Content uploaded by Asha De Vos
Author content
All content in this area was uploaded by Asha De Vos on Oct 30, 2014
Content may be subject to copyright.
Biogeosciences, 11, 5909–5930, 2014
www.biogeosciences.net/11/5909/2014/
doi:10.5194/bg-11-5909-2014
© Author(s) 2014. CC Attribution 3.0 License.
Surface circulation and upwelling patterns around Sri Lanka
A. de Vos1,2,3, C. B. Pattiaratchi1, and E. M. S. Wijeratne1
1School of Civil, Environmental and Mining Engineering & The UWA Oceans Institute, University of Western Australia,
35 Stirling Highway, Crawley, Western Australia 6009, Australia
2The Centre for Ocean Health, The University of California Santa Cruz, CA 95060, USA
3The Sri Lankan Blue Whale Project, 131 W.A.D. Ramanayake Mawatha, Colombo 2, Sri Lanka
Correspondence to: A. de Vos (asha.devos@lincoln.oxon.org)
Received: 15 August 2013 – Published in Biogeosciences Discuss.: 11 September 2013
Revised: 26 August 2014 – Accepted: 5 September 2014 – Published: 30 October 2014
Abstract. Sri Lanka occupies a unique location within the
equatorial belt in the northern Indian Ocean, with the Ara-
bian Sea on its western side and the Bay of Bengal on its
eastern side, and experiences bi-annually reversing monsoon
winds. Aggregations of blue whale (Balaenoptera musculus)
have been observed along the southern coast of Sri Lanka
during the northeast (NE) monsoon, when satellite imagery
indicates lower productivity in the surface waters. This study
explored elements of the dynamics of the surface circulation
and coastal upwelling in the waters around Sri Lanka using
satellite imagery and numerical simulations using the Re-
gional Ocean Modelling System (ROMS). The model was
run for 3 years to examine the seasonal and shorter-term
(∼10 days) variability. The results reproduced correctly the
reversing current system, between the Equator and Sri Lanka,
in response to the changing wind field: the eastward flow-
ing Southwest Monsoon Current (SMC) during the south-
west (SW) monsoon transporting 11.5Sv (mean over 2010–
2012) and the westward flowing Northeast Monsoon Current
(NMC) transporting 9.6Sv during the NE monsoon, respec-
tively. A recirculation feature located to the east of Sri Lanka
during the SW monsoon, the Sri Lanka Dome, is shown to
result from the interaction between the SMC and the island
of Sri Lanka. Along the eastern and western coasts, during
both monsoon periods, flow is southward converging along
the southern coast. During the SW monsoon, the island de-
flects the eastward flowing SMC southward, whilst along the
eastern coast, the southward flow results from the Sri Lanka
Dome recirculation. The major upwelling region, during both
monsoon periods, is located along the southern coast, re-
sulting from southward flow converging along the southern
coast and subsequent divergence associated with the offshore
transport of water. Higher surface chlorophyll concentrations
were observed during the SW monsoon. The location of the
flow convergence and hence the upwelling centre was depen-
dent on the relative strengths of wind-driven flow along the
eastern and western coasts: during the SW (NE) monsoon,
the flow along the western (eastern) coast was stronger, mi-
grating the upwelling centre to the east (west).
1 Introduction
Sri Lanka is situated within the equatorial belt in the northern
Indian Ocean, with the Arabian Sea on its western side and
the Bay of Bengal on its eastern side (Fig. 1). In an oceano-
graphic sense, the location of Sri Lanka is unique, with its
offshore waters transporting water with different properties
through reversing ocean currents driven by monsoon winds.
The northern Indian Ocean is characterised by bi-annually
reversing monsoon winds resulting from the seasonal differ-
ential heating and cooling of the continental land mass and
the ocean. The Southwest (SW) monsoon generally operates
between June and October, and the Northeast (NE) monsoon
operates from December through April (Tomczak and God-
frey, 2003). The transition periods are termed the first inter-
monsoon (May) and the second inter-monsoon (November).
During the SW monsoon, the Southwest Monsoon Current
(SMC) flows from west to east, transporting higher salinity
water from the Arabian Sea, whilst during the NE monsoon,
the currents reverse in direction, with the Northeast Mon-
soon Current (NMC) transporting lower salinity water orig-
inating from the Bay of Bengal from east to west (Schott
and McCreary, 2001). During the SW monsoon, increased
Published by Copernicus Publications on behalf of the European Geosciences Union.
5910 A. de Vos et al.: Surface circulation and upwelling patterns
27
826
827
Fig. 1: Study area showing bathymetry and model domain. Numbers represent tide stations used for 828
model validation. 1. Point Pedro 2. Kayts 3. Delft Island 4. Kalpitiya 5. Chilaw 6. Colombo 7. Galle 829
8. Dondra 9. Kirinda 10. Oluwil 11. Batticaloa 12. Trincomalee. Wind speed and direction data was 830
from the Hambantota Meteorological Station on the southeast coast. 831
Figure 1. Study area showing the bathymetry and model domain.
Numbers represent tide stations used for model validation. 1. Point
Pedro 2. Kayts 3. Delft Island 4. Kalpitiya 5. Chilaw 6. Colombo
7. Galle 8. Dondra 9. Kirinda 10. Oluwil 11. Batticaloa 12. Trin-
comalee. Wind speed and direction data were from the Hambantota
meteorological station on the southeastern coast.
chlorophyll concentrations (>5mgm−3) have been recorded
around Sri Lanka, particularly along the southern coast
(Vinayachandran et al., 2004), which appears to be a major
upwelling region. These elevated chlorophyll concentrations
persist for more than four months and have been attributed
to coastal upwelling, advection by the SMC and open ocean
Ekman pumping (Vinayachandran et al., 2004). During the
SW monsoon, where the winds blow parallel to the coast,
winds are upwelling favourable in terms of Ekman dynamics.
Chlorophyll concentrations during the NE monsoon appear
to be low, but there is evidence of high productivity through
the documented feeding aggregations of blue whales (Bal-
aenoptera musculus) along the southern coast of Sri Lanka
(de Vos et al., 2014). Field observations have shown that ∼6–
10 whales are sighted per day along the southern coast of
Sri Lanka (data from 2009 to 2011). The region also has a
well-developed whale watching tourism industry. At present,
there is a lack of information regarding the environmental
features that influence the distribution of blue whales in the
waters of Sri Lanka. Therefore, the aim of this paper was
to examine the oceanographic features that may influence
the distribution of the blue whales off the southern coast of
Sri Lanka. Due to a paucity of field data, previous research
has focused on the analysis of satellite imagery and coarse-
resolution models designed to simulate basin-scale features.
In this paper we use satellite imagery and a high spatial reso-
lution numerical model (ROMS) with realistic and idealised
forcing to investigate the flow patterns and upwelling mech-
anisms, particularly off the southern coast of Sri Lanka.
The continental shelf around Sri Lanka is narrower,
shallower and steeper than is average for the world
(Wijeyananda, 1997). Its mean width is 20km, and it is nar-
rowest on the southwestern coast, where it is less than 10km
(Shepard, 1963; Swan, 1983; Wijeyananda, 1997). The con-
tinental slope around Sri Lanka is a concave feature that ex-
tends from 100m to 4000m in depth. The continental slope
on the southern and eastern coasts has an inclination of 45◦,
which is one of the steepest recorded globally (Sahini, 1982).
The abyssal plain around the island is 3000–4000m deep
(Swan, 1983).
The seasonal difference in sea surface salinity (>2 ppt)
around Sri Lanka is highly significant compared to other re-
gions (Levitus et al., 1994). Salinity in the Bay of Bengal is
generally lower (<33 ppt), whilst salinities in the Arabian
Sea are higher, with maxima up to 36.5ppt due to high evap-
oration and negligible freshwater input. The Bay of Bengal
receives ∼1500km3yr−1of freshwater through freshwater
run-off, whilst the total freshwater input into the Arabian Sea
is ∼190km3yr−1(Jensen, 2001). Including evaporation and
rain, the Arabian Sea experiences a negative freshwater sup-
ply of about 1myr−1, whereas there is a positive freshwa-
ter supply of about 0.4myr−1to the Bay of Bengal (Jensen,
2001).
The mean sea level pressure (SLP) in the northern Indian
region is at a maximum from December to January and at
a minimum from June to July, with a mean seasonal range
of 5–10hPa (Wijeratne, 2003). There is significant seasonal
variation in sea level in the northeastern Indian Ocean, with
a range in the inner Bay of Bengal of ∼0.80–0.90m, de-
creasing to the south (Wijeratne, 2003); hence, the mean sea
level is 0.05m lower in January compared to July, due to the
inverse barometric effect. The seasonal sea level variability
around Sri Lankan waters is around 0.2–0.3m, with max-
ima during June through the action of the SW monsoon (Wi-
jeratne et al., 2008). The tides around the island are mixed
semidiurnal with a maximum spring tidal range of ∼0.70m.
The surface circulation of the northern Indian Ocean may
be described after Schott and McCreary (2001). A schematic
of the circulation in the northern Indian Ocean in the vicinity
of Sri Lanka during the SW monsoon is shown in Fig. 2b.
