JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. C6, 3052, 10.

1029/2000JC000679, 2002

Differences in heat budgets of the near-surface

Arabian Sea and Bay of Bengal: Implications
for the summer monsoon
S. S. C. Shenoi, D. Shankar, and S. R. Shetye
Physical Oceanography Division, National Institute of Oceanography, Goa, India
Received 24 October 2000; revised 30 November 2001; accepted 19 December 2001; published 15 June 2002.

[1] An analysis of the heat budgets of the near-surface Arabian Sea and Bay of Bengal
shows significant differences between them during the summer monsoon (June–
September). In the Arabian Sea the winds associated with the summer monsoon are
stronger and favor the transfer of heat to deeper layers owing to overturning and turbulent
mixing. In contrast, the weaker winds over the bay force a relatively sluggish oceanic
circulation that is unable to overturn, forcing a heat budget balance between the surface
fluxes and diffusion and the rate of change of heat in the near-surface layer. The weak
winds are also unable to overcome the strong near-surface stratification because of a low-
salinity surface layer. This leads to a shallow surface mixed layer that is stable and
responds quickly to changes in the atmosphere. An implication is that sea surface
temperature (SST) in the bay remains higher than 28
C, thereby supporting large-scale
deep convection in the atmosphere during the summer monsoon. The atmospheric heating
associated with the convection plays a critical role in sustaining the monsoon winds, and
the rainfall associated with it, not only over the bay but also over the Indian subcontinent,
maintains a low-salinity surface layer. In the Arabian Sea the strong overturning and
mixing lead to lower SST and weak convective activity, which in turn, lead to low rainfall
and runoff, resulting in weak stratification that can be overcome easily by the strong
monsoon winds. Thus, in both basins, there is a cycle with positive feedback, but the
cycles work in opposite directions. This locks monsoon convective activity primarily to
the bay. INDEX TERMS: 4504 Oceanography: Physical: Air/sea interactions (0312); 4572 Oceanography:
Physical: Upper ocean processes; 4599 Oceanography: Physical: General or miscellaneous; 9340 Information
Related to Geographic Region: Indian Ocean; KEYWORDS: Arabian Sea, Bay of Bengal, heat budget, sea
surface temperature, monsoon, mixed layer processes

1. Introduction similarities between the two basins: both are located in

the same latitude band and are semienclosed basins that
[2] The north Indian Ocean, a tropical basin bounded in
open into the equatorial Indian Ocean in the south. There
the north, comes under the influence of the seasonally
are also meteorological similarities: both are forced by the
reversing monsoon winds, making it unique among the
changing monsoon winds and receive similar amounts of
world oceans. The monsoon winds reverse twice a year,
solar radiation at the top of the troposphere.
blowing generally from the southwest during the summer
[4] In spite of this, there are striking dissimilarities
monsoon (June –September) and from the northeast during
between the two basins. First, the winds over the two basins
the winter monsoon (November– February). March – May
are different, this difference being most striking during the
and October are the months of transition between the two
summer monsoon (Figure 2). This is a consequence of the
monsoons. The summer monsoon winds are much stronger
Arabian Sea being bounded on its west by the highlands of
than the winter monsoon winds; hence the annual mean
East Africa. These highlands serve as the western boundary
winds over the basin are southwesterly. This temporal
for the low-level atmospheric flow, leading to the formation
asymmetry extends to rainfall, with most (
80%) of the
of an atmospheric ‘‘western boundary current,’’ just like in
rainfall over the north Indian Ocean and the Indian subcon-
the oceans [Anderson, 1976]. The resulting low-level jet,
tinent occurring during the summer monsoon.
called the Findlater Jet [Findlater, 1969], makes the winds
[3] The Indian peninsula splits the north Indian Ocean
over the Arabian Sea more than twice as strong as those
into two basins, the Arabian Sea in the west and the Bay of
over the bay, which does not experience a similar western
Bengal in the east (Figure 1). There exist geographical
boundary effect. Second, precipitation exceeds evaporation
in the bay; evaporation exceeds precipitation in the Arabian
Copyright 2002 by the American Geophysical Union. Sea. Annual rainfall over the bay varies from 1 m off
0148-0227/02/2000JC000679 southeast India to more than 3 m in the Andaman Sea and

5-1
5-2 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL

Figure 1. The geography of the north Indian Ocean. The

two major rivers, Ganga and Brahmaputra, that debouch into
the northern Bay of Bengal, are indicated by RG and RB,
respectively. The hatched areas show the two control
volumes used for estimating the heat budgets; the southern
boundary is fixed at 6
N, which roughly marks the southern
tip of Sri Lanka, and the northern boundary is fixed at 25
N.
The dark strips along the western boundary of the basins
represent the coastal strips used for computing the contribu-
tion of coastal pumping (see 3.2.2) to the heat budget.

the coastal region north of it [Baumgartner and Reichel,

1975]; annual rainfall over the Arabian Sea is barely 1 m.
The bay receives an annual runoff of 1.5  1012 m3 from
rivers flowing into it, the runoff from the Ganga and the
Brahmaputra into the northern bay being the fourth largest
in the world [Martin et al., 1981; Shetye, 1993]; runoff from
rivers into the Arabian Sea is meager, and most of it is
confined to its eastern boundary. Therefore the surface layer
in the bay is much fresher than that in the Arabian Sea; the
average salinity of the top 50 m in the Arabian Sea exceeds
that in the bay by nearly 3 psu. As a consequence, typical
profiles of temperature and salinity in the two basins differ
considerably (Figure 3). The large inflow of freshwater from
precipitation and runoff results in strong near-surface strat-
ification in the bay. The rapid increase in salinity with depth
near the surface, commonly observed in the bay, is not seen
in the Arabian Sea, where salinity in the top 50 m is greater
than that below. The profile in the bay also shows striking
temperature inversions, these being most common during
winter [Shetye et al., 1996].

Figure 3. Typical profiles of temperature (
C), salinity

(psu), and sq (kg m3) in the Arabian Sea (19
460N,
64
360E; 6 March 1994) and the Bay of Bengal (18
570N,
88
440E; 24 December 1991). Note the sharp decrease in
salinity near the surface in the bay.

