and Atmospheric Sciences (IAMAS), Institute of Atmospheric
Physics (IAP) and Science Press and Springer-Verlag Berlin Heidelberg 2012
NO. 6 ZHAO ET AL. 1361

supply of nutrients (Anis and Singhal, 2006) and, sub- surface ML in Taihu Lake.
sequently, algal growth. Turbulence in the ML can also
control the movement and distribution of phytoplank- 2. Materials and methodology
ton, which in turn affects their photosynthesis rate, as
well as the frequency, duration, and intensity of algal 2.1 Materials
blooms (Obata et al., 2009). Therefore, the mixed
The observation site, a station in the Global Lake
layer depth (MLD) is critical to species succession,
Ecological Observatory Network (GLEON), is located
distribution, and the primary productivity of phyto-
at 31.419◦ N and 120.213◦ E in the observation plot of
plankton in the water column (Raybov et al., 2010);
the Taihu Ecosystem Research Station on the south-
plus, it plays an important role in aquatic ecosystem
east side of Meiliang Bay, Taihu Lake. The observa-
dynamics as well as in climate change (Nagai et al.,
tion site is 150 m away from the shore and is linked
2005; Obata et al., 2009).
to the shore through its observation trestle. The wa-
Given the important role that the surface ML plays
ter depth was 2.33 m during the observation period.
in the aquatic ecosystem, the characteristics of the ML
Water temperature, at depths of 0.1 m, 0.63 m, 1.1
of oceans and deep lakes have been studied extensively
m, 1.63 m and 2.13 m, were monitored using a TS110
(Hadfield and Sharples, 1996; Lehman, 2002; Rao and
water temperature chain (NexSense, USA).The mete-
Schwab, 2007; Kara et al., 2009). However, the effect
orological parameters were monitored using a Vaisala
of variation of the MLD in large shallow lakes with
WXT520 automatic weather system deployed above
strong surface mixing action has yet to be studied.
the water surface. The measured parameters included
Taihu Lake, for example, is often assumed to be verti-
wind speed, wind direction, air temperature, relative
cally homogenous because it is a typical shallow large
humidity, barometric pressure and downwelling solar
lake with an area of 2,338 km2 and an average depth
short-wave radiation.
of approximately 2 m (Qin et al., 2004a).
The data were recorded every 10 minutes from 0500
Unlike deep lakes, seasonal stratification is rare in
LST 13 August 2008 to 2300 LST 18 August 2008.
large shallow like Taihu Lake due to intensive mix-
In this study, the data acquired at the integral time
ing. However, Zhang et al. (2008a, 2008b) sug-
points were used. The cloud cover and evaporation
gested that stratification may be established and main-
level were monitored by the nearby Yixing Meteorolog-
tained during daytime, and then destroyed at night;
ical Station, which is 12 km from the northeast shore
namely, diurnal stratification. Thus, large shallow
of Taihu Lake.
lakes can also form a thin ML under appropriate
A lake-wide survey including 29 stations in Taihu
conditions of wind speed/direction, water/atmosphere
Lake was conducted during 15–18 August 2008. At
heat exchange, and surrounding terrain (Lövstedt and
each station, the water samples were collected using
Bengtsson, 2008). The rapid establishment and de-
Niskin bottles. These samples were immediately pre-
struction of stratification can lead to frequent replen-
served at 2◦ C –4◦ C and transported to the lab for
ishment of nutrients in the water column (Qin et al.,
analysis within the same day. In the lab, the samples
2004b; Pang et al., 2007; Lövstedt and Bengtsson,
were filtered with GF/C filters (after the filters had
2008). The mixing process can also promote algal
been dried at 105◦ C for 4 h and preweighed), dried
growth under suitable conditions of MLD and pho-
(105◦ C for 4 h) and weighed to determine the total
tosynthetic active radiation.
suspended solid (TSS).
In recent years, the effects of the ML on the eco-
environment may be more significant with the rapid in- 2.2 Methodology
crease of air temperature, decreasing precipitation and
increasing sunshine duration in the Taihu Lake region 2.2.1 Water temperature simulation
(Shang et al., 2010), but previous studies have largely A finite-volume coastal ocean circulation model
focused on analyses of data measured at a single ob- (FVCOM), which is a 3-D unstructured–grid, free-
servation station (Zhang et al., 2008a, b). Moreover, surface, primitive equation model developed by the
high-frequency changes in wind direction and heat ex- Woods Hole Oceanographic Institution, Woods Hole,
change have not been taken into account in previous MA, USA (Chen, 2003), was used to explore the reg-
studies on hydrodynamic mechanisms for Taihu Lake ularities and mechanisms of the temporal and spatial
(Pang et al., 2008). In this context, the present work variation of the surface ML driven by the meteoro-
aims to provide evidence demonstrating the impact of logical field in Taihu Lake. The model was set up
the physical environment on aquatic ecosystems and with a triangle mesh to better fit the boundary, and
the mechanisms underlying algal blooms with an em- its numerical calculation was configured to use the
phasis on the spatial and temporal distribution of the finite-volume and finite-difference methods, which pro-
1362 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