Along India and Sri Lanka, the eastern boundary current,
or West Indian Coastal Current (WICC) in the Arabian
Sea, flows southwards along the western Indian coastline
and joins the eastward flowing Southwest Monsoon Cur-
rent (SMC). Shankar et al. (2002) also postulated a westerly
flow from the southern–central Arabian Sea entraining wa-
ter into the SMC. The presence of the anti-clockwise Lak-
shadweep eddy off the southwestern coast of India modi-
fies the current flow in this region. The SMC flows along
the southern coast of Sri Lanka from west to east (Schott et
al., 1994), transporting ∼8Sv (1Sv=106m3s−1) between
the Equator and Sri Lanka. After passing the coast of Sri
Lanka, the currents form an anti-clockwise eddy defined as
the Sri Lanka Dome (SLD) centered around 83◦E and 7◦N
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5911
28
832
833
Fig. 2: Circulation patterns around Sri Lanka and southern India for (a) Northeast monsoon and (b) 834
Southwest monsoon. WICC – West Indian Coastal Current; EICC – East Indian Coastal Current; 835
SMC – South Monsoon Current; NMC – North Monsoon Current; SD– Sri Lanka Dome. 836
837
Figure 2. Circulation patterns around Sri Lanka and southern In-
dia for the (a) Northeast monsoon and the (b) Southwest monsoon.
WICC – West Indian Coastal Current; EICC – East Indian Coastal
Current; SMC – South Monsoon Current; NMC – North Monsoon
Current; SD – Sri Lanka Dome.
(Vinayachandran and Yamagata, 1998). The western arm of
this eddy drives a southward current along the eastern coast
of Sri Lanka, whilst the remainder flows northward along
the eastern Indian coast as the East Indian Coastal Current
(EICC).
During the NE monsoon, the currents reverse direction
(Fig. 2a). Along the eastern Indian coast, the EICC flows
southward past Sri Lanka and joins the Northeast Monsoon
Current (NMC) flowing from east to west, transporting about
12Sv (Schott et al., 1994). The currents then flow around
the clockwise Lakshadweep eddy and northward along the
western Indian coastline as the West Indian Coastal Current
(WICC).
One of main features to note from this description from
the perspective of Sri Lanka is the reversal of currents
along the western and southern coasts and the north-to-south
flow along the eastern coast. This circulation pattern was
confirmed by Shankar et al. (2002). However, Varkey et
al. (1996) and Shankar and Shetye (1997) both provide a
different interpretation, and suggest that currents along the
eastern coast of Sri Lanka flow south to north irrespective of
season. However, using altimeter data, Durand et al. (2009)
have shown a seasonal reversal of the currents along the east-
ern coast of Sri Lanka.
Sri Lanka is a relatively large island (length 440km; width
225km), extending offshore into the Indian Ocean, similar
to a headland. This allows the island to interact with the
seasonally reversing monsoon. Many studies have reported
the influence of flow interaction with islands and headlands
leading to enhanced primary production – termed the island
mass effect (IME) by Doty and Oguri (1956). These stud-
ies have included different spatial scales using laboratory
and field experiments to understand circulation and enhanced
productivity. They include those in the vicinity of oceanic
islands: Johnston Atoll (Barkley, 1972), Aldabra and Cos-
moledo atolls (Heywood et al., 1990), Barbados (Bowman
et al., 1996; Cowen and Castro, 1994), the Canary Islands
(Barton et al., 2000), the Kerguelen Islands (Bucciarelli et
al., 2001), Madeira (Caldeira et al., 2002), the Galapagos Is-
lands (Palacios, 2002), Hawaii (Hafner and Xie, 2003), Santa
Catalina (Dong and McWilliams, 2007); and, in continen-
tal shelf and coastal regions, Wolanski et al. (1984), Pat-
tiaratchi et al. (1987) and Alaee et al. (2007). Many scal-
ing arguments have been proposed to define the circulation
patterns in the lee of islands based on the Reynolds num-
ber which appears to reproduce the observed circulation in
the lee of the island/headland (Tomczak, 1988; Wolanski et
al., 1984). The Reynolds number for the deep ocean is de-
fined as (Tomczak, 1988): Re=U L / Kh, where Uis the ve-
locity scale, La length scale, and Khthe horizontal eddy
viscosity. The nature of the wake downstream of an island
can be predicted using the Reynolds number. For low val-
ues of Re(∼1), there is no perceptible wake with the flow
attached to the island (the “attached” flow condition; Alaee
et al., 2007). For Rebetween 1 and 40, the wake consists
of two attached eddies. At higher values of Re, the wake
becomes increasingly unstable, and counter-rotating eddies
form a vortex street (Tomczak, 1988). Flow past a curved
coastline can also lead to secondary circulation: here, as a
result of the curvature-induced centrifugal acceleration, the
surface waters move offshore and are replaced by water from
the sub-surface (Alaee et al., 2004).
Alaee et al. (2004) examined the secondary circulation in-
duced by both the flow curvature and the Coriolis effect, for
quasi-steady oceanic flows. Using scaling of the transverse
momentum equation, Alaee et al. (2004) developed a flow
regime diagram to predict the strength of the secondary flow
Unfor different flow regimes and also to provide informa-
tion on the relative importance of the flow curvature and the
Coriolis effect in the generation of the secondary flow.
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5912 A. de Vos et al.: Surface circulation and upwelling patterns
The upwelling off the southern coast of Sri Lanka usu-
ally appears and intensifies during the summer months, when
the SW monsoon prevails, and is said to be due to a com-
bination of wind-driven Ekman transport, advection by the
SMC and open ocean Ekman pumping (McCreary Jr. et al.,
2009; Vinayachandran et al., 1999, 2004). Monthly satellite
image composites of chlorophyll analysed by Yapa (2009)
show high-productivity waters with mean chlorophyll con-
centrations of more than 5mgm−3along the southern and
western regions during the months of June to August that are
accompanied by a 2◦to 3◦C decrease in sea surface tem-
perature (SST) corresponding to regions where high chloro-
phyll a concentrations are detected. To illustrate this rela-
tionship, MODIS images indicate the strong relationship be-
tween higher chlorophyll and cooler SSTs (Fig. 3). Data col-
lected during the Dr. Fridtjof Nansen cruises between 1978
and 1989 provide evidence that the SW monsoon bloom re-
sults from upwelling that begins closer to the coast, and
progresses further offshore as it develops over subsequent
months (Saetersdal et al., 1999). Michisaki et al. (1996) con-
firmed high primary productivity when they recorded maxi-
mum nitrate concentrations of approximately 10µM in mid-
June, accompanied by maximum chlorophyll concentrations
of 0.9mgm−3off the western coast of Sri Lanka.
The aim of this paper is to define the seasonal changes in
circulation and upwelling patterns around Sri Lanka using a
high-resolution numerical model (ROMS) including realistic
forcing complemented by satellite imagery. The motivation
for the paper is the observation of blue whale (Balaenoptera
musculus) feeding aggregations off the southern coast of Sri
Lanka during the NE monsoon period (de Vos et al., 2014),
despite satellite imagery indicating lower productivity in the
surface waters. This paper is organised as follows: in Sect. 2,
we describe the numerical model configuration and valida-
tion, Sect. 3 presents the results from analysis of the wind
fields, satellite imagery and numerical model output includ-
ing idealised simulations to examine upwelling generation
mechanisms, and the results are discussed in Sect. 4, with
overall conclusions given in Sect. 5.
2 Methodology
The main approach for the study is the use of a numerical
model to identify the mean circulation patterns and upwelling
around Sri Lanka. There is a lack of field data from this
region, and some of the available public domain data have
been accessed and presented in this paper. The data include
wind speed and direction data from a coastal meteorolog-
ical station located at Hambantota (Fig. 1), meteorological
information from ECMWF ERA interim data which were
also used for model forcing, and MODIS satellite imagery
(ocean colour and SST) accessed from the ocean colour web-
site (Feldman and McClain, 2013).
2.1 ROMS configuration and validation
The Regional Ocean Modelling System (ROMS) is a three-
dimensional numerical ocean model based on the non-
linear terrain following coordinate system of Song and
Haidvogel (1994). ROMS solves the incompressible, hydro-
static, primitive equations with a free sea surface, horizontal
curvilinear coordinates, and a generalised terrain-following
s-vertical coordinate that can be configured to enhance reso-
lution at the sea surface or seafloor (Haidvogel et al., 2008).
The model formulation and numerical algorithms are de-
scribed in detail in Shchepetkin and McWilliams (2005), and
have been used to simulate the circulation and upwelling pro-
cesses in a range of ocean basins (e.g. Di Lorenzo et al.,
2007; Dong et al., 2009; Haidvogel et al., 2008; Marchesiello
et al., 2003; Xu et al., 2013).
The model grid (Fig. 1) configured for this study included
the continental shelf and slope waters surrounding Sri Lanka
as well as the deeper ocean, and consisted of a horizontal
grid with resolution less than 2km, with 30 vertical lay-
ers in a terrain-following s-coordinate system. The minimum
model depth was set to −15m, i.e. coastal regions shallower
than 15m were set to 15m. The model was driven by di-
rect air–sea heat and freshwater fluxes, momentum fluxes, in-
verted barometric effects, tide/sea levels, transport and trac-
ers at open boundaries. The forcing data were interpolated
onto the corresponding model grid points to create initial
and forcing files. The model was driven with 3-hourly atmo-
spheric forcing and daily surface heat and freshwater fluxes
using ECMWF ERA interim data. The heat and freshwater
fluxes were also specified using ECMWF ERA data. The net
heat flux at the air–sea interface was estimated based on the
balance of incoming solar radiation, outgoing long waves,
and sensible and latent heat fluxes, respectively. Freshwa-
ter fluxes were estimated using precipitation and evaporation
data from ECMWF ERA data, and the river inputs were ig-
nored. HYCOM global ocean model (Bleck, 2002) daily out-
puts of salinity, temperature, and horizontal velocities were
used to specify the open boundary section 3-D tracers and
transport. Open boundary barotropic velocities were esti-
mated by vertically averaging the eastward (u) and northward
(v) component data, which were interpolated at the bound-
ary sections. At open boundaries, ROMS offers a wide array
of conditions. We used a combination of nudging and radi-
ation conditions for 3-D transport and tracers at the model
open boundaries. The model forcing tides were derived from
the TPX07.2 global tidal model and monthly climatologi-
cal mean sea levels derived from the AVISO database. The
tides were provided as complex amplitudes of earth-relative
sea-surface elevation and tidal currents for eight primary
harmonic constituents (M2,S2,N2,K2,K1,O1,P1,Q1).