[5] These differences extend to the evolution of sea

surface temperature (SST) in the two basins. Prior to the
Figure 2. Wind stress over the north Indian Ocean during onset of the summer monsoon, during April – May the north
July [Josey et al., 1996]. The units are N m2. Indian Ocean becomes the warmest area among the world
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL 5-3

Figure 4. (left) SST and (right) average temperature over the top 50 m in the Arabian Sea (solid) and
Bay of Bengal (dashed). The temperatures were averaged over the region between 6
and 25
N in both
basins (see Figure 1 for the domain used for averaging). The SST was computed from the weekly SST
data of [Reynolds and Smith, 1994]; the average temperature over 50 m was computed from the
climatology of Levitus and Boyer [1994]. The units are degrees Celsius.

oceans [Joseph, 1990]. Soon after the onset of the monsoon We conclude by discussing the implications of the heat
in June, the winds strengthen, and SST decreases. The budget and energetics for the coupling of the ocean to the
Arabian Sea cools rapidly, but SST in the bay remains atmosphere (section 5).
higher than 28
C (Figure 4), the threshold for deep con-
vection in the atmosphere over tropical oceans [Gadgil et al., 2. Evolution of SST in the North Indian Ocean
1984; Graham and Barnett, 1987; Sud et al., 1999].
[8] The evolution of SST in the north Indian Ocean, as
[6] This difference in SST is reflected in the convective
seen in weekly climatology, is shown in Figure 6. The
activity in the atmosphere, with the convection over the bay
being perhaps the largest in the tropics during summer. The
convective activity associated with the summer monsoon,
which has been recognized as an important source of
energy for the tropical circulation [Krishnamurti, 1971] is
maximum in the northern bay. The number of low-pressure
systems (LPSs) that form there by far exceed those in the
rest of the north Indian Ocean [Mooley and Shukla, 1989]
(Figure 5). These LPSs move westward over India [Sikka,
1977; Goswami, 1987; Mooley and Shukla, 1989] and bring
rainfall to central and northern India. (The condition of a
closed isobar imposed by Mooley and Shukla [1989]
reduces the number of identifiable LPSs in the central
bay, where convective activity is not as low compared to
the northern bay as might be suggested by Figure 5
(S. Gadgil, personal communication, 2000).)
[7] Given that SST > 28
C is a necessary condition for
deep convection in the tropics, it is likely that the higher
precipitation in the Bay of Bengal and the genesis of a large
number of LPSs there are related to its being warm enough
to support convection. The processes that ensure this,
however, remain to be elucidated. It is this issue that we
address in this paper; simultaneously, we examine the
processes that leave the Arabian Sea cooler. The paper is
organized as follows. After describing the evolution of SST Figure 5. Total number of LPSs (low-pressure systems)
in the north Indian Ocean (section 2), we analyze the heat that formed during the summer monsoon (June –September)
budget of the upper ocean (section 3) and the energetics of over 4
 4
blocks of the north Indian Ocean and the
the mixed layer (section 4). This is done using climatologies Indian subcontinent during 1888 – 1983 [adapted from
of air-sea fluxes, radiative fluxes, winds, and hydrography. Mooley and Shukla, 1989].
5-4 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL

Figure 6. Evolution of SST in the north Indian Ocean. Weekly SST data [Reynolds and Smith, 1994] for
1982 –1998 were combined to form a weekly climatology. The data set has a spatial resolution of 1

1
. Contours are drawn only for SST greater than the deep convection threshold of 28
C; the contour
interval is 1
C.
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL 5-5

Figure 7. Seasonal variation of mixed layer depth (meters) in the north Indian Ocean. The mixed layer
depth is assumed to be the depth at which st is 0.2 kg m3 greater than that at the surface. The contour
interval is 5 m. The computations are based on the seasonal climatologies of Levitus and Boyer [1994]
and Levitus et al. [1994].

climatology, which is based on the weekly SST data of
Reynolds and Smith [1994], is defined over 1982 –1998.
Only SST above the convection threshold of 28
C is shown.
[9] During January – February, SST in the north Indian
Ocean is <28
, except for a small region off southwest India
[Shenoi et al., 1999]. SST rises gradually as the Sun moves
poleward over the Northern Hemisphere, with the Bay of
Bengal warming faster than the Arabian Sea. SST exceeds
30
C over most of the north Indian Ocean by May; for the
week centered on 15 May, surface water having temper-
atures >30
C occupies 40% (63%) of the total area of the
Arabian Sea (Bay of Bengal). With the onset of the summer
monsoon in June, SST starts falling in both basins and
continues to fall until the end of August. For the week
centered on 14 August, surface water having temperatures
>28
C occupies about 17% (100%) of the Arabian Sea (Bay
of Bengal). Once the summer monsoon starts collapsing in
September, both basins start warming again. The cold, dry
continental winds that blow from the northeast cool the
basins again in November, this phase lasting till February.
[10] Three conclusions can be drawn from the evolution
of SST described in Figure 6. First, though both the Arabian
Sea and the Bay of Bengal show a bimodal SST distribution
in time (Figure 4), the latter remains warmer throughout the
year. Second, the northern bay is warmer than the rest of the
bay during the summer monsoon. Third, SST in the Arabian
Sea exhibits a distinct spatial gradient, with SST in the west
being less than that in the east.
3. Heat Budget of the Upper Ocean
[11] To determine the causes of the higher SST in the Bay
of Bengal, we analyze the heat budget of the upper ocean. A
preliminary heat budget was computed for the summer
monsoon for the Arabian Sea by Düing and Leetmaa
[1980]. They concluded that the upper ocean loses heat
during this period because the heat lost owing to advection
across the equator and cooling due to upwelling exceeds the
heat gained from the atmosphere; coastal upwelling was
found to be the dominant term during the summer monsoon.
An analysis based on observations, like that of Düing and
Leetmaa [1980], is not available for the Bay of Bengal.
Loschnigg and Webster [2000] used a numerical model to
confirm that advection across the equator is important for
the heat budget of the north Indian Ocean. Düing and
Leetmaa [1980] did not have access at that time to the kind
of data sets that are available today; the availability of more
accurate and reliable climatologies makes possible a more
detailed heat budget analysis, which we present in this
section. We define the region of interest (Figure 1) to
exclude the area south of 6
N, which marks the southern
tip of the Indian subcontinent, and the area north of 25
N.
To define the control volume completely, we have to fix its
lower boundary; we fix this at 50 m on the basis of the
observed mixed layer depths in the two basins (Figure 7).
Sensitivity tests show that changing the depth of the control
volume to, say, 75 m does not influence the conclusions
drawn in this paper. Also, the average temperature over 50
m follows a bimodal distribution (Figure 4), like SST,
implying that the choice of control volume is reasonable
for identifying the processes that control the heat budget of
the near-surface ocean.
[12] The rate of change of heat in the control volume is
balanced by the fluxes of heat through its boundaries owing
to advective and nonadvective processes. Starting with the
equation for conservation of heat, assuming the fluid to be
5-6 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
incompressible, and using Gauss’s theorem, we can express
this balance as
Z Z
1 @   1   1