vide a better representation of the mass, momentum Hlat = ρa Le U κlat (qa − qsat ) , (4)
and heat conservation in coastal and estuarine regions
with complex geometry (Chen, 2003). The model was where ρa represents the atmospheric density, ζpa is the
closed physically and mathematically (Chen, 2003), atmospheric heat capacity, κsat is the transforming co-
using the Mellor and Yamada level 2.5 turbulence clo- efficient of sensible heat, κlat is the transforming coef-
sure scheme for vertical mixing and the Smagorin- ficient of latent heat, qa is the specific humidity, and
sky turbulence closure scheme for horizontal mixing qsat is the saturation specific humidity.
(Chen, 2003). Horizontal and vertical turbulence dif-
fusion coefficients were introduced as functions instead 2.2.3 Determination of the mixed layer depth
of constants to overcome the deficiencies in the previ-
According to the optimal definition of the ML de-
ous hydrodynamic model of Taihu Lake (Qin and Fan,
veloped by Kara et al. (2000), the MLD of Taihu Lake
2002). In addition, the bottom boundary condition for
was determined as follows:
temperature in most finite-difference models is simpli-
(1) The temperature at 10 cm below the water sur-
fied to ∂T /∂z = 0, which is generally sound for much
face was considered to be the initial reference tempera-
of the continental shelf in coastal oceans where the
ture (Tref ) to avoid the variation caused by evaporation
bottom topography is smooth with a small gradient.
at the water surface;
In this paper, the δ-coordinate was used in the verti-
(2) To determine the regions with similar tempera-
cal direction and divided into 14 levels: 0.0 m, 0.166
ture, we compared the temperatures (Tn , Tn+1 ) of the
m, 0.332 m, 0.498 m, 0.664 m, 0.83 m, 0.996 m, 1.162
two adjacent water layers (hn , hn+1 ). If the difference
m, 1.328 m, 1.494 m, 1.66 m, 1.826 m, 1.992 m, and
was less than one-tenth of the standard value (0.8◦ C),
2.1528 m.
the two layers were considered to be regions with the
2.2.2 Quantitative analysis of the net heat flux re- same temperature; thus, we continued to search deeper
ceived by the water of Taihu Lake layers. If the difference was larger than 0.08◦ C, the hn
The net heat flux indicates the heat received by the was considered as MLD.
water during the exchange process between water and
2.2.4 Determination of the Richardson Number
atmosphere. Its components include net short-wave
radiation flux, long-wave radiation flux, sensible heat In Taihu Lake, the inflows and outflows are small;
flux and latent heat flux, and the components were an- therefore, it can be regarded as a closed lake, and the
alyzed quantitatively as described previously by Mag- mixing effect caused by outflows and withdrawals may
giore et al. (1998), Churchill and Kerfoot (2007) and be omitted (Pang and Pu, 1996).
Kim and Cho (2011). In addition to the abovementioned processes, the
Determination of the net short-wave radiation flux following also affect mixing: (1) the vertical shear of
into the lake was calculated according to: current velocity induced by wind, which leads to heat
transfer to the lower layer and contributes to the mix-
Qs = (1 − α)Qs0 (1) ing of the layers; and (2) heat exchange across the
air–water interface (Tuan et al., 2009). The second
where Qs0 represents short-wave radiation monitored process may have an opposite effect on the mixing.
using the Vaisala WXT520 automatic meteorological On the one hand, thermal stratification will be pro-
system, α is the albedo of the water surface, and Qs is duced and maintained when wind is unable to mix the
the short-wave radiation into the water (W m−2 ). heated surface water over the whole water column. On
The net long-wave radiation flux was calculated us- the other hand, the thin MLD will become deeper or
ing the following formula: be destroyed when the net heat flux directed into the
water is not large enough to cause stratification in the
HLW = εσTw4 (0.05ea −0.39)Bc −4εσTw3 (Tw −Ta ) , (2)
lake (Tuan et al., 2009). Together, these two processes
where HLW is the net long-wave radiation flux (W determine the variation of water mixing. The Richard-
m−2 ); ε is the emissivity of the water; σ is the Stephen son Number is a measure of the relative importance of
Boltzmann constant; Tw and Ta are the temperatures turbulence induced by wind and thermal effects in the
(K) of the water and atmosphere, respectively; and Bc water column. This research focused on the tempo-
is the cloud correction factor, Bc = (1 − 0.72C), where ral and spatial variation of the Richardson Number
C represents the cloud cover. to explore the spatial differences in the mechanisms
The heat influx due to sensible and latent heat was controlling the MLD. We calculated the Richardson
calculated as follows: Number as follows:
The oscillation frequency (N ) of stable stratifica-
Hsen = ρa ζpa U κsat (Tw − Ta ) , (3) tion is defined as:
NO. 6 ZHAO ET AL. 1363