These harmonics were introduced in ROMS through the open
boundaries elevation using the Chapman and current ellipse
variables using the Flather condition (see Marchesiello et al.,
2001). Model hindcast simulations were undertaken to obtain
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5913
29
Fig. 3: (a) Surface Chlorophyll concentration (SCC) obtained on 12 December 2010; (b) SCC on
21 January 2011; (c) SCC on 19 January 2013; (d) SCC on 8 August 2011; (e) Sea Surface
Temperature (SST) on 19 October 2003; (f) SCC on 19 October 2003.
Figure 3.
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5914 A. de Vos et al.: Surface circulation and upwelling patterns
30
Fig. 3: cont: (g) SST on 1 August 2012; (h) SCC on SST on 1 August 2012; (i) SST on 19 June 838
2013; (j) SCC on 19 June 2013; (k) SST on 30 Sep 2013; (l) SCC on 30 Sep 2013. The chlorophyll 839
and temperature scales are on (k) and (l) respectively and apply to all images. 840
841
Figure 3. (a) Surface chlorophyll concentration (SCC) obtained on 12 December 2010, (b) SCC on 21 January 2011, (c) SCC on 19 Ja-
nuary 2013, (d) SCC on 8 August 2011; (e) sea surface temperature (SST) on 19 October 2003, (f) SCC on 19 October 2003. (g) SST on
1 August 2012, (h) SCC on SST on 1 August 2012, (i) SST on 19 June 2013, (j) SCC on 19 June 2013, (k) SST on 30 September 2013,
(l) SCC on 30 September 2013. The chlorophyll and temperature scales are in (k) and (l) respectively and apply to all images.
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5915
31
842
843
844
Fig. 4: Comparison of the shipboard ADCP measured (red) and ROMS (black) currents at 30 m 845
depth. a) ADCP from 25 to 31 Dec 1997 and ROMS for 31 Dec 2010, b) ADCP from 01 to 05 846
March 1995 and ROMS for 03 March 2011, c) ADCP from 21 to 30 July 1993 and ROMS for 26 847
July 2011 and d) ADCP from 01 to 03 Aug 2005 and ROMS for 03 Aug 2011. 848
849
Figure 4. Comparison of the shipboard ADCP measured (red) and
ROMS (black) currents at 30m depth. (a) ADCP from 25 to 31 De-
cember 1997 and ROMS for 31 December 2010, (b) ADCP from 1
to 5 March 1995 and ROMS for 3 March 2011, (c) ADCP from 21
to 30 July 1993 and ROMS for 26 July 2011 and (d) ADCP from 1
to 3 August 2005 and ROMS for 3 August 2011.
optimal model results. Three-dimensional variables (salinity,
temperature and velocity components) were output at daily
intervals, with sea surface heights at hourly intervals.
2.2 Experimental setup
In addition to realistic simulations to examine the seasonal
circulation patterns and upwelling, numerical experiments
were also designed to address the following: (1) the role
of the land-mass effect contribution to upwelling around Sri
Lanka, (2) variability in the upwelling centre in response to
the magnitude and direction of winds along the western and
eastern sides of the island, and (3) mechanisms for the forma-
tion of the Sri Lanka Dome located to the east of Sri Lanka.
In order to address (1), model simulations were undertaken,
including and excluding the Coriolis term, whilst model runs
with synthetic wind fields were undertaken to examine (2),
with different wind stresses on the western and eastern sides
of Sri Lanka. Mechanisms for the formation of the Sri Lanka
Dome (3) were undertaken by forcing the model with con-
stant westerly winds of different magnitudes (2, 4, 6 and
8ms−1). A simulation was also undertaken to investigate the
flow patterns in the absence of Sri Lanka. Additional model
runs (not presented here) were also undertaken to investigate
whether the tides played a role in the upwelling process.
2.3 Model validation
Model hindcasts were undertaken over a 3-year period (2010,
2011 and 2012) using realistic surface and boundary forc-
ing (Sect. 2.1). The first year (2010) was considered as
spin up, and the results presented here are from the second
year (2011) of simulations, although results from the third
year of the simulations are also presented (Table 2). It should
be noted that 2011 was a strong La Niña year. However, com-
parison of the seasonal winds observed in 2011 with the pre-
vious 4 years indicated that the wind field in 2011 is not sig-
nificantly different from other years. In the absence of de-
tailed field measurements from the region, predicted surface
currents and temperature distributions were compared with
available data as well as with sea level data.
2.3.1 Tide and mean sea level
The predicted hourly sea levels at each grid point were
subjected to harmonic analysis using the T-Tide MATLAB
toolbox (Pawlowicz et al., 2002). To visualise and interpret
model results obtained around Sri Lanka, co-tidal charts for
the main tidal constituents, M2,S2,K1, and O1, were pro-
duced (not shown). The predicted amplitudes and phases
from the simulation are in close agreement with measured
data for four major tidal constituents (Table 1) and those of
Wijeratne (2003).
2.3.2 Large-scale circulation
Shipboard ADCP current measurements for the region are
available from the World Ocean Circulation Experiments
(WOCE). However, it is important to note that there were
no observations during the model simulation period (2010
and 2011). We compared the model results and observa-
tions based on the time of the year, as shown in Fig. 4. It
is clear that there is good qualitative agreement between the
predicted and observed currents throughout the ship tracks.
The model results also reproduce some of the observed cir-
culation features. For example, seasonal reversal of currents
along the southern coast during the two monsoon periods is
reproduced: during the NE monsoon, the currents flow to-
wards the west (Fig. 4a, b), whilst during the SW monsoon,
they flow to the east (Fig. 4c, d). The reversing current pat-
tern to the east of Sri Lanka during the NE monsoon with
southward currents close to the coast and northward currents
further offshore is also reproduced (Fig. 4a). The model also
reproduced fine-scale features that were represented in the
ADCP transect such as the transition from westward to east-
ward currents closer to the coast (Fig. 4c).
2.3.3 Satellite imagery
Suspended material (such as sediment, chlorophyll, etc.) in
the surface waters may be used as a passive tracer to fol-
low flow patterns using satellite imagery (Pattiaratchi et
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5916 A. de Vos et al.: Surface circulation and upwelling patterns
Table 1. Tidal constituents at different stations along the coastline of Sri Lanka. Tide gauge data are denoted in regular font and model data
are denoted in italic font. Phase refers to local time.
Station M2S2K1O1
a (m) goa (m) goa (m) goa (m) go
Trincomalee 0.18 0.19 238 252 0.06 0.08 268 276 0.07 0.06 332 338 0.02 0.01 304 343
Batticaloa 0.14 0.14 235 227 0.07 0.06 265 279 0.05 0.04 330 310 0.02 0.02 300 267
Oluwil 0.07 0.05 230 248 0.03 0.02 260 245 0.05 0.06 332 324 0.01 0.01 280 253
Kirinda 0.07 0.06 92 88 0.056 0.06 130 112 0.031 0.03 29 358 0.01 0.01 17 10
Galle 0.16 0.17 56 48 0.11 0.09 99 112 0.05 0.06 21 69 0.01 0.03 73 112
Colombo 0.18 0.18 45 48 0.12 0.12 93 115 0.07 0.06 32 48 0.03 0.03 58 76
Chilaw 0.18 0.19 045 0.57 0.11 0.12 092 114 0.09 0.08 43 55 0.03 0.04 058 86
Delft Island 0.11 0.11 38 42 0.04 0.035 77 84 0.11 0.08 76 57 0.05 0.06 006 24
Kayts 0.03 0.02 63 78 0.01 0.008 96 126 0.12 0.09 61 145 0.02 0.03 77 123
Point Pedro 0.16 0.18 242 256 0.09 0.08 270 245 0.05 0.06 328 277 0.01 0.03 003 347
Figure 5. Typical summer upwelling frontal features around
the southern part of Sri Lanka obtained on 12 October 2003:
(a) satellite-derived surface chlorophyll concentration, and (b) pre-
dicted near-surface current vectors and temperature. Note that the
colour scales are given in Fig. 3.
al., 1987). In regions of upwelling (for example, see Fig. 3),
there is also a correspondence between regions of higher sur-
face chlorophyll concentrations (SCC) and lower sea sur-
face temperatures (SST). Thus, ocean colour imagery may be
used qualitatively to validate numerical model outputs. Com-
parison between model-predicted SST and satellite-derived
SCC indicates that the model reproduced observed patterns,
particularly the higher chlorophyll “tongue” feature, and
sharp fronts (Fig. 5).