rw Cp T dV ¼ 
rw Cp T u  ndS 
A @t V A S A
where T is the temperature, u is the velocity, F represents
the nonadvective fluxes, n is the unit vector normal to the
surfaces bounding the control volume (directed outward),
and the integral on the left-hand side is evaluated over the
control volume;
rw ¼ 1026 kg m3 is the mean density of
the upper layer, and Cp = 3902 J kg 1 K 1 is the specific
heat of seawater. The temperature T in equation (1) (and in
other equations below) is based on the climatology of
Levitus and Boyer [1994], which provides vertical profiles
of temperature on a 1
 1
spatial grid. The nonadvective
term F consists of the surface fluxes and diffusion through
the bottom of the control volume; diffusion through the
lateral boundaries, including the open southern boundary, is
negligible in comparison. We find it convenient to divide all
terms by A, the area of the control volume at the ocean
surface, because it allows us to represent the surface fluxes
in W m2, the units used in most climatologies.
[13] The control volume is heated by the surface fluxes.
We consider diffusion of heat through the bottom and the
advective fluxes together under the appellation ‘‘oceanic
processes’’ because they are processes intrinsic to the ocean.
We discuss the surface fluxes and oceanic processes sepa-
rately. Unlike in equation (1), in the discussion below we
consider the fluxes positive if the control volume, or the
upper ocean, gains heat.
3.1. Surface Fluxes
[14] The net flux of heat through the surface is
Qsf ¼ Rs þ Rl þ Ql þ Qs ;
where Rs, Rl, Ql, and Qs represent the net shortwave
radiation, net longwave radiation, latent heat flux, and
sensible heat flux through the surface, respectively. A
refined version of the Comprehensive Ocean-Atmosphere
Data Set (COADS) compiled by Josey et al. [1996] (often
called the Southampton Oceanographic Center Climatol-
ogy), having a resolution of 1
 1
, is used to estimate the
heat and radiative fluxes at the surface. This data set is
preferred to a similar climatology based on COADS
compiled by Oberhuber [1988] and to a climatology
derived from the National Centers for Environmental
predication/National Center for Atmospheric Research
Reanalysis Programme [Kalnay et al., 1996] because it is
more recent, uses better methods for generating the
climatology, and has been compared extensively with
observations [Josey et al., 1999; Weller et al., 1999].
[15] The climatological monthly mean surface fluxes are
shown in Figure 8 (top). Like SST (Figure 4), Rs and Ql in
both basins show a bimodal distribution. Except during
December – February, Rs is greater by
30 W m2 over the
Arabian Sea. It is minimum during the summer monsoon
(owing to clouds) and during winter. Ql is maximum during
the onset of the summer monsoon in June, when the winds
strengthen and humidity increases rapidly in the lower
troposphere, and during winter, when the winds are weaker
Z
but humidity is low. Except during winter, when the latent
heat lost by the Arabian Sea is greater by
25 W m2, both
F  ndS;
S basins lose roughly the same amount of heat owing to Ql.
ð1Þ Loss of heat owing to longwave radiation, Rl, is also the
same in both basins. This loss is minimum during the
summer monsoon because of an increase in the downward
longwave radiation from the atmosphere due to increased
humidity and cloudiness and is maximum during winter
when the skies are clear. Qs is
5 W m2 in both basins and
contributes little to the heat budget.
[16] Qsf is positive, except during December – January,
when it is negligible; this gain of heat by the north Indian
Ocean also exhibits a bimodality like the SST. Except
during January – March the Arabian Sea receives more heat
(
20 W m2) than the Bay of Bengal from the atmosphere.
For example, during July, which represents the peak of the
summer monsoon, the Arabian Sea (Bay of Bengal)
receives 240 (212) W m2 through net shortwave radiation,
of which
70% (75%) is returned to the atmosphere
through longwave radiation and latent and sensible heat
fluxes. Thus the shortwave radiation into the bay is a little
less than that into the Arabian Sea, but both basins return
roughly the same amount of heat to the atmosphere.
Nevertheless, the bay remains much warmer than the
Arabian Sea throughout the year, implying a major role
for oceanic processes in the heat budget of the north Indian
Ocean.
3.2. Oceanic Processes
[17] Diffusion of heat through the bottom and advection
of heat constitute the oceanic processes. We split the advec-
tive processes into two parts. First, there is flow through the
open southern boundary, with this mass flux being balanced
by flow through the bottom of the control volume; this is the
meridional overturning cell that exports heat out of the north
ð2Þ Indian Ocean [Lee and Marotzke, 1997; Garternicht and
Schott, 1997]. Second, cross-shore flow at the western
boundary layer results in upwelling (downwelling) in the
western boundary regime; this vertical mass flux through the
bottom of the coastal strip (Figure 1) is balanced by a vertical
mass flux through the bottom of the rest of the control
volume. We call this process coastal pumping. It also causes
overturning and removes heat from the control volume, but
unlike meridional overturning, it does not remove heat from
the north Indian Ocean. Vertical mass flux at the northern
and eastern boundaries is ignored in comparison to that at the
western boundary. Also ignored is the contribution of the
negligible flow between the Red Sea and the Persian Gulf
and the control volume defining the Arabian Sea [Düing and
Leetmaa, 1980]. Therefore the net flux of heat due to oceanic
processes is (Figure 8 (middle))
Qop ¼ Qmo þ Qcp þ Qd ; ð3Þ
where Qmo, Qcp, and Qd are the heat fluxes due to
meridional overturning, coastal pumping, and diffusion,
respectively.
3.2.1. Meridional overturning
[18] The transport across the southern boundary is bal-
anced by a mass flux through the bottom of the control
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL 5-7

Figure 8. The heat budget of the upper ocean in (left) the Arabian Sea and (right) the Bay of Bengal.
(top) Climatological monthly mean net shortwave radiation (Rs), net longwave radiation (Rl), latent heat
flux (Ql), and sensible heat flux (Qs) and their sum, the net surface flux (Qsf). (middle) The fluxes due to
oceanic processes: meridional overturning (Qmo = Qmoe + Qmog), coastal pumping (Qcp = Qcpe + Qcpg),
and diffusion (Qd); the sum of the fluxes due to these processes (Qop) is also shown. The individual
contributions of the Ekman and geostrophic components to meridional overturning and coastal pumping
are shown in Figure 9. (bottom) The rate of change of heat (qt) in the control volumes and the net flux of
heat (Q = Qsf + Qop) into or out of them. The estimated error on qt (23 W m2), based on the standard
error of 0.3
C on the annual mean temperature [Levitus and Boyer, 1994], is also shown (shading). The
southern (northern) boundary of the control volumes is at 6
N (25
N), and their depth is 50 m. The units
are W m2 and a positive flux implies that the control volumes gain heat.