s water temperature of Taihu Lake was simulated by

g ∂ρ coupling the heat exchange between the atmosphere
N= − , (5) and water with FVCOM.
ρ0 ∂z
In order to analyze the temporal and spatial vari-
where ρ0 and ρ represent the mean and actual density ation of the MLD in Taihu Lake using the three-
of water, respectively; g is gravitational acceleration; dimensional water temperature data, the result of the
and z is the depth of the water. simulation had to be calibrated using measured water
Thus, the Richardson Number (Ri ) can be ex- temperature data. Thus, the meteorological data de-
pressed as: scribed in Section 2.1 were used as the input data to
simulate the water temperature during 13–18 August
N2
Ri = µ ¶2 µ ¶ 2 , (6) 2008. Finally, we interpolated the simulation results
∂u ∂v into the observation points and the depths in order to
+
∂z ∂z validate the simulation and analyzed the changes of
the temperature vertical profile.
where u and v represent the horizontal velocity com- The variation of surface water temperature with
ponents, respectively. time in the simulation was consistent with that of the
2.2.5 Estimation of the euphotic depth measured data. There was a minimum relative error of
1.37% and a maximum relative error of 5.4% (Fig. 1),
According to the relationship between the con-
which showed that our model was suitable for the sim-
centration of TSS and the euphotic depth of PAR
ulation of surface water temperature.
[Deu (PAR)] (Zhang et al., 2006), the euphotic depths
To further validate the simulated temperature re-
of PAR in Taihu Lake were estimated before the spatial
sults, the relationship between the simulated and mea-
distribution of TSS was calculated using the method of
sured data at various depths was analyzed and dis-
inverse distance weighted interpolation. The formula
played in a scatter plot (Fig. 2). The results showed
is as follows:
that these two sets of data were distributed along a
4.605 straight line y = 0.997x, with a correlation coefficient
Deu (PAR) = . (7)
0.0626TSS + 1.60608 of 0.83 (n=695). The mean relative error was 1.3%,
and the maximum relative error was 6.3%. These re-
3. Results sults further confirmed that our model was suitable for
the simulation of the three-dimensional water temper-
3.1 Simulation and calibration of the water ature of large shallow lakes.
temperature in Taihu Lake
3.2 Error analysis
As a large shallow freshwater lake, the stratifica-
tion of Taihu Lake is mainly determined by the verti- There were two possible causes of water tempera-
cal temperature profile. Hence, the three-dimensional ture simulation errors. First, the one-hour interval be-

36
35 Temperature(simulated)
34 Temperature(measured)
33
Temperature(oC)

32
31
30
29
28
27
26
13_05:00

14_00:00

14_19:00

15_14:00

16_09:00

17_04:00

17_23:00

18_18:00

Time(hour)

Fig. 1. Temporal variation between measured and simulated water

temperature at 10 cm.
1364 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

34
y = 0.9973x
R² = 0.69

Water temperature measured(oC)

33

32

31

30

29

28
      

Water temperature simulated(oC)

Fig. 2. Correlation between measured and simulated water tempera-

ture (n=695).

1000
Time Interval of 10 Minutes
Time Interval of 1 Hour
800
Short Wave(Wm-2)

600

400

200

0
13_05:00

13_21:40

14_14:20

15_07:00

15_23:40

16_16:20

17_09:00

18_01:40

18_18:20

Time(hour)

Fig. 3. Time series of short-wave radiation flux.

tween two continuous meteorological data points, such 3.3 Temporal and spatial distribution of the
as net short-wave radiation flux, was too long to ac- MLD
curately depict the time series of the meteorological
The factors involved in water stratification and
data. For instance, at 1300 LST 15 August, the net
mixing were as follows: (1) heat exchange between the
short-wave radiation flux value was 425 W m−2 lower
water and atmosphere; and (2) the turbulence that
than the measured value (730 W m−2 ) at 1320 LST 15
resulted from the vertical shearing of the horizontal
August (Fig. 3). Second, the water albedo may affect
velocity induced by wind. In this section, we focus
the accuracy of the water temperature simulation. In
mainly on elucidating the influence of heat exchange
our model, the albedo of the water surface was set as
on water stratification and mixing. The second factor
a constant (Gill, 1982; Guarini et al., 1997). However,
will be discussed later, in section 4.2.
it can be modified by variation in the atmospheric tur-
The processes of heat exchange between the wa-
bidity and solar altitude. The atmospheric turbidity
ter and atmosphere involves short-wave and long-wave
impacts direct-beam irradiances, subsequently affect-
radiation, as well as sensible and latent heat. The
ing the water albedo, and the solar altitude exhibits
analysis of the time series of the heat flux compo-
obvious daily variations (Payne, 1972). Thus, the con-
nents (Figs. 4–6) revealed that the fluctuation ampli-
stant might cause error in the simulation of water tem-
tude of the net heat flux directed into the lake (pos-
perature.
NO. 6 ZHAO ET AL. 1365

1000
Short Wave Radia!ve Flux
800
Net Heat Flux
600

Heat Flux (Wm -2) 400

200

0
13_05:00

13_21:00

14_13:00

15_05:00

15_21:00

16_13:00

17_05:00

17_21:00

18_13:00
-200

-400

-600
Time(Hour)

Fig. 4. Time series of net heat flux and short wave radiation flux.

800
Latent Heat Flux
600 Net Heat Flux
Heat Flux (Wm -2)

400

200

0
13_05:00

13_21:00

14_13:00

15_05:00

15_21:00

16_13:00

17_05:00

17_21:00

18_13:00

-200

-400

-600 Time(Hour)

Fig. 5. Time series of net heat flux and latent heat flux.