3 Results
3.1 The wind field
The monsoon and inter-monsoon periods occur at similar
times during the year. However, there is an inter-annual vari-
ability in the onset of these climatic events, and thus the tim-
ing of each monsoon can vary by up to 1–2 months. Wind
data recorded in 2010 from a coastal meteorology station lo-
cated along the southeastern coast of Sri Lanka (Hambantota,
Fig. 1) reflect changes in the wind field, in accordance with
the monsoons (Fig. 6): winds blew from between the north
and east (0–90◦) from December to April, whilst the winds
were predominantly from the southwest and west (225–270◦)
between April and November (Fig. 6b). Wind speeds were
∼8ms−1between mid-January and mid-March, correspond-
ing to the peak of the NE monsoon: less than 6ms−1be-
tween mid-March and mid-May (waning NE monsoon and
first inter-monsoon), increasing to more than 6ms−1from
June until October, reflecting the SW monsoon, and decreas-
ing to less than 6ms−1during the second inter-monsoon pe-
riod in mid-November.
In addition to the temporal changes in the wind field,
there is also significant spatial distribution, as revealed by
the ECMWF ERA interim data (Fig. 7). One of the factors
influencing the spatial wind field is the local land topography
of Sri Lanka and southern India. Coastal regions around Sri
Lanka are relatively flat and surround the elevated central re-
gion that increases to a maximum elevation of 2500m. Simi-
larly, southern India consists of elevated terrain that exceeds
1000m (Luis and Kawamura, 2000). During the NE mon-
soon (Fig. 7a, f), winds are predominantly from the northeast
across the study region, with stronger winds in the Gulf of
Mannar (Fig. 1) as a result of the local land topography. Here,
the northeasterly winds are funneled through the elevated to-
pography between southern India and Sri Lanka, resulting
in strong winds over the Gulf of Mannar (Luis and Kawa-
mura, 2000). Off the southern coast of Sri Lanka, the winds
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5917
33
855
Fig. 6: Time series for wind speed and direction (0 winds from the north) for Sri Lanka in 2010. 856
The red line and dots indicate daily data collected at 0830 hrs and 1730 hrs and black line and dots 857
indicate the daily averaged data. Data were obtained from the Hambantota meteorological station, 858
southeast Sri Lanka. 859
860
861
862
863
864
Figure 6. Time series for wind speed and direction (0 winds from the north) for Sri Lanka in 2010. The red line and dots indicate daily
data collected at 08:30 and 17:30LT and the black line and dots indicate the daily averaged data. Data were obtained from the Hambantota
meteorological station, southeastern Sri Lanka.
are weaker and are mainly offshore during the NE mon-
soon (Fig. 7a, f). During the first inter-monsoon, the east-
ern coast of Sri Lanka experiences onshore winds (easterly)
with northeasterly winds along the western coast, and winds
off the southern coast remaining offshore (Fig. 7b). Along
the western and southern coasts of Sri Lanka, during the SW
monsoon, the winds are westerly (Fig. 7c, d and e) and, due
perhaps to the local topography, they veer northwards off the
eastern side of the island (southwesterly winds). As such,
both the temporal and spatial wind fields influence the ocean
circulation patterns around the island.
3.2 Seasonal circulation
3.2.1 Satellite imagery
The seasonal circulation around Sri Lanka was examined
through the use of surface chlorophyll concentration (SCC)
climatology data (resolution of 4km from Feldman and Mc-
Clain, 2013) as a passive tracer and to understand seasonal
variability in surface chlorophyll concentrations.
In January, the Northeast Monsoon Current (NMC) flows
from east to west (Fig. 8a). This is reflected in the SCC
data with slightly higher concentrations to the west of Sri
Lanka. However, the more pronounced feature is the “stir-
ring” caused by the NMC flowing from east to west past the
Maldives island chain, with enhanced SCC to the west of the
island chain. During this period, the monsoon drift is shal-
low and will generally only have a minimal effect on the wa-
ters below the thermocline (Wyrtiki, 1973). In March, during
the monsoon transition period, SCC decreased to less than
0.20mgm−3(Fig. 8b) in the whole study region. There is
an absence of a “concentration wake” in the vicinity of the
Maldive Islands, indicating weak currents lacking unidirec-
tionality in this region. Similar conditions were observed in
April (not shown). In May, during the onset of the SW mon-
soon (Fig. 7c), a band of high SCC (∼2.5mgm−3) water was
present along the southern coast of Sri Lanka (Fig. 8c) and
also in the Gulf of Mannar. SCC levels along the southern
coast were 10 times higher than they were in April, but low
concentrations were present to the east of Sri Lanka. In July,
enhanced SCC to the east of the Maldive islands and the
plume of elevated SCC to the southeast of Sri Lanka con-
firmed the eastward flow of the Southwest Monsoon Current
(Fig. 8). In June (not shown), the high SCC patch off south-
ern India begins to extend to the east across the entrance
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5918 A. de Vos et al.: Surface circulation and upwelling patterns
34
865
Fig. 7: Bi-monthly wind speed and direction for Sri Lanka in 2011 from ECMWF ERA interim 866
data. Each plot represents a bi-monthly average with length of arrow correlating with speed. (a) 867
January (northeast monsoon); (b) March (first inter-monsoon); (c) May, (d) July and (e) September 868
(southwest monsoon); (f) November (second inter-monsoon). 869
870
Figure 7. Bi-monthly wind speed and direction for Sri Lanka in 2011 from ECMWF ERA interim data. Each plot represents a bi-monthly
average, with the length of the arrow correlating with speed. (a) January (northeast monsoon), (b) March (first inter-monsoon), (c) May, (d)
July, (e) September (southwest monsoon) and (f) November (second inter-monsoon).
to the Gulf of Mannar, whilst the surrounding areas expe-
rienced decreased SCC. In July, the high SCC plume gener-
ated by the SW monsoon current flowing past the Maldive
Islands merged with the high SCC patch off southern In-
dia and the higher SCC waters off the western coast of Sri
Lanka (Fig. 8d). The SCC is ∼5mgm−3along the western
and southern coasts of Sri Lanka. A plume of higher SCC wa-
ter originating from the southern coast of Sri Lanka extended
to the east and shows evidence of an eddy – most likely the
Sri Lankan Dome (Fig. 2). There is also a band of lower SCC
water adjacent to the eastern coast of Sri Lanka, which is due
to the southward flow of water along this coast at this time
of year. In September, the SCC patterns were similar to that
in July (Fig. 8e), except that the maximum SCC was lower
in the range of 0.20 = −0.40mgm−3and extended over a
larger area, particularly to the south and east of Sri Lanka.
In November, the SCC levels decreased almost to those ob-
served in January, the difference being that the plume from
the Maldive Islands was present to the east, indicating that
the SMC was still flowing eastwards (Fig. 8f).
In general, chlorophyll a concentrations around Sri Lanka
were relatively lower during the NE monsoon compared
to the SW monsoon (Kabanova, 1968). This seasonality is
maintained year to year, but with inter-annual variability
(Fig. 9). A Hovmöller diagram of monthly mean SSC be-
tween the southern coast of Sri Lanka (6◦N) and the Equator
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5919
35
Fig. 8: Climatological monthly mean surface chlorophyll concentrations around Sri Lanka: (a) 871
January; (b) March; (c) May, (d) July and (e) September; (f) November. The white line in March 872
represents the location of the data extracted for the Hovmöller diagram in Fig. 9. 873
874
January March
May July
September November
Figure 8. Climatological monthly mean surface chlorophyll concentrations around Sri Lanka: (a) January, (b) March, (c) May, (d) July,
(e) September and (f) November. The white line in March represents the location of the data extracted for the Hovmöller diagram in Fig. 9.
indicates higher values closest to the Sri Lankan coast ex-
tending ∼270 km offshore on average. In 2002 and 2006, the
influence of this upwelling can be observed extending to the
Equator. Although interannual variability is not within the
scope of this paper, it is interesting that 2002 and 2006 reflect
El Niño and positive Indian Ocean dipole years (Sreenivas et
al., 2012).
3.2.2 Numerical modelling
Numerical model results reproduce the general patterns iden-
tified in previous studies (Fig. 2) and from ocean colour im-
agery (Fig. 8). The seasonal mean currents show significant
spatial variability due to the spatial and temporal changes in
the wind climate (Fig. 10). This is evident when comparing
the mean currents during the NE monsoon (Fig. 10a) and the
instantaneous currents at the end of December (Fig. 4a, b).
The reversing currents to the south of Sri Lanka, easterly dur-
ing the SW monsoon and westerly during the NE monsoon,
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5920 A. de Vos et al.: Surface circulation and upwelling patterns
36
875
876
Fig. 9: Hovmöller diagram displaying seasonality and inter-annual variability of surface chlorophyll 877
concentrations off the southern coast of Sri Lanka. The location of transect is shown on Figure 8. 878
879
Figure 9. Hovmöller diagram displaying seasonality and interannual variability of surface chlorophyll concentrations off the southern coast
of Sri Lanka. The location of the transect is shown in Fig. 8.
Table 2. Volume transport of water between Sri Lanka and the
Equator along 80.5◦E for the three simulation years.
Simulation Southwest Monsoon Northeast Monsoon
year Current (SMC) Current (NMC)
2010 10.25Sv 8.27Sv
2011 12.78Sv 10.69Sv
2012 11.36Sv 9.87Sv
Mean 11.46Sv 9.61Sv
are reproduced by the model. The currents during the SW
monsoon are stronger than those during the other seasons,
reflecting the stronger winds during this period (Fig. 10c).