volume. The difference between the vertical average of flux is estimated as

temperature at the southern boundary and the basin-wide Z
average temperature at the bottom of the control volume 1
Qmo ¼
rw Cp ^vTmo dx; ð4Þ
results in a net flux of heat owing to this process; this heat A x
5-8 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
Figure 9. Flux of heat (W m2) due to the Ekman and geostrophic components of meridional
overturning (Qmo) and coastal pumping (Qcp) in the (left) Arabian Sea and (right) Bay of Bengal. The
Ekman components of the two processes are important for the heat budget of the Arabian Sea.
where ^v is the transport (per unit distance along the southern
boundary) through the southern boundary,
Tmo ¼ T sb  hT i;
and the integral is evaluated along the southern boundary.
T sb is the vertically averaged temperature (over the depth of
the control volume) at the southern boundary, and hTi is the
basin-wide average of the temperature at the bottom of the
control volume. Both Ekman and geostrophic flow con-
tribute to ^v; that is, Qmo = Qmoe + Qmog, and ^v ¼ ^ve þ ^vg ,
where
tx gD @n
^ve ¼ ^vg ¼ : is contrary to the conclusion of Düing and Leetmaa [1980],

rw f f @x
The subscript e refers to Ekman flow, and the subscript g
refers to geostrophic flow. The geostrophic current is
assumed to be constant over the depth of the control volume.
In the above equations, tx is the zonal wind stress, h is the sea
level, f is the Coriolis parameter, g = 9.81 m s2 is the
acceleration due to gravity, and D is the depth of the control
volume.
[19] This approach differs from those used earlier. Düing
and Leetmaa [1980] considered only the geostrophic inflow
in the western boundary regime of Somalia to estimate the
contribution of advection; they ignored the contribution of
Ekman flow and assumed that outflow across the rest of the
southern boundary compensated for the Somali Current.
Garternicht and Schott [1997] and Levitus [1987] consid-
ered only the Ekman flow, which was assumed to be
compensated by a barotropic return flow; the net heat flux
was due to the difference in temperature betweem the
Ekman flow and the return flow. The difficulty with this
approach lies in estimating the temperature of the return
flow, and it seems more suited to control volumes whose
depth is comparable to that of the basin, not to shallow,
surficial control volumes like the one defined here.
[20] Wind stress from the climatology of Josey et al.
[1996] is used to estimate Qmoe (Figure 9). It is stronger in
the Arabian Sea, from which it removes heat (
40 W m2)
during the summer monsoon, but is not a major factor in the
heat budget during the rest of the year, when it brings heat
into the control volume. In the bay, Qmoe removes heat
(
25 W m2) during the summer monsoon.
[21] The 10 day repeat cycle data of sea level anomalies
from the TOPEX/Poseidon altimeter during 1993 – 1997 [Le
Traon et al., 1995; Le Traon and Ogor, 1998; Le Traon et al.,
1998] are used to construct a monthly climatology of sea
level anomalies (h0), which is used to estimate the climato-
logical, monthly mean geostrophic current at the southern
boundary. These anomalies exclude the time mean sea level
h), which has been estimated using the 0/1000 m dynamic
(

height computations from the annual mean profiles of
temperature [Levitus and Boyer, 1994] and salinity [Levitus
et al., 1994]. The total sea level h ¼ h
þ h0 is used to
compute the geostrophic current at the southern boundary.
The contributions of the annual mean and time-varying
components of this process are comparable. Unlike Qmoe,
however, Qmog is not a major contributor to the heat budget
of either basin; it does not exceed 10 W m2 (Figure 9). This
who assumed that the inflow in the Somali Current was
cooler than the outflow across the rest of their southern
boundary and found the geostrophic advection to be an
important process for cooling the Arabian Sea. Hence Qmo
is important for removing heat from the Arabian Sea during
the summer monsoon but is not important for the heat budget
of the Bay of Bengal (Figure 8).
3.2.2. Coastal pumping
[22] Cross-shore flow at the western boundary forces a
vertical mass flux in the western boundary regime (repre-
sented by the dark coastal strips in Figure 1) through the
bottom of the control volume. This has to be balanced by a
vertical mass flux through the bottom of the control volume
over the rest of the basin. The difference in average temper-
ature between the coastal strip and the rest of the control
volume at the bottom results in a flux of heat. We call this
process coastal pumping and evaluate the flux of heat due to
it as
Z
1
Qcp ¼ 
rw Cp Tcp ^  ndl;
u ð5Þ
A l
where u^ is the cross-shore transport (per unit distance along
the coast), n is the unit vector normal to the western
boundary (positive into the coast), and l is the coordinate
along the coast; the integral is evaluated along the coastal
strip. T is the average difference in temperature at the
bottom of the control volume between the coastal strip and
the rest of the basin; that is,
Tcp ¼ hTc i  hTi i;
where hTci (hTii) is the average temperature of the coastal
strip (interior of the basin) at the bottom of the control
volume (the angle brackets indicate a spatial average).
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
[23] Like meridional overturning, coastal pumping also
has Ekman and geostrophic components, i.e., Qcp = Qcpe +
^¼u
Qcpg and u ^e þ u^g ;
knt gD
^e  n ¼
u ^g ¼
u rh;