800
Long wave Radia!ve Flux
600 Sensible Heat Flux
Net Heat Flux
Heat Flux (Wm -2)

400

200

0
13_05:00

13_21:00

14_13:00

15_05:00

15_21:00

16_13:00

17_05:00

17_21:00

18_13:00

-200

-400

-600
Time(Hour)

Fig. 6. Time series of net heat flux, sensible heat and long-wave radi-
ation flux.
1366 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

Fig. 7. Times series of water temperature at different depths.

itive) was less than the case of short-wave radiation (A) The temperature increased during daytime and
flux; the phase of net heat flux exhibited a similar decreased at night;
changing trend compared with short-wave radiation (B) The vertical temperature difference was rel-
flux (Fig. 4), which suggested that the diurnal trend atively significant during daytime and very small at
of the net heat flux was determined principally by the night. These results suggested that the water mixed
short-wave radiation flux when the net heat flux was well at night on account of the conditions being favor-
positive. The sensible heat flux, latent heat flux, and able for convection to prevail due to heat loss from the
the long-wave radiation flux mainly contributed nega- water surface (Anis and Singhal, 2006). During the
tively to the net heat flux. The variation of the nega- daytime of 13–16 August, the stratification was rela-
tive net heat flux was consistent with that of the latent tively more significant than that during 17–18 August;
heat flux (Figs. 5 and 6), suggesting that variations of (C) The time of the peak temperature in the lower
latent heat flux played a key role in the process of heat layer lagged behind that in the upper by 2–4 h;
loss from the water surface. (D) The water temperature dropped on 17 and 18
We analyzed the strength and phase of water August and, especially on the latter of these two days,
stratification and mixing on the basis of the tempo- the water temperature varied little among the depths,
ral changes of water temperature at different depths indicating that the water was well mixed according
(Fig. 7) and found that water temperature exhibited a to the optimal definition for MLD (Kara et al., 2000)
diurnal variation. The variation was as follows: and the mechanism analyses on diurnal stratification
NO. 6 ZHAO ET AL. 1367

a: 6:00 b: 10:00

c: 14:00 d: 16:00

Fig. 8. Spatial distribution of MLD on 17 August 2008 at (a) 0600 LST; (b) 1000 LST; (c) 1400
LST; and (d) 1600 LST(units: m).

(Zhang et al., 2008a). LST, and 1600 LST, 17 August, with the sunrise at
A comparison between the trends shown in Figs. 7 0526 LST (Figs. 8a–d). Combined with the net la-
and 4 demonstrated that the water stratification was tent heat flux and other processes, the water that was
mainly determined by the net heat flux. When the well-mixed at 0600 LST 17 August had started mixing
net heat flux was positive, the water temperature in- from the moment the net heat flux was negative on
creased due to receiving heat. 16 August, indicating that the ML of Taihu Lake was
The data from 18 August (Fig. 4) showed that the relatively deep (Fig. 8a). Compared with the depth of
positive net heat flux was normal, but the time span Taihu Lake (Fig. 9), the MLD was close to the bot-
of the positive net heat flux was short. This finding tom of the lake (Fig. 8a). As time passed, the net heat
suggested that the water received only a tiny amount flux gradually increased. Absorption of solar radiation
of heat from the atmosphere, which in turn lowered within a relatively thin surface layer produced thermal
the water temperature and weakened the stratification stratification due to the strong absorption by high-tur-
compared with that during 13–16 August. Further-
more, during the period between sunset on 17 August
and sunrise on 18 August, the net heat flux exhibited
a large negative value. This negative value probably
indicated a great heat loss (Figs. 4–6) and showed that
the atmosphere was slightly colder during this period.
A north wind during this period would also contribute
to this result. Together, these factors caused the wa-
ter temperature to decline markedly and resulted in a
strong mixing of the water column during the daytime.
The spatial distributions of the MLD were ana-
lyzed based on the MLD at 0600 LST, 1000 LST, 1400
1368 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

