1.

INTRODUCTION
The importance of lakes as stores and emitters of heat has been pointed out by different authors (see
[1]
for a review) and it is known that for any given place in the world,
their role depends mainly on their bathymetry [2]. Recently, it has also been observed that temperature increases in large la
kes
at mid to high latitudes of the northern
hemisphere are greater than in other locations [
3].
Models have been and are useful to quantify the effects of warming from other stressors [
4].
and to investigate the
feedback between the lake and the atmosphere
[5].
Physical mechanisms in lakes associated with the multiscale variability of the basin climate have been studied
[6]
and
on the other hand lake parameterization schemes have been implemented into numerical weather prediction models to improve the
ir
performance [8]. Complementary to
different studies about the accuracy of the lake models and their sensitivity to climate change ([9], [6], [7], [10], [11]),
we
present modeling results to describe how cloud
cover variation can affect the lake variables, which in turn drive atmospheric feedback processes. Fractional cloud cover is
sti
ll a poorly determined variable that is included
in the air

sea bulk formulas of momentum and of heat surface fluxes. On the other hand, and in the context of climate change, cl
oud cover trends have been documented
[12], [13]).
2.

OBJECTIVES
We present the preliminary results of a study, based on numerical simulations, of the effect of cloud cover on the sea surfac
e t
emperature (SST) and on the different
components of the net surface heat flux, taking into account different morphometric characteristics of the lake and different
wi
nd forcings. The aim of this work is:
1.

to evaluate the impact of inaccurate values of the cloud cover on the model predictions, and
2.

to discuss the effects, due to cloud cover variation, of feedback between the lake and the atmosphere.
•
For this study we have implemented, in the Princeton Ocean Model (
http://www.aos.princeton.edu
), the momentum and heat
surface fluxes described in the POM user’s guide (June 2004) after correcting typographical errors in the document.
•
For
sensible heat
we used the formula
Q
s
=
r
A
C
PA
C
H
│U
10
(T
10

T
W
)│, where
r
A
is the air density, C
PA
is the specific heat, C
H
is
the Stanton number, U
10
is the wind speed, T
10
is the temperature of the air at 10 m above the water surface, and T
W
is the
temperature of the water surface.
•
For
the laten heat
Q
L
=Lr
A
C
E
U
10
(q
w

q
10
) where L is the latent heat of evaporation ((2.501

0.002T
w
)
×
10
6
J Kg

1
), C
E
is the
Dalton number, q
w
is the specific humidity and q
10
is the air humidity at the 10 m height.
•
For
long

wave radiation
we used the formula
LWR=0.98s(T
w
+273)4(0.39

0.05e
10
1/2
)(1

0.8n)

s[(T
w
+273)
4

(T
10
+273)
4
].
•
Short

wave radiation
has been parameterized as SWT=SC(sin
a d
csc
a
)(1

0.71n)(1

b
) where SC=1370 W m

2
is the solar
constant,
d
=0.85 is the atmospheric transmission coefficient, α the sun’s altitude, and
b
the albedo.
•
The temperature surface boundary condition is given by the term K
H
(dT/dz) which comes from the heat balance on a control
surface layer, which is represented in
Figure 1
. Note that short

wave radiation is absent from the balance shown in Figure 1
since this absorption takes place in the consecutive layers underlying the control volume, after being partially reflected, a
nd
doesn’t contribute to the balance in the upper control surface layer. The attenuation of the incoming solar flux in the water
column is modeled with a ia Jerlov type parameterization [16].
In this work, done in the context of the CLIMSEAS project, we considered separate bathymetries of the two
lobes of the Large Aral Sea [14] to not have to take into account the important exchanges between the two
lobes [15]. Accordingly, the results cannot be thought as real for the case of the Aral Sea, but rather as
information to be discussed in a more general context. Contours of the bathymetric maps of both lobes
(
Figure 2
) have been slightly modified so that the ratio of the mean depth of cells and their standard
deviation is the same in both lobes. That is, for the western lobe the mean depth is 26.27
8.55 m and for the
eastern it is 2.14
0.69 m. Sigma levels have been chosen so that, given the mean depth of both lobes, the
bulk sea surface temperature will be obtained at 1.3 cm from the surface in both cases. We took 28 levels
growing exponentially toward the bottom for the western lobe and 12 levels for the eastern lobe. Sensitivity
of the model to the thickness of the first layer has been checked for the range below 6 cm and the results
are not dependent on its vertical extension. The horizontal resolution of the bathymetric matrix along the
meridian and the parallel is 0.0067º and 0.0087 º, respectively.
3.1.

The model
3.

METHODS
Francesc Forcat
(1),
Elena Roget
(1),
Lydia

R Dmitrieva

Arrago
(2),
Marina V. Shatunova
(2),
Valentina M. Khan
(2)
(1):
University of Girona, Catalonia, Spain
.
elena.roget@udg.edu
(2): Hydrometeorological Center of Russia, Russia Federation.
Figure 1
.

Heat balance in a control
layer at the surface from where the
temperature surface boundary
condition is obtained.
3.2

The bathymetry
Figure 2
.