rw f f Sea, but we use a constant value of k.
where t is the wind stress vector, k is the unit vector normal
to the Earth’s surface, D is the depth of the control volume,
and h is the sea level.
[24] Düing and Leetmaa [1980] showed that upwelling
forced by coastal Ekman pumping along the western boun-
dary is the dominant term in the heat budget of the Arabian
Sea during the summer monsoon. The winds in the western
Arabian Sea turn to blow from the southwest in May,
strengthening with the onset of the summer monsoon in
June. The axis of maximum winds, the Findlater Jet, runs
along the coast until
10
N, where it turns offshore. This
jet forces strong upwelling along the western boundary,
drawing up cooler water from the deeper layers [Schott,
1983]. Since Tcp
3.5
C [Levitus and Boyer, 1994],
coastal Ekman pumping cools the Arabian Sea (
50 W
m2) during the summer monsoon; it is not important
during the rest of the year (Figure 9).
[25] Contribution of coastal Ekman pumping is insignif-
icant in the bay even during the summer monsoon (Figure
9), when the winds along the east coast of India favor
upwelling. Hydrographic observations show that the
upwelling along this coast is weak and restricted to a
shallow surface layer, with downwelling indicated below
it [Shetye et al., 1991]. This is due to the low-salinity
surface layer [Shetye et al., 1991] and remote forcing from
the equator and cyclonic curl of wind stress in the interior
bay, both of which force a downwelling favorable equator-
ward current along the east coast of India [McCreary et al.,
1993; Shankar et al., 1996; McCreary et al., 1996; Vinaya-
chandran et al., 1996]. The strong alongshore winds over-
whelm the remotely forced downwelling, but only near the
surface. Hence the resulting flux of heat does not exceed
8 W m2.
[26] Sea level anomalies from TOPEX/Poseidon are used
together with the dynamic height computations from the
annual mean profiles of temperature [Levitus and Boyer,
1994] and salinity [Levitus et al., 1994] for estimating u ^g .
The contribution of coastal geostrophic pumping (Figure 9)
is stronger in the Arabian Sea, but it is small compared to
the Ekman component; it is maximum during August –
October, when it removes heat from the control volume
(
15 W m2). It is negligible in the bay (<2 W m2).
Hence coastal pumping is important for removing heat from
the Arabian Sea during the summer monsoon (Figure 8).
3.2.3. Diffusion
[27] The flux of heat through the bottom of the control
volume owing to diffusion is computed as
Z the north Indian Ocean, averaged over the top 50 m, is
1 dT
Qd ¼ 
rw Cp k dS;
A S dz
where k = 2  104 m2 s1 is the eddy diffusivity
coefficient, dT/dz is evaluated at 50 m depth, and the integral
is over the surface area of the basins. The choice of k has
5-9
been guided by Zhang and Talley [1998], who estimated that
diathermal diffusivity at 50 m depth in the Indian Ocean
varied from <1  104 m2 s1 at the northern boundary to
3.5  104 m2 s1 at
6
N. They estimated that the eddy
diffusivity coefficient is larger in the bay than in the Arabian
[28] Diffusion always results in a loss of heat from the
control volume (Figure 8). The loss is a minimum during
February – March (
20 W m2 for the Arabian Sea and
35
W m2 for the bay) and a maximum during October –
December (
50 W m2 for the Arabian Sea and
60 W
m2 for the bay).
3.2.4. Net flux due to oceanic processes
[29] In the Arabian Sea the dominant oceanic processes
are coastal Ekman pumping, meridional overturning due to
Ekman flow, and diffusion. Of these, the first two, which are
directly influenced by the winds and cause overturning, are
important primarily during the summer monsoon, when
they remove heat from the control volume. Diffusion, the
only oceanic process not directly influenced by the winds, is
important throughout the year. In the bay these wind-forced
processes have a minor impact on the heat budget; diffusion
overwhelms other oceanic processes (Figure 8).
[30] The reason for this lies in the asymmetry of the wind
field in the north Indian Ocean (Figure 2). The weaker
winds over the bay force a relatively sluggish oceanic
circulation there, making it difficult for overturning or
coastal pumping to remove heat from the control volume
there.
3.3. Rate of Change of Heat
[31] The resultant heat flux owing to surface fluxes and
oceanic processes changes the heat content in the control
volume. The rate of change (qt) is estimated using the left-
hand side of equation (1).
[32] Like SST (Figure 4), qt and the resultant of all the
fluxes through the control volume, Q = Qsf + Qop, show a
bimodal distribution in the two basins, with minima during
winter and the summer monsoon (Figure 8 (bottom)). The
upper ocean in the Arabian Sea loses heat to the atmos-
phere and to the deeper ocean during November – January
and June – August, gaining heat during the other months.
The bay behaves similarly, except that the magnitude of qt
is lower there; for example, the bay loses
20 W m2
during June– August, in contrast to
85 W m2 lost by the
Arabian Sea during July. Q follows the same pattern and is
close to qt, the difference between them not exceeding

30 W m2.
[33] The errors associated with the surface fluxes are
lower than those associated with qt or Qop, which are
primarily a consequence of the errors in the temperature
profiles. For the climatology of Josey et al. [1996] the errors
associated with the surface fluxes are <10 W m2 [Josey
et al., 1999; Weller et al., 1999]. From Levitus and Boyer
[1994] the standard error on the annual mean temperature in
0.3
C, which translates to a standard error of
23 W m2
on qt (shown by the shaded band Figure 8 (bottom)). Since
the standard error on the monthly temperature is likely to be
much higher than that on the annual mean, this implies that
the heat budget for the two control volumes is balanced
within acceptable error bars.
5 - 10 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
Figure 10. Seasonal distribution of ATKE over the north Indian Ocean. The units are 105 W m2, and
the contour interval is 40 units.
[34] Thus, in the Arabian Sea the heat gained at the
surface is balanced by the rate of change of heat in the
control volume and by overturning, coastal pumping, and
diffusion. In the bay the heat gained at the surface is
balanced by the rate of change of heat in the control volume
and the loss of heat by diffusion. It is the difference in the
structure and magnitude of winds that keeps the mean
temperature of the top 50 m of the bay warmer than that
of the Arabian Sea. Though this helps in understanding why
SST in the bay is greater than in the Arabian Sea during the
summer monsoon, there is another aspect we need to
examine: the difference in near-surface stratification and
its implications for the SST. We do this by analyzing the
energetics of the surface mixed layer.
4. Energetics of the Mixed Layer
[35] The seasonal distribution of mixed layer depth
(Figure 7) shows that the mixed layer in the Arabian Sea
is thicker than that in the bay and that it varies much more
with season and in space than the mixed layer in the bay.
During the summer monsoon the mixed layer in the bay
shallows, but that in the Arabian Sea deepens, especially in
the region of anticyclonic Ekman pumping to the south and
east of the Findlater Jet (Figure 2). The energetics of the
mixed layer in the north Indian Ocean is examined by
comparing the turbulent kinetic energy available for mixing
with the energy required for mixing the stratified water
column to a constant depth of 50 m.
[36] The excess of generated turbulent kinetic energy
(TKE) over that dissipated within the mixed layer deter-
mines the energy available for mixing. TKE is generated in
two ways: through the wind stress that acts on the sea
surface and through the loss of potential energy due to
cooling at the surface owing to the exchange of heat with
the atmosphere. We estimate the available TKE (ATKE)
following Kim [1976] and Shetye [1986] and using the
climatologies of Levitus and Boyer [1994], Levitus et al.
[1994], and Josey et al. [1996]. The seasonal distribution of
ATKE (Figure 10) shows it is of the order of 105 W m2 in
both basins, except during the summer monsoon, when the
winds are strong. During the summer monsoon, ATKE over
the Arabian Sea, particularly along the axis of the Findlater
Jet, exceeds 5  103 W m2, while that over the bay is