than that in Gonghu Bay. Furthermore, the MLD was

1.6 m in Zhushan Bay, which was 0.2 m shallower than
that in the central lake. At the two points in west of
Xishan Island, the MLD was 1.8 m and 1.4 m, respec-
tively, while it was approximately 1.4 m in Xukou Bay
and East Taihu Bay.
During the period when the net heat flux was pos-
itive, the temporal variations of the MLD in these re-
gions were as follows: (1) In the daytime from 13–
18 August, the MLD trend displayed a double val-
ley, which was consistent with the bimodal variation
in the net heat flux (Figs. 4–6; Figs. 10a–d); (2) From
13–16 August, the minimum value of MLD was simi-
lar among the different lake regions; (3) On 17 and 18
August, the MLD was significantly different among the
different lake regions. The smallest depth occurred in
Fig. 9. Spatial distribution of water depth. Gonghu Bay, with values of 0.4 m and 0.6 m, followed
by that in Xukou Bay and East Taihu Bay (the depths
bidity water. The thermal stratification first appeared were similar: 0.5 m and 0.7 m for the two days, respec-
in bay regions, such as Zhushan Bay, Gonghu Bay, and tively), Meiliang Bay and Zhushan Bay (the depth on
Xukou Bay. The area of the relatively thin ML grad- both days was 0.8 m), and the center of the lake and
ually became larger (Fig. 8b). With the short-wave the western part of Xishan Island (approximately 1.2
radiation flux increasing and accumulating, the MLD m).
tended to become shallower throughout the whole lake, In summary, stratification disappeared at night be-
especially in the abovementioned bay regions. The cause of hydrodynamics and the declining effect of the
maximum MLD was less than 1 m, and the depth in net heat flux. In the daytime, the establishment of
the center of the lake was greater than that in other stratification depended on the effect of the net heat
regions. From 1600 LST, the net heat flux gradually flux. The MLD in large shallow lakes like Taihu Lake
decreased; thus, the temperature difference among the shows not only an apparent diurnal variation, but also
different depths decreased, and the mixing process in- significant spatial differences. The mechanism of strat-
tensified. The overall MLD of Taihu Lake at 1600 LST ification is complex and will be discussed in detail be-
(Fig. 8d) was greater than that at 1400 LST (Fig. 8c), low.
and the spatial difference persisted. At all four time
points, the MLD was the greatest in the center of the 4. Discussion
lake, followed by that in Meiliang Bay, the group of
Zhushan Bay and Gonghu Bay, and the group of East Taihu Lake exhibits significant temporal and spa-
Taihu Bay and Xukou Bay. tial variation in its MLD. The mechanisms underlying
this variation are discussed in this section.
3.4 Temporal variation of MLD in different
lake regions 4.1 Mechanisms underlying the temporal
variation of the MLD in Taihu Lake
To further analyze the temporal changes in the
MLD, eight observation points (#1–8) were selected: In an ocean or a deep lake in winter or summer,
Meiliang Bay(#1); Gonghu Bay (#2); Zhushan Bay there is a relatively stable stratification that affects
(#3); the center of the lake (#4); the west (#5) and phytoplankton production, algal blooms and nutri-
southwest parts of Xishan Island (#6); Xukou Bay ent cycling (Sverdrup, 1953). However, the seasonal
(#7); and the East Taihu Bay (#8). Four subgraphs stratification may be a limiting factor for phytoplank-
in Fig. 10, each consisting of two points, were employed ton photosynthesis because the nutrients in the up-
to clearly depict the temporal changes of the MLD at per layer may be depleted and the stratification pre-
these points. vents the supplementation of nutrients from the lower
In the different lake regions, the water was well layer. In large shallow lakes, diurnal stratification has
mixed during the period with a negative net heat flux, a stronger impact on the ecological environment than
especially between sunset and sunrise on the following in the ocean or deep lakes. The fact that stratifica-
day. The MLD was different in different regions. In tion is established and destroyed quickly is beneficial
Meiliang Bay, it was 1.8 m, which was 0.4 m deeper to nutrient supplementation to the upper layer.
NO. 6 ZHAO ET AL. 1369

Fig. 10. Time series of MLD at different observation points.

The vertical profile of water temperature in any solar radiation, for instance, on 18 August 2008. The
body of water is determined by the combined effects of process of heat exchange may lead to diurnal stratifi-
solar short-wave radiation heating, water surface cool- cation when solar heating is large enough to overcome
ing, and wind-induced mixing. Because solar short- the mixing effect induced by wind. Alternatively, the
wave radiation reduces exponentially with an increase stratification could be destroyed or prevented from be-
in water depth and strong absorption in turbid water, ing established, as evidenced by the stratification and
heating can be concentrated in a thin layer, and the mixing phenomena monitored during 13–18 August
vertical gradient becomes greater. (Figs. 7 and 10). Second, the momentum transferred
Temperature gradients will become significant and from the atmosphere decreases along with an increase
probably lead to the establishment of stratification in water depth, so the vertical shear of the horizon-
when surface temperature increases quickly. In ad- tal movement may lead to water mixing when it over-
dition, because the quickly heated layer is thin on comes the effect resulting from the vertical tempera-
account of strong absorption and there is relatively ture gradient. Additionally, mixing could arise from
strong heat loss, the stratification would be quickly de- a combination of the vertical shear of the horizontal
stroyed during a short period when the net heat flux movement and the negative net heat flux process. As
is negative or when the effect of mixing induced by a result, the water mixing will be strengthened, which
wind is strong. It can be concluded that there may in turn can lead to the disappearance of stratification
be diurnal stratification in Taihu Lake, rather than at night, or even in daytime, making it difficult to es-
seasonal stratification, because the time span of the tablish diurnal stratification.
seasonal stratification for other lakes is comparatively Without the hydrodynamic effects, diurnal stratifi-
long (Huang et al., 1999) compared to Taihu Lake. cation can form in Taihu Lake under the heating effect
In addition, there are two processes that can de- of short-wave radiation. The effect of net heat flux on
stroy the stable stratification after the establishment such diurnal stratification can be confirmed by the re-
of the temperature gradient. First, the net heat loss lationship between net heat flux and the surface MLD
can decrease the water temperature. As a result, the at all of the previously mentioned points (Table 1).
surface water temperature is reduced to form unstable The correlation between net heat flux and surface
conditions in which the upper water is cold while the
lower water is warm, which in turn results in convec-
tive mixing and an increase in the MLD. This phe-
nomenon may be caused by a cooling process when
the net heat flux is directed out of the lake (negative)
resulting from such events as a cold air break and low
1370 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