Schott and McCreary (2001) estimated that off southern
Sri Lanka (north of the Equator), transport rates resulting
from the SMC and NMC were 8Sv and 12Sv respectively,
with SMC transport rates lower than those for the NMC.
The numerical model output indicates transport rates rang-
ing from 10.25 to 12.78Sv (mean=11.5Sv) and from 8.27
to 10.69Sv (mean=9.6Sv) for the SMC and NMC respec-
tively (Table 2). The estimate for the NMC transport is simi-
lar to that reported by Schott and McCreary (2001), but SMC
transport is higher, as expected due to the stronger winds ex-
perienced during this period.
During the NE monsoon, currents along the eastern coast
of Sri Lanka flow southwards closer to the coast and north-
wards further offshore, separated by a shear zone (Fig. 4a,
b). The presence of the shear zone is confirmed by a
shipborne ADCP transect (Fig. 4a). The currents closer to
the shore follow the coastline, flowing to the west along
the southern coast and northward along the western coast
(Fig. 10a; Fig. 4a, b). Currents in the Gulf of Mannar flow
towards the southwest and mirror the direction of the wind
(Fig. 10a). During the first inter-monsoon, current speeds de-
crease (Fig. 10b); currents along both the eastern and west-
ern coasts were converging off southern Sri Lanka (south of
∼6.5◦N). The presence of the Sri Lanka Dome centered at
84◦E and 8◦N can be identified. Strong northward currents
along the northeastern coast extending along the southern In-
dian coastline are predicted. This flow pattern is similar to
that shown on satellite images by Legeckis (1987) postulat-
ing a western boundary current in the Bay of Bengal.
Under SW monsoon conditions, currents are higher across
the entire region, particularly along the southern and south-
eastern coasts of Sri Lanka (Fig. 10c). As a result, the Sri
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5921
37
880
881
Fig. 10: Seasonal surface circulation from ROMS: (a) January (northeast monsoon); (b) March (first 882
inter-monsoon); (c) May, (d) July and (e) September (southwest monsoon); (f) November (second 883
inter-monsoon). 884
885
Figure 10. Seasonal surface circulation from ROMS: (a) January
(northeast monsoon), (b) March (first inter-monsoon), (c) May,
(d) July, (e) September (southwest monsoon) and (f) November
(second inter-monsoon).
Lanka Dome shifts to the north – now centered at 84◦E
and 9.5◦N. Southward flowing water along the eastern coast
converges with water eastward of the SMC. There is also
a stronger band of currents flowing past the southern tip of
India and west of Sri Lanka (Fig. 10c), which explains the
merging of the SCC between the southern regions of the In-
dian coast and Sri Lanka (Fig. 8d). The weakest currents are
predicted during the second inter-monsoon, with no evidence
of the Sri Lanka Dome (Fig. 10d).
Flow patterns such as those described in Fig. 10 provide
no indication of regions and periods of coastal upwelling
around Sri Lanka. Therefore, the model-predicted sea surface
temperature (SST) and flow fields were examined at shorter
timescales, with the assumption that cooler waters (compared
to the surrounding water) represented upwelling. Analysis of
model output revealed that upwelling occurs on a seasonal
basis and/or during shorter period sporadic events along dif-
ferent parts of the coastline. This is highlighted in Fig. 11,
with the corresponding wind fields shown in Fig. 12. During
the NE monsoon, cooler SSTs were observed along the west-
ern and southern coasts, with warmer water along the eastern
coast of Sri Lanka (Fig. 11a). The latter is due to the down-
welling regime in this region, with onshore winds (Fig. 12a)
38
886
Fig. 11: Predicted near-surface current vectors plotted on the surface temperature showing major 887
upwelling regions around Sri Lanka: (a) 16-20 January 2011; (b) 10-14 April 2011; (c) 12-16 July 888
2011 and (d) 18-22 August 2011. 889
890
891
Figure 11. Predicted near-surface current vectors plotted on the
surface temperature showing major upwelling regions around Sri
Lanka: (a) 16–20 January 2011, (b) 10–14 April 2011, (c) 12–
16 July 2011 and (d) 18–22 August 2011.
during the NE monsoon reflected in a band of narrow warm
southward moving water. Colder waters were found in re-
gions of divergence in the flow field, where there was mainly
offshore transport of water (Fig. 11a), reflecting the possibil-
ity that perhaps processes other than wind-driven upwelling
may be responsible for the upwelling. There was negligible
colder surface water present during the first inter-monsoon
period, except perhaps along the extreme north of Sri Lanka
(Fig. 11b). The southern coastal regions of both India and Sri
Lanka experienced colder SST throughout the SW monsoon,
indicating strong upwelling during this period (Figs. 11c, d).
There was also advection of colder water from the southern
tip of India to the western coast of Sri Lanka during the SW
monsoon (Fig. 11c, d). The most notable feature is the shape
of the cold water regions to the south and southeast of Sri
Lanka (Fig. 11c, d, respectively). This shape is clearly vis-
ible on satellite images as a result of the associated higher
SCC (Figs. 2 and 5), and occurs in regions of convergence: in
July 2011 (Figs. 3 and 11c), water flowing southwards along
both the eastern and western coasts converges to the south
and is transported offshore, resulting in a colder water patch
near the coast. In August, this colder water patch migrates to
the east, and is present off the southeastern coast of Sri Lanka
(Fig. 11d). This feature is very similar to that observed on the
August 2012 satellite image (Fig. 3c, d). These features indi-
cate that wind-driven upwelling through Ekman dynamics is
most likely not responsible for upwelling along the southern
coast of Sri Lanka.
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5922 A. de Vos et al.: Surface circulation and upwelling patterns
39
892 893
Fig. 12: Wind fields corresponding to the predicted currents shown on Fig. 11. 894
895
Figure 12. Wind fields corresponding to the predicted currents
shown in Fig. 11.
Time series of volume transport along a transect to the
south of Sri Lanka and SST at three locations over one
year (2012) indicated the seasonal variability (Fig. 13). Be-
tween January and April (NE monsoon and the first inter-
monsoon), the volume transport was to the west, with rates
up to 0.08Sv and intermittent reversal of the currents to the
west (Fig. 13b). The flow was generally towards the east be-
tween May and October (SW monsoon and the second inter-
monsoon) before reversing towards the west during Novem-
ber and December (Fig. 13b). During the SW monsoon, the
transport increased from June to the end of July, when it
reached a maximum of 0.23Sv and then generally decreased
in strength (Fig. 13b). The monsoon pattern was clearly re-
flected in the SST time series (Fig. 13c), with relatively lower
SST in January/February (NE monsoon) and higher SST be-
tween March and April (first inter-monsoon), with the tem-
peratures decreasing in May and then being relatively low
between June and September (SW monsoon), and higher val-
ues between October and November (second inter-monsoon)
and then decreasing during December (NE monsoon). Al-
though the SST at all three locations was very similar during
the inter-monsoon period, with relatively higher values, there
were differences during the monsoon period due to upwelling
(cooler waters). In January/February (NE monsoon), the wa-
ters to the southwest were relatively cooler when compared
to the southeast, whilst between June and September (SW
monsoon), the waters in the southeast were cooler (Fig. 13c).
This shift in the colder water may indicate a migration of the
upwelling centre from the southwest to southeast between
the two monsoon periods.
3.3 Temporal (10-day) variability
In order to assess the shorter period spatial variability of sur-
face circulation and upwelling around Sri Lanka, model out-
puts for surface currents and temperature averaged over a
10-day period were examined. Initially, during the NE mon-
soon (January 2011), southward currents flowed along both
the eastern and western coasts of Sri Lanka, with easterly
currents along the southern coast (Fig. 14a). The currents ap-
pear to converge along the southeastern corner, as indicated
by the presence of colder water. Over the next 10 days, the
currents along the eastern coast increased due to stronger
winds, and this was accompanied by a reversal in the cur-
rents along the southern coast, which flow eastwards, causing
the convergence zone (and colder water due to upwelling) to
shift towards the southeast (Fig. 14b). During the subsequent
10-day period, there is colder water along the entire west-
ern coast of Sri Lanka including the Gulf of Mannar, due
to upwelling and a contribution through cooling due to air–
sea fluxes (e.g. Luis and Kawamura, 2000). Analysis of scat-
terometer (NSCAT) winds by Luis and Kawamura (2000)
indicated a 15-day periodicity in the wind field, and these
changes in the circulation patterns may reflect the temporal
changes associated with the wind field. Shorter period spa-
tial variability during January and July is shown in Figs. 9
and 10, respectively. ROMS simulations suggest that a small
change in the direction of the currents incident on the island
can change the nature of the current patterns around the is-
land and the location of the upwelling centre. This will be
analysed further in the next section.
During the SW monsoon, the eastward flowing SMC dom-
inates the region. However, there is a similarity in the cur-
rent fields to those observed during the NE monsoon: cur-
rents along both the western and eastern coasts flow south-
wards, with a region of convergence either along the south-
ern or southeastern coast of Sri Lanka (Fig. 15). At the be-
ginning of the sequence (12 July 2012), the currents along
both coasts converge at the centre of the southern coast of Sri
Lanka (Fig. 15a), and over the subsequent 40-day period, this
convergence zone progressively migrated along the southern
coast to the eastern coast (Fig. 15b–d). As a result, the cold
water region associated with the convergence of the currents
also migrated to the east. The SST patterns predicted by the
model were very similar to those observed in the satellite im-
agery (cf. Figs. 15 and 3c, d).