4  104 W m2. Thus the energy available for mixing is
almost an order of magnitude greater in the Arabian Sea
than in the bay.
[37] The energy required for mixing (ERM) the water
column to 50 m is the difference between the potential
energy of a stratified 50 m column and that of the same
column when it is unstratified (or mixed vertically). The
density required to compute the ERM is estimated using
the climatologies of Levitus and Boyer [1994] and Levitus
et al. [1994]. ERM in the bay, especially in the north, is
much greater than that in the Arabian Sea (Figure 11).
During June –December, ERM in the northern bay is >
12
 103 J m2, but it is <
3  103 J m2 in the Arabian
Sea. This large difference between the two basins is a
consequence of the low salinity of the surface waters in the
bay (Figure 3).
[38] The energy available for mixing in the Arabian Sea
is an order of magnitude greater than that in the bay, but the
converse is true for the energy required for mixing the
surface waters to 50 m. During the summer monsoon a
typical value of ATKE for the Arabian Sea (Bay of Bengal)
is 3  103 W m2 (1  103 W m2) (Figure 10), and that
of ERM is 1  103 J m2 (5  103 J m2) (Figure 11).
Therefore the winds during the summer monsoon will take
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
Figure 11. Seasonal distribution of the ERM the water column to a depth of 50 m in the north Indian
Ocean. The units are 103 J m2, and the contour interval is 1 unit. Values >5  103 J m2 are shaded.
only
4 days, which is comparable to the inertial period at
these low latitudes, to mix the Arabian Sea to 50 m but will
take
2 months in the bay. Thus it is not surprising that the
mixed layer in the bay does not deepen during the summer
monsoon, unlike in the Arabian Sea; the weaker winds in
the bay are incapable of mixing the highly stratified surface
waters beyond the shallow mixed layer depths of
10– 20
m observed there. This mechanism is similar to that in the
barrier layer in the western Pacific and other areas of the
tropical oceans [Godfrey and Lindström, 1989; Lukas and
Lindström, 1991; Sprintall and Tomczak, 1992; Godfrey
et al., 1998; You, 1998], except that the salinity (stratifica-
tion) in the bay is much lower (stronger).
5. Discussion
[39] The main reason for the difference in the heat budget
and energetics between the Arabian Sea and the Bay of
Bengal is the asymmetry in the wind field, which is caused
by geography and atmospheric dynamics on an equatorial b
plane: the existence of the Findlater Jet is a direct conse-
quence of the East African highlands and atmospheric
dynamics [Anderson, 1976]. This difference between the
heat budgets of the two basins has implications for large-
scale air-sea interaction over the north Indian Ocean.
[40] In the Arabian Sea the winds are strong, and strat-
ification is weak owing to evaporation exceeding precip-
itation and the runoff from rivers being meager, except
along the west coast of India. This leads to strong over-
turning and mixing, and the wind-forced processes transport
heat received from the atmosphere to deeper layers. Hence
SST in most of the Arabian Sea is lower than 28
C, below
the threshold for sustaining deep convection in the atmos-
phere. The result is weak convection, which in turn, leads to
5 - 11
low rainfall and runoff, creating a self-sustaining cycle with
positive feedback.
[41] In the bay the weaker winds cannot force the over-
turning necessary to remove heat from the upper ocean, and
the strong stratification inhibits turbulent mixing. Hence
diffusion, a slow process, is the only oceanic process
available for removing heat from the control volume in
the bay. On short timescales of the order of a week the only
way the bay can restrict a rise in SST is through manipu-
lation of surface fluxes. This can be done in two ways: by
increasing latent heat flux, which can lead to the formation
of clouds, and by decreasing the net shortwave radiation
through the medium of clouds [Sud et al., 1999]. This is
best accomplished by LPSs, which involve cloud clusters
and are effective in reducing the net shortwave radiation
received at the surface [Rao et al., 1985; Sanilkumar et al.,
1994].
[42] During the summer monsoon, winds strengthen
over the Bay of Bengal when an LPS forms, leading to
an increase in the available TKE. At the same time, rainfall
increases because of the presence of the LPS, increasing
near-surface stratification. The increase in runoff due to an
LPS lags behind because this occurs some time after the
LPS moves over land. The increase in stratification inhibits
mixing, and the mixed layer remains shallow. Hence the
heat influx from the atmosphere is trapped in the shallow
mixed layer, leading to SST higher than 28
C. Since this
exceeds the threshold for sustaining convective activity in
the atmosphere, there is strong convection over the bay.
Strong atmospheric convection, in turn, leads to higher
precipitation and to the formation of LPSs, which move
westward and northwestward over the Indian subcontinent,
bringing rain to large parts of central and northern India
[Sikka, 1977; Mooley and Shukla, 1989]. Rainfall over land
5 - 12 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL

Figure 12. A schematic of the feedback cycles that lead to the Bay of Bengal being warmer than the
Arabian Sea and sustaining organized convection in the atmosphere, leading to the disproportionate
number of low-pressure systems that form in the northern bay (Figure 5) [Mooley and Shukla, 1989].
PE denotes the difference between precipitation and evaporation. In the Arabian Sea the winds
associated with the summer monsoon are stronger and favor the transfer of heat to deeper layers owing to
overturning and turbulent mixing. The strong overturning and mixing lead to lower SST and weak
convective activity, which in turn, lead to low rainfall and runoff, resulting in weak stratification that can
be overcome easily by the strong monsoon winds. In contrast, the weaker winds over the bay force a
relatively sluggish oceanic circulation that is unable to overturn, forcing a heat budget balance between
the surface fluxes and diffusion and the rate of change of heat in the near-surface layer. The weak winds
are also unable to overcome the strong near-surface stratification due to a low-salinity surface layer. This
leads to a shallow surface mixed layer that is stable and responds quickly to changes in the atmosphere.
Therefore SST in the bay remains higher than 28
C, thereby supporting large-scale deep convection in
the atmosphere during the summer monsoon. The atmospheric heating associated with the convection
plays a critical role in sustaining the monsoon winds, and the rainfall associated with it, not only over the
bay but also over the Indian subcontinent, maintains a low-salinity surface layer. Thus, in both basins,
there is a cycle with positive feedback, but the cycles work in opposite directions. This locks monsoon
convective activity primarily to the bay.