Table 1. Relationship between the net radiation flux and result in a weakening or rapid disappearance of di-
the surface MLD. urnal stratification and an increase in the MLD. The
Richardson Numbers at different points of Taihu Lake
Sample size
were determined using the method described above
point # 139 138 137 136 (Fig. 11), and they were calculated based on the up-
1# −0.44 −0.63 −0.67 −0.59
per five layers of water where the difference in water
2# −0.55 −0.69 −0.70 −0.60 temperature was relatively significant.
3# −0.50 −0.66 −0.67 −0.58 The Richardson Number can decrease with increas-
4# −0.41 −0.58 −0.62 −0.55 ing vertical shear of horizontal velocity or decreasing
5# −0.40 −0.57 −0.62 −0.56 vertical stratification of the density. Hence, the tem-
6# −0.50 −0.63 −0.61 −0.50 poral and spatial variation of the Richardson Number
7# −049 −0.63 −0.65 −0.56 can account for the mechanically and thermodynam-
8# −0.58 −0.71 −0.70 −0.58 ically driven variation in the surface MLD in Taihu
Lake.
MLD, as well as their temporal relationship, is shown The temporal variation in the Richardson Num-
in Table 1. By comparing the correlation coefficients ber at different points showed that when net heat flux
between net heat flux and the MLD with a 0-hour was positive, the Richardson Numbers in Gonghu Bay,
delay (n=139), 1-hour delay (n=138), 2-hour delay Zhushan Bay, Xukou Bay, East Taihu Bay, Meiliang
(n=137), and 3-hour delay (n=136) respectively, we Bay, and the southwestern part of Xishan Island were
found that the MLD with a 2-hour delay exhibited the greater than those in the center of the lake, indicating
best correlation with net heat flux data for all of the that the water in these former regions was easily strat-
lake regions. These results not only verified that net ified (Fig. 11; each subgraph contains the Richardson
heat flux was the driving force for the establishment Numbers at two points). During diurnal stratifica-
of diurnal stratification, but also showed that the vari- tion, the MLD in the lake centre was relatively deeper
ation of net heat flux was prior to the change of the than those in other regions, which was consistent with
MLD. the results of Figs 8a–d. However, in the center of
the lake, the Richardson Number was small, suggest-
4.2 Mechanisms underlying the spatial varia- ing that the strong disturbance caused by the shear
tion of the MLD in Taihu Lake of the horizontal flow field tended to mix the wa-
ter. In addition, because this study was performed
The effects of thermodynamic processes on the es- in summer when there were predominantly southerly
tablishment of the thin ML only illustrated how energy and northerly winds (Fig. 12), the wind-driven effects
influenced the vertical distribution of water tempera- were relatively apparent in Meiliang Bay. Thus, the
ture; they could not explain the spatial variation in Richardson Number in Meiliang Bay was small and
the MLD mentioned previously (Table 1, Figs. 8 and
10). Here, we address this question in detail. Table 2. The mean Richardson Number during the simu-
Taihu Lake is defined as a large shallow lake based lation.
on the ratio between its area and depth, which cov- point # mean Richardson number
ers approximately 50×60 km2 , so the atmosphere over
the area is approximately uniform horizontally, and #1 0.291
the net heat flux is almost even for the whole lake. #2 1.063
#3 1.865
However, the flow field of water exhibited significant
#4 0.191
spatial heterogeneity because it depended on the fetch #5 0.185
length, duration and water depth. Hence, the spatial #6 0.244
difference in the three-dimensional flow field may be #7 0.651
the only cause that led to the spatial variation of the #8 0.523
MLD.
The transfer of momentum from the air to the wa-
ter causes the vertical shear of the horizontal current in
the lake. Only when such an effect is greater than the
buoyancy resulting from stable stratification will the
turbulent mixing of surface water appear. Combined
with the relatively weak temperature stratification in
Taihu Lake, the turbulent mixing of surface water can
NO. 6 ZHAO ET AL. 1371

Fig. 11. Time series of the Richardson number in different regions of Lake Taihu.

Fig. 12. Time series of wind speed and wind direction.

Table 3. The relationship between the Richardson Num- in Meiliang Bay tended to mix uniformly and showed a
ber and the MLD.
small Richardson Number, which was consistent with
Correlation coefficient our previous analysis (Fig. 11 and Table 2).
point # (n=139) We further confirmed that the Richardson Number
could represent changes in turbulence through a cor-
#1 −0.53
relation analysis between the MLD and Richardson
#2 −0.28
#3 −0.41 Number at the different points mentioned previously
#4 −0.57 (Table 3).
#5 −0.53 4.3 Analysis of the effect of the mixing pro-
#6 −0.40
cess on algal production
#7 −0.33
#8 −0.51 Algal development is thought to be strongly influ-
enced by surface irradiance, vertical mixing and tem-
perature (Tirok and Gaedke, 2007). The effect of the
there was relatively strong mixing action. On 18 Au-
ML is more prevalent than the effect of temperature
gust, due to the short duration of the positive net heat
because the ML governs both the light climate and
flux and subsequently tiny energy received, the water
temperature in the ML (Berger et al., 2007). In ad-
1372 SIMULATION AND EXPLORATION OF SURFACE MIXED LAYER DEPTH VOL. 29