A similar time sequence in 2012 also indicated a simi-
lar process (Fig. 16). Initially, the currents were strongly to-
wards the east associated with the well-established SMC, and
southerly flow along the eastern coast was weak. The sur-
face currents and SST shown in Fig. 16a and b corresponded
to the peak of the SMC in August 2012 (Fig. 13b). How-
ever, at the end of August, the flow along the eastern coast
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5923
40
896
897
898
899
900
901
902
Fig. 13: Time series of volume transport across a transect located to the south of Sri Lanka (b) and 903
sea surface temperature time (SST) series along the south-west (black), south (blue) and south-east 904
(red) coast of Sri Lanka. Locations of the volume transport time series transect and the SST time 905
series are shown in (a). 906
(a)
(b)
Figure 13. Time series of volume transport across a transect located to the south of Sri Lanka (b) and sea surface temperature (SST) time
series along the southwestern (black), southern (blue) and southeastern (red) coasts of Sri Lanka. Locations of the volume transport time
series transect and the SST time series are shown in (a).
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5924 A. de Vos et al.: Surface circulation and upwelling patterns
41
907
908
909
910
Fig. 14. Predicted near-surface current vectors plotted on the surface temperature showing sporadic 911
upwelling (a) 01-10 January 2011 (b) 11-20 January 2011 and (c) 21 -31 January 2011. 912
Figure 14. Predicted near-surface current vectors plotted on the
surface temperature showing sporadic upwelling in (a) 1–10 Ja-
nuary 2011, (b) 11–20 January 2011 and c) 21–31 January 2011.
strengthened, and a convergence zone was present along the
southern coast, with colder temperatures (Fig. 16c). This is
also reflected in the time series of SST, where the SST at the
southern and southeastern stations decreased during this pe-
riod (Fig. 13c). In the subsequent 10 days, the convergence
zone migrated to the east (Fig. 16d) and was reflected in the
time series (Fig. 13b).
The model results and satellite imagery for both the NE
and SW monsoon periods indicated that, in general, south-
ward currents flow along both coasts of Sri Lanka, resulting
in a convergence region along the southern half of the is-
land. During the NE monsoon, this convergence region mi-
grates to the west (Fig. 14), whilst during the SW monsoon,
42
913
Fig. 15. Predicted near-surface current vectors plotted on the surface temperature showing 914
migration of the upwelling centre to the east: (a) 12 July 2011; (b) 22 July 2011; (c) 02 August 915
2011; (d) 12 August 2011. 916
917
Figure 15. Predicted near-surface current vectors plotted on the sur-
face temperature showing migration of the upwelling centre to the
east: (a) 12 July 2011, (b) 22 July 2011, (c) 2 August 2011 and
(d) 12 August 2011.
the convergence region migrates to the east (Figs. 15 and
16). Idealised model runs were undertaken to investigate the
mechanisms causing this migration, which was hypothesised
to be due to different wind stresses on each of the coasts.
Three idealised model runs were undertaken with constant
northerly winds as follows: (1) wind stress of 0.28Pa off
the eastern coast and 0.14Pa along the western coast; (2)
wind stress of 0.28Pa along both coasts; and (3) wind stress
of 0.14Pa off the eastern coast and 0.28Pa along the west-
ern coast (i.e. the opposite of 1). The results indicated that
when the wind stress was equal along both coasts, the up-
welling region was located directly off the southern coast
(Fig. 17), whilst when the wind stress was stronger on the
western (eastern) coast, the upwelling region migrated to the
east (west). Thus, the location of the upwelling appears to
be controlled by the relative strengths of the winds along
each coast, which changes with season due to the changing
monsoon. However, the surface currents and upwelling were
much stronger during the SW monsoon compared to during
the NE monsoon, due to the increased wind strengths (see
also Fig. 13).
3.4 Sri Lanka Dome
One of the major features observed during the SW monsoon
period is the presence of the Sri Lanka Dome (SLD, Fig. 2).
Here, the SMC flows eastward along the southern coast of
Sri Lanka and creates a recirculation in the lee (east) of Sri
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5925
43
918
919
Fig. 16. Predicted near-surface current vectors plotted on the surface temperature showing 920
development of upwelling during the south-west monsoon in 2012: (a) 10 August 2012; (b) 20 921
August 2012; (c) 1 September 2012; (d) 10 September 2012. 922
923
Figure 16. Predicted near-surface current vectors plotted on the surface temperature showing the development of upwelling during the
southwest monsoon in 2012: (a) 10 August 2012, (b) 20 August 2012, (c) 1 September 2012 and (d) 10 September 2012.
Lanka, with the western arm creating a southward current
along the eastern coast. The features of the dome were iden-
tified in the satellite climatology (Fig. 8d) and in the nu-
merical model output (Fig. 10c). Analysis of the climato-
logical thermal structure along 85◦E by Vinayachandran and
Yamagata (1998) indicated well-developed upward doming
isotherms, and they attributed the presence of the dome to
open ocean Ekman pumping. The SLD is analogous to flow
patterns in the lee of headlands and islands, with the island
of Sri Lanka acting as a headland interacting with the east-
ward flowing SMC (e.g. Pattiaratchi et al., 1987). A series
of idealised model runs were undertaken to examine the hy-
pothesis that the SLD is formed through the interaction be-
tween the SMC and the topography. Here, 15-day model runs
with constant westerly winds of 2, 4, 6 and 8ms−1were un-
dertaken. An additional model was also undertaken with the
removal of the land mass of Sri Lanka. Here, water depths
less than 500m surrounding Sri Lanka and the land area
were set to 500m water depth. The wind speeds selected
were based on observed winds (Fig. 6), and westerly winds
were prescribed in the model, as this was the main direction
of winds to the west of Sri Lanka during the SW monsoon
(Fig. 7). The results indicate that, under wind forcing, a re-
circulation occurred in the lee of Sri Lanka. In contrast, for
the case where the Sri Lanka land mass was absent, there
was no re-circulation in the lee of the island. The recircula-
tion strengthened (increased in vorticity) with an increase in
the wind speed, although the location of the centre remained
at the same location around 84◦E and 7–8◦N, with a slight
migration to the east with increasing wind stress. These re-
sults indicate that the primary formation mechanism of the
SLD is the interaction between the SMC and the land mass
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5926 A. de Vos et al.: Surface circulation and upwelling patterns
44
924
925
926
Fig. 17. Predicted locations of the convergence region and associated upwelling region with respect 927
to different stress along each coast. The contours represent sea surface temperature (SST) with the 928
colour representing the magnitude of the wind stress. 929
930
931
932
Figure 17. Predicted locations of the convergence region and asso-
ciated upwelling region with respect to different stresses along each
coast. The contours represent sea surface temperature (SST) with
the colour representing the magnitude of the wind stress.
of Sri Lanka. This does not rule out the possibility that Ek-
man pumping may play a role in strengthening the dome.
4 Discussion
The seasonal and shorter term (∼10 days) changes in the
surface circulation and upwelling patterns around Sri Lanka
were examined using satellite imagery (mainly ocean colour)
and a high spatial resolution numerical model (ROMS) con-
figured to the study region and forced with ECMWF interim
data. The model reproduced all of the documented major cir-
culation features in the region: reversing monsoon currents in
response to the changing wind field and the Sri Lanka Dome.
The model predictions of sea surface temperature patterns
were similar to those observed by satellite imagery. Model
output was used to update the transport rates of the SMC
and NMC between Sri Lanka and the Equator. Using a cur-
rent meter array located to the south of Sri Lanka, Schott
et al. (1994) estimated transport rates of 8 and 12Sv for
SMC and NMC. These values are contradictory in that with
stronger SW monsoon winds it would be expected that SMC
transport rates are higher than those for the NMC. The nu-
merical model output indicates mean transport rates of 11.5
and 9.6Sv for SMC and NMC respectively (Table 2). The
values for the NMC are similar to those estimated by Schott
et al. (1994), but that for the SMC is now higher. It should be
noted that the estimates by Schott et al. (1994) were through
the analysis of moored current meters, which did not sample
the top 30m of the water column.
The predicted flow patterns around Sri Lanka, created
through the interaction between SMC and the island leads,
are indicative of flow patterns observed in other regions both
in deep and shallow water; however, due to the reversing
flow, two distinct patterns can be identified:
1. During the SW monsoon, the SMC interacts with the
island, which acts more as a headland, as there is mini-
mal flow through the Palk Strait, the channel between
India and Sri Lanka. The flow follows the curvature
of the southern coast of Sri Lanka and generates a lee
eddy in the form of the Sri Lanka Dome. Using val-
ues of L∼200 km, U∼0.8ms−1, and Kh∼104m2s−1
yields a Reynolds number (Re=U L/Kh; see Sect. 1)
of ∼20 which predicts an attached eddy, which is the
Sri Lanka Dome (Fig. 1). This is confirmed by the ide-
alised model runs with constant westerly winds which
predict a stronger eddy with increasing wind (flow)
speeds (Fig. 15). Here, the Reynolds numbers range
from 5 (Fig. 15a) to 20 (Fig. 15d), and the model runs
indicate an attached eddy in the lee of the island, with
its strength increasing with increasing westerly winds.