leads to increased runoff from large rivers like the Ganga
and the Brahmaputra, which debouch into the northern bay,
creating a self-sustaining cycle with positive feedback. The
feedback cycles in the two basins are summarized in
Figure 12.
[43] A critical component of these feedback cycles is the
convection threshold of 28
C. If this threshold were lower
by 0.5
C, more than 60% of the Arabian Sea would still
have SST below the threshold and the mean SST in the
basin (Figure 4) would be lower than the threshold during
July, when the monsoon peaks. Hence it would still be
incapable of supporting deep convection in the atmosphere.
Increasing the threshold by 0.5
C implies that the mean SST
in the Arabian Sea would be lower than the threshold during
most of the summer monsoon. In the bay, however, increas-
ing the threshold does not have a similar effect. It appears
from the study of Sud et al. [1999] that the 28
C threshold is
more robust as a lower bound. Hence it is likely that the
feedback cycles described above will hold even if the
threshold is varied; a quantitative study of the sensitivity
of the cycles to the threshold is, however, beyond the scope
of this paper.
[44] In computing the budget for the Arabian Sea we
have averaged over the two highly dissimilar regimes in the
west and the east (section 2). The eastern Arabian Sea is
bounded by the Indian west coast, where rainfall exceeds
200 cm during the year, with most of it occurring during
the summer monsoon. Though this is attributed to orog-
raphy [Sarkar, 1966; Sarkar, 1967], SST in this region
(east of, say, 65
E) also exceeds 28
C until August
(Figure 6). It is likely that the high rainfall over the sea
and the runoff from the small rivers that flow down the
western ghats into the eastern Arabian Sea will lead to
stratification similar to that in the bay during the summer
monsoon, causing a positive feedback that leads to deep
convection in the atmosphere. An analysis of this possi-
bility, however, requires a separate heat budget computa-
tion for the eastern Arabian Sea.
 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
[45] Nevertheless, the conclusions drawn above hold for
the Arabian Sea on the average. Thus, in both the Arabian
Sea and the Bay of Bengal, there is positive feedback
associated with the air-sea interaction processes during
the summer monsoon, leading to self-sustaining cycles,
except that they work in opposite directions in the two
basins. Though there must also be other reasons for the
large difference in the number of LPSs that form in the bay
and the Arabian Sea (orography, upper atmospheric flow,
the large-scale tropical circulation, etc.), it is evident that
the feedback cycles must play a significant role in locking
monsoon convective activity to the Bay of Bengal. In
section 1 we listed a few special features of the north
Indian Ocean that make it a natural laboratory to study
phenomena related to coupling between the tropical atmos-
phere and the ocean. The interplay among SST, spatial
structure of winds, oceanic near-surface stratification, and
atmospheric convection described above is another such
phenomenon.
[46] Acknowledgments. We thank Simon Josey for making available
the Southampton Oceanography Centre climatology and the Collecte
Localisation Satellites (CLS) Space Oceanography Division, Paris, for the
altimeter products. Critical comments from John Toole and two anonymous
reviewers helped improve the manuscript considerably. Ferret, GMT, and
GIMP were used for analysis and graphics; G. S. Michael helped with
Figures 3 and 5. The Department of Ocean Development, New Delhi,
provided financial support. This is NIO contribution 3723.
References
Anderson, D. L. T., The low-level jet as a western boundary current, Mon.
Weather Rev., 104, 907 – 921, 1976.
Baumgartner, A., and E. Reichel, The World Water Balance: Annual Glo-
bal, Continental and Maritime Precipitation, Evaporation and Runoff,
Elsevier Sci., New York, 1975.
Düing, W., and A. Leetmaa, Arabian Sea cooling a preliminary heat budget,
J. Phys. Oceanogr., 10, 307 – 312, 1980.
Findlater, J., A major low-level air current near the Indian Ocean during the
northern summer, Q. J. R. Meteorol. Soc., 95, 362 – 380, 1969.
Gadgil, S., P. V. Joseph, and N. V. Joshi, Ocean-atmosphere coupling over
monsoon regions, Nature, 312, 141 – 143, 1984.
Garternicht, U., and F. Schott, Heat fluxes of the Indian Ocean from a
global eddy-resolving model, J. Geophys. Res., 102, 21,147 – 21,159,
1997.
Godfrey, J. S., and E. Lindström, The heat budget of the western equatorial
Pacific surface mixed layer, J. Geophys. Res., 94, 8007 – 8017, 1989.
Godfrey, J. S., R. A. Houze, R. H. Johnson, R. Lukas, J.-L. Redelsperger,
A. Sumi, and R. Weller, Coupled Ocean-Atmosphere Response Experi-
ment (COARE): An interim report, J. Geophys. Res., 103, 14,395 –
14,450, 1998.
Goswami, B. N., A mechanism for the west-northwest movement of mon-
soon depressions, Nature, 326, 376 – 377, 1987.
Graham, N. E., and T. P. Barnett, Observations of sea-surface temperature
and convection over tropical oceans, Science, 238, 657 – 659, 1987.
Joseph, P. V., Warm pool over the Indian Ocean and monsoon onset, Tro-
pical Ocean and Atmosphere Newsletter, 53, 1 – 5, 1990.
Josey, S. A., E. C. Kent, D. Oakley, and P. K. Taylor, A new global air-sea
heat and momentum flux climatology, Int. WOCE Newsl., 24, 3 – 5, 1996.
Josey, S. A., E. C. Kent, and P. K. Taylor, New insights into the ocean heat
budget closure problem from analysis of the SOC air-sea flux climatol-
ogy, J. Clim., 12, 2856 – 2880, 1999.
Kalnay, E., et al., The NCEP/NCAR 40-year reanalysis project, Bull. Am.
Meteorol. Soc., 77, 437 – 471, 1996.
Kim, J.-W., A generalized bulk model of the oceanic mixed-layer, J. Phys.
Oceanogr., 6, 686 – 695, 1976.
Krishnamurti, T. N., Tropical east-west circulations during northern sum-
mer, J. Atmos. Sci., 28, 1342 – 1347, 1971.
Lee, T., and J. Marotzke, Inferring meridional mass and heat transports of