dition to temperature, the turbulence of the ML con- depth in Zhushan Bay and Meiliang Bay met the fa-
trols the vertical movement of algae, which in turn vorable conditions for algal production, but it needs
affects algal production and destruction by controlling to be further validated by future observations.
the light climate. Therefore, the ratio between the eu-
photic depth and the MLD is one of the major factors 5. Conclusion
affecting the growth rate of algae (Sverdrup, 1953).
In summary, owing to heat exchange between
The favorable physical conditions for algal production
the water and atmosphere through short-wave radia-
tion, long-wave radiation and sensible and latent heat,
Taihu Lake is able to establish diurnal stratification
and maintain a thin surface ML during periods when
net heat flux is large. The variation of the MLD was
found to be well correlated to net heat flux with a two-
hour delay. Under the effects of hydrodynamic and
thermodynamic forces, the MLD exhibited apparent
spatial variation. Specifically, the thin ML was es-
tablished and maintained more easily in Gonghu Bay,
Zhushan Bay and East Taihu Bay compared to Meil-
iang Bay. However, compared with the center of the
lake it formed more easily in Meiliang Bay. It was
relatively difficult for diurnal stratification to be es-
tablished in the center of the lake because the me-
chanical effects are very strong in this region, which
in turn results in a strong mixing action and a greater
Fig. 13. Location of sampling stations. MLD. The spatial differences in hydrodynamic action
are probably the major cause of the spatial variation in
the MLD in Taihu Lake. The ratio between the MLD
 
  

and euphotic depth in Meiliang Bay and Zhushan Bay

  
is favorable for algal production.

Acknowledgements. The authors wish to thank

  


 

  

Chris McBride, Linlin Zhao and Jingchen Xue for their


maintenance of the Global Lake Ecological Observatory
  

Network (GLEON) monitoring system in Taihu and data

processing. The study was supported by the National Nat-
  

  

ural Science Foundation of China (Grant Nos. 41071070,





  
41165001, and 41171368), the Special Scientific Research
Project of the China Commonwealth Trade (meteorology)
 

(GYHY201006054).

 


                         

REFERENCES

Fig. 14. Spatial distribution of euphotic depth on 15 Au- Anis, A., and G. Singhal, 2006: Mixing in the surface
gust 2008 (Units: m). boundary layer of a tropical freshwater reservoir. J.
Mar. Syst., 63, 225–243.
is a deep euphotic depth and a thin ML regardless of Berger, S. A., and Coauthors, 2007: Water temperature
and mixing depth affect timing and magnitude of
the effect of nutrients, temperature, and so on.
events during spring succession of the plankton. Oe-
Based on the concentrations of TSS at 29 sampling cologia, 150, 643–654.
stations in Taihu Lake (Fig. 13), during 15–18 August Chen, C. S., 2003: Marine Ecosystem Dynamics and Mod-
2008, the euphotic depths were estimated. The eu- eling. Higher Education Press, 307–329. (in Chinese)
photic depths in Meiliang Bay and Zhushan Bay were Churchill, J. H., and W. C. Kerfoot, 2007: The impact of
relatively deep Compared to other regions (Fig. 14). surface heat flux and wind on thermal stratification
In contrast, the MLDs in the bays, such as Zhushan in Portage Lake, Michigan. Journal of Great Lakes
Bay, Gonghu Bay and Meiliang Bay, were relatively Research, 33, 143–155.
shallow. The ratio between the MLD and euphotic Doyon, P., B. Klein, R. G. Ingram, L. Legendre, J.-E.
NO. 6 ZHAO ET AL. 1373