2. During both the SW and NE monsoons, the model
results indicated southward flow along both the east-
ern and western coasts, converging along the southern
coast. In this case, circulation is similar to that of an is-
land with no discernible wake – defined as attached flow
(e.g. Alaee et al., 2004). The currents are now weaker,
and using values of L∼100km, U∼0.1ms−1, and
Kh∼104m2s−1yields a Reynolds number Re= ∼ 1 in
agreement with the theoretical predictions.
Flow along the southern coast of Sri Lanka in both mon-
soons is subject to curvature which can lead to secondary
circulation (Alaee et al., 2004). Here, as a result of the
curvature-induced centrifugal acceleration, the surface wa-
ters move offshore and are replaced by water from the sub-
surface. In the case of Sri Lanka, although located close to the
Equator, scaling reveals that the Coriolis force is important
in the dynamics (Rossby number Ro<1) and that, accord-
ing to the flow regime proposed by Alaee et al. (2004), flow
curvature is negligible in the generation of the secondary
circulation when compared to the Coriolis force (regime B,
where Ro<1 and Re>1; Alaee et al., 2004). To investigate
the importance of the Coriolis term further, model simula-
tions were undertaken with the inclusion and exclusion of the
Coriolis force during the SW monsoon. The results indicate
that when the Coriolis force was omitted, there was no up-
welling (colder water) to the west of Sri Lanka, particularly
off the southern Indian coast (Fig. 16). The upwelling feature
with convergent flow to the southeast of the island is present
in both simulations, but is enhanced and pronounced in the
model run with the inclusion of the Coriolis force. Hence,
although the Coriolis force is important in the dynamics of
the region, it does not appear to play a major role in the up-
welling along the southern coast of Sri Lanka.
In terms of upwelling patterns, case (1) clearly indicates
the presence of higher SCC within the Sri Lanka Dome
(Fig. 8), and Vinayachandran and Yamagata (1998) indicated
well-developed upward doming isotherms in a climatological
cross section of the dome. The main upwelling observed in
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5927
45
933
934
Fig.18. Predicted surface currents under constant westerly winds at (a) 2 ms-1; (b) 4 ms-1; (c) 6 ms-1; 935
and, (d) without the Sri Lankan island. 936
937
938
939
Figure 18. Predicted surface currents under constant westerly winds at (a) 2ms−1,(b) 4ms−1,(c) 6ms−1and (d) without the Sri Lankan
island.
the satellite imagery, both in terms of climatology (Fig. 3)
and individual dates (Fig. 8), indicates the dominant up-
welling regions along the southern coast of Sri Lanka. Ex-
amining the climatological monthly means indicates a wide
band of higher SCC offshore of the southern coast which
could be attributed to wind-driven coastal upwelling due to
Ekman dynamics. However, individual satellite images and
numerical model outputs indicate that the mechanism of up-
welling is more complicated. Located in the tropics, the re-
gion is frequently under cloud cover, and cloud-free satel-
lite imagery is very limited. Examination of the complete
10-year archived daily images in the ocean colour imagery
database (Feldman and McClain, 2013) yielded fewer than
10 cloud-free images for the region. However, these images
often indicate similar patterns of upwelling where there is
a “tongue” (triangular shape) of high SCC water, with the
wider section attached to the coast and tapering offshore
(Fig. 3). The location of this tongue varied along the south-
ern coast, and was present during both the SW and NE mon-
soon periods. Similarly high SCC patterns were reported by
Vinayachandran et al. (2004) (Fig. 3). Although the numer-
ical model did not include a biophysical model to simu-
late phytoplankton growth (chlorophyll), the predicted SST
distribution was remarkably similar to the higher SCC pat-
terns and the associated SST patterns observed by satellites
(Fig. 3). The model output indicated that the lower SST pat-
terns were associated with regions of convergence: currents
from both the eastern and western coasts converged in the
upwelling centre defined by lower SST, and the idealised
model runs indicated that the location of the upwelling cen-
tre was dependent on the relative wind stress along each
coast. During the NE monsoon, the upwelling centre was
shifted to the west, whilst during the SW monsoon, the up-
welling centre was shifted to the east (Fig. 17). It should
also be noted that the southern coast of Sri Lanka has a nar-
row continental shelf; hence, shelf processes as a primary
mechanism for upwelling may be neglected. There are no
previous studies which have addressed this type of circula-
tion pattern and upwelling: interaction between convergent
flows around an island leading to upwelling. The island of
Taiwan has a similar oceanographic setting, with northward
currents along both coastlines converging to the north of the
island, with upwelling along the northeastern corner (Chang
et al., 2010). However, numerical experiments indicate that
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5928 A. de Vos et al.: Surface circulation and upwelling patterns
46
940
941
Fig. 19. Predicted near-surface current vectors plotted on the surface temperature field on 15 June 942
2011. (a) simulation excluding Coriolis forcing; and, (b) simulation including Coriolis forcing. 943
944
945
946
Figure 19. Predicted near-surface current vectors plotted on the surface temperature field on 15 June 2011. (a) Simulation excluding Coriolis
forcing, and (b) simulation including Coriolis forcing.
there is recirculation to the north of the island, and the up-
welling is due mainly to the Kuroshio current encroaching
onto the shelf (Chang et al., 2010). On a smaller scale, Mag-
nell et al. (1990) show enhanced upwelling at Cape Mendo-
cino resulting from converging currents at the tip of the cape.
Through continuity, horizontal divergence at the sea surface
results in vertical upwelling of water from depth. The numer-
ical model results, confirmed by the high SCC patterns, con-
firm this process: the currents flowing parallel to the eastern
and western coasts converge along the southern coastand are
deflected offshore. As the water is flowing offshore, there is
divergence of water at the coast, which results in upwelling
of colder water from depth. This was confirmed by the nu-
merical model output, which indicated a lower sea surface
height at the centre of upwelling.
The observation that blue whales (Balaenoptera muscu-
lus) feed off the southern coast of Sri Lanka during the NE
monsoon period (de Vos et al., 2014) provided the motivation
for this study. The NE winds, under Ekman dynamics, would
generate a downwelling system (onshore Ekman flow) along
the southern coast of Sri Lanka, resulting in a low primary
productive system. The results of this study are able to ex-
plain that the upwelling system along the southern coast of
Sri Lanka is not driven by Ekman dynamics, but rather by an
interaction of the wind-driven circulation around the island.
This results in a converging coastal current system that flows
offshore, creating a divergence at the coastline, resulting in
upwelling which is able to maintain a relatively higher pro-
ductivity system during both monsoon periods.
5 Conclusions
This paper has explored the elements of the dynamics of
the surface circulation and coastal upwelling in the waters
around Sri Lanka, located in the northern Indian Ocean, a
region influenced by seasonally reversing monsoon winds,
through satellite imagery and a numerical model. Numeri-
cal model predictions compared well with the limited field
data and satellite observations. The main conclusions may
be summarised as follows:
1. The results confirmed the presence of the eastward flow-
ing Southwest Monsoon Current (SMC) during the SW
monsoon and the westward flowing Northeast Monsoon
Current (NMC), respectively. The predicted mean trans-
ports over the period 2010–2012 for the SMC and NMC
were 11.5 and 9.6Sv, respectively.
2. Sri Lanka Dome, a recirculation feature located to the
east of Sri Lanka during the SW monsoon, is the result
of the interaction between SMC and the island, resulting
in a recirculation eddy. It is possible that the eddy is
enhanced through wind stress curl.
3. During both monsoon periods, the flow along the east-
ern and western coasts was southward, converging
along the southern coast. During the SW monsoon, the
island deflected the eastward flowing SMC southward,
whilst along the eastern coast, the southward flow re-
sults from the Sri Lanka Dome recirculation.
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
A. de Vos et al.: Surface circulation and upwelling patterns 5929
4. The major upwelling region, during both monsoon peri-
ods, is located along the southern coast, and results from
flow convergence and the associated offshore transport
of water. Higher SCC values were observed during the
SW monsoon. The location of the flow convergence
and hence the upwelling centre was dependent on the
relative strengths of wind-driven flow along the east-
ern and western coasts: during the SW (NE) monsoon,
the flow along the western (eastern) coast was stronger,
and hence the upwelling centre was shifted to the east
(west).
Acknowledgements. A. de Vos was supported by a UWA Scholar-
ship for International Research Fees (SIRF).
Edited by: R. Hood
References
Alaee, J. M., Ivey, G., and Pattiaratchi, C.: Secondary circulation
induced by flow curvature and Coriolis effects around headlands
and islands, Ocean Dynam., 54, 27–38, 2004.
Alaee, J. M., Pattiaratchi, C. B., and Ivey, G. N.: Numerical simula-
tion of the summer wake of Rottenest Island, Western Australia,
Dynam. Atmos. Oc., 43, 171–198, 2007.
Barkley, R. A.: Johnston Atoll’s wake, J. Mar. Res., 30, 210–216,
1972.
Barton, E. D., Basterretxea, G., Flament, P., Mitchelson-Jacob, E.
G., Jones, B., Arístegui, J., and Herrera, L.: region of Gran Ca-
naria, J. Geophys. Res., 105, 173–117, 2000
Bleck, R.: An oceanic general circulation model framed in hybrid
isopycnic-cartesian coordinates, Oc. Modell., 4, 55–88, 2002.
Bowman, M. J., Dietrich, D. E., and Lin, C. A.: Observations and
modelling of mesoscale ocean circulation near a small island,
in: Small Islands: Marine Science and Sustainable Development,
Coastal and Estuarine Studies, edited by: Maul, G., American
Geophysical Union, Washington DC, 18–35, 1996.