the Indian Ocean by fitting a general circulation model to climatological
data, J. Geophys. Res., 102, 10,585 – 10,602, 1997.
Le Traon, P. Y., and F. Ogor, ERS-1/2 orbit improvement using TOPEX/
Poseidon: The 2 cm challenge, J. Geophys. Res., 103, 8045 – 8057,
1998.
5 - 13
Le Traon, P. Y., P. Gaspar, F. Bouyssel, and H. Makhmara, Using TOPEX/
Poseidon data to enhance ERS-1 data, J. Atmos. Ocean. Tech., 12, 161 –
170, 1995.
Le Traon, P. Y., F. Nadal, and N. Ducet, An improved mapping method of
multi-satellite altimeter data, J. Atmos. Oceanic Technol., 25, 522 – 534,
1998.
Levitus, S., Meridional Ekman heat fluxes for the world oceans and indi-
vidual ocean basins, J. Phys. Oceanogr., 17, 1484 – 1492, 1987.
Levitus, S., and T. P. Boyer, World Ocean Atlas 1994, vol. 4, Temperature,
NOAA Atlas NESDIS, vol. 4, 129 pp., Natl. Oceanic and Atmos. Admin.,
Silver Spring, Md., 1994.
Levitus, S., R. Burgett, and T. P. Boyer, World Ocean Atlas 1994, vol. 3,
Salinity, NOAA Atlas NESDIS, vol. 3, 111 pp., Natl. Oceanic and Atmos.
Admin., Silver Spring, Md., 1994.
Loschnigg, J., and P. J. Webster, A coupled ocean-atmosphere system of
SST modulation for the Indian Ocean, J. Clim., 13, 3342 – 3360,
2000.
Lukas, R. B., and E. J. Lindström, The mixed layer of the western equator-
ial Pacific Ocean, J. Geophys. Res., 96, 3343 – 3357, 1991.
Martin, J. M., J. D. Burton, and D. Eisma, River Inputs to Ocean Systems,
United Nations Press, New York, 1981.
McCreary, J. P., P. K. Kundu, and R. L. Molinari, A numerical investigation
of the dynamics, thermodynamics and mixed-layer processes in the In-
dian Ocean, Prog. Oceanogr., 31, 181 – 244, 1993.
McCreary, J. P., W. Han, D. Shankar, and S. R. Shetye, Dynamics of the
East India Coastal Current, 2, Numerical Solutions, J. Geophys. Res.,
101, 13,993 – 14,010, 1996.
Mooley, D. A., and J. Shukla, Main features of the westward-moving low
pressure systems which form over the Indian region during the summer
monsoon season and their relation to the monsoon rainfall, Mausam, 40,
137 – 152, 1989.
Oberhuber, J. M., An atlas based on the COADS data set: The budgets of
heat, buoyancy and turbulent kinetic energy at the surface of the global
ocean, Tech. Rep. MPI Rep. 15, 19 pp., Max-Plank-Inst. für Meteorol.,
Bundestrasse, Germany, 1988.
Rao, R. R., K. V. S. Ramam, D. S. Rao, and M. X. Joseph, Surface heat
budget estimates at selected areas of north Indian Ocean during Mon-
soon-77, Mausam, 36, 21 – 32, 1985.
Reynolds, R. W., and T. M. Smith, Improved global sea surface temperature
analysis using optimum interpolation, J. Clim., 7, 929 – 948, 1994.
Sanilkumar, K. V., N. Mohankumar, M. X. Joseph, and R. R. Rao, Genesis
of meteorological disturbances and thermohaline variability of the upper
layers in the head Bay of Bengal during Monsoon Trough Boundary
Layer Experiment (MONTBLEX-90), Deep Sea Res., Part I, 41,
1569 – 1581, 1994.
Sarkar, R. P., A dynamical model of orographic rainfall, Mon. Weather Rev.,
94, 555 – 572, 1966.
Sarkar, R. P., Some modifications in a dynamical model of orographic
rainfall, Mon. Weather Rev., 95, 673 – 684, 1967.
Schott, F., Monsoon response of the Somali current and associated upwel-
ling, Prog. Oceanogr., 12, 357 – 381, 1983.
Shankar, D., J. P. McCreary, W. Han, and S. R. Shetye, Dynamics of the
East India Coastal Current, 1, Analytic solutions forced by interior Ek-
man pumping and local alongshore winds, J. Geophys. Res., 101,
13,975 – 13,991, 1996.
Shenoi, S. S. C., D. Shankar, and S. R. Shetye, On the sea surface tem-
perature high in the Lakshadweep Sea before the onset of the southwest
monsoon, J. Geophys. Res., 104, 15,703 – 15,712, 1999.
Shetye, S. R., A model study of the seasonal cycle of the Arabian Sea
surface temperature, J. Mar. Res., 44, 521 – 542, 1986.
Shetye, S. R., The movement and implications of the Ganges-Brahmaputra
runoff on entering the Bay of Bengal, Curr. Sci., 32 – 38, 1993.
Shetye, S. R., S. S. C. Shenoi, A. D. Gouveia, G. S. Michael, D. Sundar,
and G. Nampoothiri, Wind-driven coastal upwelling along the western
boundary of the Bay of Bengal during the southwest monsoon, Cont.
Shelf Res., 11, 1397 – 1408, 1991.
Shetye, S. R., A. D. Gouveia, D. Shankar, S. S. C. Shenoi, P. N. Vinaya-
chandran, D. Sundar, G. S. Michael, and G. Nampoothiri, Hydrography
and circulation in the western Bay of Bengal during the northeast mon-
soon, J. Geophys. Res., 101, 14,011 – 14,025, 1996.
Sikka, D. R., Some aspects of the life history, structure of movement of
monsoon depressions, inMonsoon Dynamics, PAGEOPH, vol. 115, edited
by T. N. Krishnamurti, pp. 1501 – 1529, Birkhäuser Boston, Cambridge,
Mass., 1977.
Sprintall, J., and M. Tomczak, Evidence of the barrier layer in the surface
layer of the tropics, J. Geophys. Res., 97, 7305 – 7316, 1992.
Sud, Y. C., G. K. Walker, and K.-M. Lau, Mechanisms regulating sea sur-
face temperatures and deep convection in the tropics, Geophys. Res. Lett.,
26, 1019 – 1022, 1999.
Vinayachandran, P. N., S. R. Shetye, D. Sengupta, and S. Gadgil, Forcing
5 - 14 SHENOI ET AL.: HEAT BUDGET OF ARABIAN SEA AND BAY OF BENGAL
mechanisms of the Bay of Bengal circulation, Curr. Sci., 71, 753 – 763,
1996.
Weller, R. A., M. F. Baumgartner, S. A. Josey, A. S. Fischer, and J. C.
Kindle, Atmospheric forcing in the Arabian Sea during 1994 – 1995:
Observations and comparisons with climatology and models, Deep Sea
Res., Part II, 45, 1961 – 1999, 1999.
You, Y., Rain-forced barrier layer of the western equatorial Pacific warm
pool: A case study, J. Geophys. Res., 103, 5361 – 5378, 1998.
Zhang, H. M., and L. D. Talley, Heat and buoyancy budgets and mixing
rates in the upper thermocline of the Indian Ocean and global oceans, J.
Phys. Oceanogr., 28, 1961 – 1978, 1998.

D. Shankar, S. S. C. Shenoi, and S. R. Shetye, Physical Oceanography
Division, National Institute of Oceanography, Dona Paula, Goa 403 004,
India. (shankar@darya.nio.org; shenoi@darya.nio.org; shetye@darya.
nio.org)