Bathymetric models used in this study corresponding to the modified eastern and western lobes of
the Large Aral Sea in 2005
[1]
Limnol. Oceanogr., 54, 2315

2329, 2009.
[2]
Hutchison, G.E., 1957. John Willey and Sons.
[3]
Geophys. Res. Lett., 37, L22405, doi:10.1029/2010GL045059.
[4]
Limnol. Oceanogr., 53, 404

410, 2008.
[5]
Int. J. Climatol., 24, 57

75, 2004.
[6]
Mon. Weather Rev., 134,
3588

3609, 2006.
[7]
Theor Appl. Climat, 79, 55

69, 2004.
[8]
Boreal Environ. Res., 15, 218

230, 2010.
[9]
J. Geophys. Res., 115, C12076, 10.1029/2010JC006269, 2010.
[10]
J. Geophys. Res. 13(C4). C04036
[11]
Limnol. Oceanogr., 55, 2246

2261, 2010.
[12]
Science,
295, 841

844,2002.
[13]
J. Geophys. Res., 115, D15301, doi:10.1029/2009JD013606.
[14]
J. Mar. Syst., 76, 263

271, 2009.
[15]
Hydrol. Earth Sys. Sci., 13, 2265

2271, 2009.
[16]
J. Phys Ocean., 7, 952

956, 1997.
•
For all simulations the model was initialized with a temperature profile measured in the western lobe in October 2005 [14] an
d r
epresented in
Figure 3.
Temperature at the eastern lobe has been set equal to that at the same depth in the western lobe. Salinity was considered con
st
ant and
equal to 70 ppt in both lobes.
•
The time step, which satisfied constraints based on the Courant

Friedrich

Levy criterion, was 6 s for the external mode and 60 s
fot the internal
mode. Zero normal velocities were used as the lateral conditions.
•
For all the simulations the temperature and the humidity of the air considered in the bulk formulas have been set equal to th
ose
measured at
Uqly station located at a latitude of 44.583 N and a longitude of 60.933 E. The wind field velocity was set at a constant spe
ed
of 2.5 m/s and
blowing from the south for the first simulation. For the last simulation the wind field velocity was taken from the NCEP Clim
ate
Forecast System
Reanalysis data basis.
•
The model was run for 17 days starting on 27 September 2005, but the first three days of the simulations are not considered
for
the discussion.
3.2

Initial conditions, forcing and model set

up
Figure 3
.

Initial
conditions for the bulk
stratification
Acknowledgments.
We are gratefull to V. Tischenko for providing wind field data and to X. Baca for the technical support with the computationa
l
resources used for this work. Study has been done within
the framework of the CLIMSEAS project (FP7

2009

IRSES

247512).
As observed in Figures 4, for the studied cases, when the cloud cover was
set to zero, the SSF for both lobes decreased, and when the cloud cover
was one, SST increased also for both lobes. As expected, ranges of
variations of SST for the shallower bathymetry (Figure 4b) are much larger
than for the deeper one.
When the bathymetries were flattened so that the
total depth was equal to the mean depth, the results
–
not shown
–
were
the same (differences not larger than 0.1º for the SST and 1 W/m
2
for the
different components of the heat surface fluxes).
The effective surface flux (line TOTAL in
Table 1
) was found to vary in all cases by about 150 W/m
2
when cloud cover variation was considered to change from 0 to 1, but
for the shallower lobe with a variable wind field variation is even larger. Although no detailed analysis of the accuracy of
the
predicted results has been performed, the
range of variation of the effective surface flux is much larger than expected errors in the model outputs, and the direction
of
the variation is as expected. Accordingly, we
can assume that the range of variation of 150 W/m
2
suggests a variation of the order of 10 W/m
2
for a variation of 0.1 in the cloud cover.
The results presented here show that a variation in cloud cover of 0.1 can affect the SST and the effective surface heat flux
es
in a range above the model resolution.
Although differences depend on the system bathymetry, the stratification background, and the meteorological forcing, it is wo
rth
while to run several other simulations to
try to systematize the expected differences by isolating the different factors that can affect the results.
The results indicate that
a)
for large lakes setting a constant cloud cover when it is not the case can introduce relevant errors into the results, and
b)
documented cloud
cover trends can affect the lake dynamics and its effect on the local meteorology.
Figures 4.

Sea surface temperature (SST) for the two extreme conditions of cloud cover (n=0
and n=1 ) for the cases when (a) the deeper lobe and (b) the shallower were forced with the
constant southern wind and with the variablle wind field obtained from reanalysis as described
in previous section.
(b)
(a)
Figures 5.

Components of the surface heat fluxes for the two
extreme conditions of cloud cover (n=0, n=1) for the cases when the
deeper (a, c) and the shallower (b,d) lobes were forced with the
constant southern wind of 2.5 m/s (c,d) and with a variable wind field
(a, b).
Table 1.

Mean values (after
3 days of simulation )
of the
different components of the
surface heat fluxes.
T
he
effective long

wave radiation
has been separated into two
components, T1 depending
on both the cloud cover and
the SST, and T2 depending
directly only on the SST.
SWR is the short

wave
radiation.
4.

RESULTS
5.

CONCLUSIONS
(a)
(b)
(c)
(d)
Σχόλια 0
Συνδεθείτε για να κοινοποιήσετε σχόλιο