Tremblay, and J. C. Therriault, 2000: Influence of Pang Y., W. Zhuang, T. Han, Y. P. Li, and J. B.
wind mixing and upper-layer stratification on phyto- Zhai, 2008: Experiment and model simulation of sus-
plankton biomass in the Gulf of St. Lawrence. Deep- pended solids in Taihu Lake under wind-wave distur-
Sea Res. II, 47, 415–433. bance. Environmental Science, 29(10), 2743–2748.
Gill, A. E., 1982: Atmosphere-Ocean Dynamics. Aca- (in Chinese)
demic Press, 662pp. Payne, R. E., 1972: Albedo of sea surface. J. Atmos. Sci.,
Guarini, J. M., G. F. Blanchard, P. Gross, and S. J. Har- 29, 959–970 .
rison, 1997: Modelling the mud surface temperature Qin, B. Q., and C. X. Fan, 2002: Exploration of con-
on intertidal flats to investigate the spatio-temporal ceptual model of nutrient release from inner source
dynamics of the benthic microalgal photosynthetic in large shallow lake. China Enviromental Science,
capacity. Marine Ecology Progress Series, 153, 25– 22(2), 150–153. (in Chinese)
36. Qin, B. Q., W. P. Hu, and W. M. Chen, 2004a: The Pro-
Hadfield, M. G., and J. Sharples, 1996: Modeling mixed cess and Mechanism of Water Environment Evolve-
layer depth and plankton biomass off the west coast ment in Taihu Lake. Science Press, 389pp. (in Chi-
of South Island, New Zealand. J. Mar. Syst., 8, 1–29. nese)
Huang, D. J., J. L. Su, and J. O. Backhaus, 1999: Mod- Qin, B. Q., W. P. Hu, G. Gao, L. C. Luo, and J. S. Zhang,
elling the seasonal thermal stratification and baro- 2004b: Dynamics of sediment resuspension and the
clinic circulation in Bohai Sea. Continental Shelf Re- conceptual schema of nutrient release in the large
search, 19, 1485–1505. shallow Lake Taihu, China. Chinese Science Bulletin,
Kara, A. B., P. A. Rochford, and H. E. Hurlburt, 2000: 49(1), 54–64. (in Chinese)
An optimal definition for ocean mixed layer depth. Rao, Y. R., and D. J. Schwab, 2007: Transport and mix-
J. Geophys. Res., 105(c7), 16803–16821. ing between the coastal and offshore waters in the
Kara, A. B., R. W. Helber, T. P. Boyer, and J. B. Elsner, great lakes: A review. Journal of Great Lakes Re-
2009: Mixed layer depth in the Aegean, Marmara, search, 33, 202–218.
Black and Azov Sea. Part I: General features. J. Mar. Raybov, A. B., L. Rudolf, and B. Blasius, 2010: Vertical
Sys., 78, s169–s180. distribution and composition of phytoplankton un-
Kim, T. A., and Y. K. Cho, 2011: Calculation of heat der the influence of an upper mixed layer. Journal of
flux in a macrotidal flat using FVCOM. J. Geophys. Theoretical Biology, 263, 120–133.
Res., 116, C03010, doi: 10.1029/2010JC006568. Shang, Z. T., J. Ren, M. R. Qin, Y. Xia, L. He, and
Lehman, J. T., 2002: Mixing patterns and plankton Y. W. Chen, 2010: Relationships between climate
biomass of the St. Lawrence Great Lakes under cli- change and cyanobacterial bloom in Taihu Lake. Chi-
mate change scenarios. Journal of Great Lakes Re- nese Journal of Ecology, 29(1), 55–61. (in Chinese)
search, 28(4), 583–596. Sverdrup, H. U., 1953: On condition for the vernal bloom-
Lövstedt, C. B., and L. Bengtsson, 2008: The role of ing of phytoplankton. Journal du Conseil, 18, 287–
non-prevailing wind direction on suspension and re- 295.
distribution of sediments in a shallow lake. Aquatic Tirok, K., and U. Gaedke, 2007: The effect of irradiance,
Sciences, 70, 304–313. vertical mixing and temperature on spring phyto-
Maggiore, A., M. Zavatarelli, M. G. Angelucci, and N plankton dynamics under climate change: Long-term
Pinardi, 1998: Surface heat and water fluxes in the observation and model analysis. Oecologia, 150, 625–
Adriatic Sea: Seasonal and interannual variability. 642.
Phys. Chem. Earth, 23(5), 561–567. Tuan, N. V., K. Hamagami, K. Mori, and Y. Hirai, 2009:
Nagai, T., H. Yamazaki, H. Nagashima, and L. H. Kan- Mixing by wind-induced flow and thermal convec-
tha, 2005: Field and numerical study of entrainment tion in a small, shallow and stratified lake. Paddy
laws for surface mixed layer. Deep-Sea Res. II, 52, and Water Environment, 7, 83–93.
1109–1132. Zhang, Y. L., B. Q. Qin, W. P. Hu, S. M. Wang, and Y.
Obata, A., J. Ishizaka, and M. Endoh, 2009: Global ver- W. Chen, 2006: Temporal-spatial variation of euph-
ification of critical depth theory for phytoplankton hotic depth of typical lake regions in Lake Taihu and
bloom with climatological in situ temperature and its ecologyical environmental significance. Science in
satellite ocean color data. J. Geophys. Res., 101(c9), China, 49(4), 431–442. (in Chinese)
20657–20667 Zhang, Y. C., X. Qian, Y. Qian, L. Genfal, and T.
Pang, Y., and P. M. Pu, 1996: Numerical simulation of Ishikawa, 2008a: Field measurement and analysis on
three-dimension wind-driven current in Taihu Lake. diurnal stratification in Taihu Lake. Environmental
Acta Geographica Sinica, 51(4), 322–328. (in Chi- Science and Management, 23(6), 117–121. (in Chi-
nese) nese)
Pang, Y., T. Han, Y. P. Li, and J. B. Zhai, 2007: Simu- Zhang, Y. C, X. Qian, I. Tadaharu, and F. X. Kong,
lation and model computation on dynamic release of 2008b: Study on diurnal stratification in a typi-
nutrition factors of bottom mud in Taihu Lake. Chi- cal shallow Lake—Taihu Lake. Sichuan Environment,
nese Journal of Enviromental Science, 28(9), 1960– 27(3), 45–48. (in Chinese)
1964. (in Chinese)