Bucciarelli, E., Blain, S. P., and Tréguer, P.: Iron and manganese
in the wake of the Kerguelen Islands (Southern Ocean), Mar.
Chem., 73, 21–36, 2001.
Caldeira, R. M. A., Groom, S., Miller, P., Pilgrim, D., and Nezlin,
N. P.: Sea-surface signatures of the island mass effect phenomena
around Madeira Island, Northeast Atlantic, Remote Sens. Envi-
ron., 80, 336–360, 2002.
Chang, Y.-L., Oey, L.-Y., Wu, C.-R., and Lu, H.-F.: Why are there
upwellings on the northern shelf of Taiwan under northeasterly
winds?, J. Phys. Oceanogr., 40, 1405–1417, 2010.
Cowen, R. K. and Castro, L. R.: Relation of coral reef fish larval
distributions to island scale circulation around Barbados, West
Indies, Bull. Mar. Sci., 54, 228–244, 1994.
de Vos, A., Pattiaratchi, C. B., and Harcourt, R. G.: Inter-annual
variability in blue whale distribution off southern Sri Lanka be-
tween 2011 and 2012, J. Mar. Sci. Engin., 2, 534–550, 2014.
Di Lorenzo, E., Moore, A. M., Arango, H. G., Cornuelle, B. D.,
Miller, A. J., Powell, B., Chua, B. S., and Bennett, A. F.: Weak
and strong constraint data assimilation in the inverse Regional
Ocean Modeling System (ROMS): Development and application
for a baroclinic coastal upwelling system, Ocean Model. Online,
16, 160–187, 2007.
Dong, C. and McWilliams, J. C.: A numberical study of island
wakes in the Southern California Bight, Cont. Shelf Res., 27,
1233–1248, 2007.
Dong, C., Idica, E. Y., and McWilliams, J. C.: Circulation and
multiple-scale variabiltiy in the Southern California Bight, Prog.
Oceanogr., 82, 168–190, 2009.
Doty, M. S. and Oguri, M.: The island mass effect, J. Conserv., 22,
33–37, 1956.
Durand, F., Shankar, D., Birol, F., and Shenoi S. S. C.: Spa-
tiotemporal structure of the East India Coastal Current from
satellite altimetry, J. Geophys. Res. (oceans), 114, C02013,
doi:10.1029/2008JC004807, 2009.
Hafner, J. and Xie, S.-P.: Far-field simulation of the Hawaiian wake:
Sea surface temperature and orographic effects, J. Atmos. Sci.,
60, 3021–3032, 2003.
Haidvogel, D. B., Arango, H., Budgell, W. P., Cornuelle, B. D., Cur-
chitser, E., Di Lorenzo, E., Fennel, K., Geyer, W. R., Hermann,
A. J., Lanerolle, L., Levin, J., McWilliams, J. C., Miller, A. J.,
Moore, A. M., Powell, T. M., Shchepetkin, A. F., Sherwood, C.
R., Signell, R. P., Warner, J. C., and Wilkin, J.: Ocean forecast-
ing in terrain-following coordinates: Formulation and skill as-
sessment of the Regional Ocean Modeling System, J. Computat.
Phys., 227, 3595–3624, 2008.
Heywood, K. J., Barton, E. D., and Simpson, J. H.: The effects of
flow disturbance by an oceanic island, J. Mar. Res., 48, 55–73,
1990.
Jensen, T. G.: Arabian Sea and Bay of Bengal exchange of salt and
tracers in an ocean model, Geophys. Res. Lett., 28, 3967–3970,
2001.
Kabanova, J. G.: Primary production of the northern part of the In-
dian Ocean, Oceanology, 8, 214–225, 1968.
Legeckis, R.: Satellite observations of a western boundary current in
the Bay of Bengal, J. Geophys. Res. (Oceans), 92, 12974–12978,
1987.
Levitus, S., Burgett, R., and Boyer, T. P.: World ocean atlas 1994,
NOAA Atlas NESDIS 3, US Government Printing Office, Wash-
ington, DC, 1994.
Luis, A. J. and Kawamura, H.: Wintertime wind forcing and sea
surface cooling near the south India tip observed using NSCAT
and AVHRR, Remote Sens. Environ., 73, 55–64, 2000.
Magnell, B. A., Bray, M. A., Winant, C. D., Greengrove, C. L.,
Largier, J. L., Borchardt, F., Bernstein, R. L., B., and Dorman, C.
E.: Convergent shelf flow at Cape Mendocino, Oceanography, 3,
4–11, 1990.
Marchesiello, P., McWilliams, J. C., and Shchepetkin, A.: Open
boundary conditions for long-term integration of regional
oceanic models, Ocean Model. Online, 3, 1–20, 2001.
Marchesiello, P., McWilliams, J. C., and Shchepetkin, A.: Equilib-
rium structure and dynamics of the California current system, J.
Phys. Oceanogr., 33, 753–783, 2003.
McCreary Jr., J. P., Murtugudde, R., Vialard, J., Vinayachandran,
P. N., Wiggert, J. D., Hood, R. R., Shankar, D., and Shetye, S.:
Biophysical processes in the Indian Ocean, Geophys. Monogr.
Ser., 185, 9–32, 2009.
Michisaki, R. P., Chavez, F. P., Friederich, G. E., and Kelley, M.:
Evolution of chemical and biological properties in the Arabian
www.biogeosciences.net/11/5909/2014/ Biogeosciences, 11, 5909–5930, 2014
5930 A. de Vos et al.: Surface circulation and upwelling patterns
Sea and Indian Ocean during 1995 from automated surface map-
ping, American Geophysical Union Ocean Sciences meeting,
San Diego, 1996.
Palacios, D. M.: Factors influencing the island-mass effect of
the Galápagos Archipelago, Geophys. Res. Lett., 29, 2134,
doi:10.1029/2002gl016232, 2002.
Pattiaratchi, C., James, A., and Collins, M.: Island wakes and head-
land eddies: A comparison between remotely sensed data and
laboratory experiments, J. Geophys. Res. (Oceans), 92, 783–794,
1987.
Pawlowicz, R., Beardsley, B., and Lentz, S.: Classical tidal har-
monic analysis including error estimates in MATLAB using
T_TIDE, Comput. Geosci., 28, 929–937, 2002.
Saetersdal, G., Bianchi, G., Stomme, T., and Venema, S. C.: The Dr.
Fridtjof Nansen Programme 1975-1993: Investigations of fish-
ery resources in developing countries. History of the programme
and review of results, Food and Agriculture Organisation of the
United Nations, Rome, 434 pp., 1999.
Sahini, A.: The structure, sedimentation and evolution of Indian
continental margins, in: The ocean basins and margins: The In-
dian Ocean, edited by: Nairn, E. M. and Stehli, F. G., Plenum
Press, New York, 353–398, 1982.
Schott, F., Reppin, J., Fischer, J., and Quadfasel, D.: Currents and
transports of the monsoon current south of Sri Lanka, J. Geophys.
Res.-Oceans, 99, 25127–25141, 1994.
Schott, F. A. and McCreary, Jr., J. P.: The monsoon circulation of
the Indian Ocean, Prog. Oceanogr., 51, 1–123, 2001.
Shankar, D. and Shetye, S. R.: On the dynamics of the Lakshadweep
high and low in the southeastern Arabian Sea, J. Geophys. Res.-
Oceans, 102, 12551–12562, 1997.
Shankar, D., Vinayachandran, P. N., and Unnikrishnan, A. S.: The
monsoon currents in the north Indian Ocean, Prog. Oceanogr.,
52, 63–120, 2002.
Shchepetkin, A. F. and McWilliams, J. C.: The regional
oceanic modeling system (ROMS): a split-explicit, free-surface,
topography-following-coordinate oceanic model, Ocean Model.
Online, 9, 347–404, 2005.
Shepard, E. P.: Submarine geology, Harper and Row Publishers,
New York, 1963.
Song, Y. and Haidvogel, D.: A semi-implicit ocean circulation
model using a generalized topography-following coordinate sys-
tem, J. Computat. Phys., 115, 228–244, 1994.
Sreenivas, P., Gnanaseelan, C., and Prasad, K. V. S. R.: Influence of
El Niño and Indian Ocean Dipole on sea level variability in the
Bay of Bengal, Glob. Planet. Change, 80/81, 215–225, 2012.
Swan, B.: An Introduction to the Coastal Geomorphology of Sri
Lanka, National Museums of Sri Lanka, Colombo, 1983.
Tomczak, M.: Island wakes in deep and shallow water, J. Geophys.
Res., 93, 5153–5154, 1988.
Tomczak, M. and Godfrey, J. S.: Regional oceanography: An intro-
duction, 2 Edn., Pergamon, 2003.
Varkey, M. J., Murty, V. S. N., and Suryanarayana, A.: Physi-
cal oceanography of the Bay of Bengal and Andaman Sea, in:
Oceanogr. Mar. Biol. Annu. Rev., edited by: Ansell, A. D., Gib-
son, R. N., and Barnes, M., CRC Press, 1–70, 1996.
Vinayachandran, P. N. and Yamagata, T.: Monsoon response of the
sea around Sri Lanka: Generation of thermal domes and anticy-
clonic vortices, J. Phys. Oceanogr., 28, 1946–1960, 1998.
Vinayachandran, P. N., Masumoto, Y., Mikawa, T., and Yamagata,
T.: Intrusion of the southwest monsoon current into the Bay of
Bengal, J. Geophys. Res.-Oceans, 104, 11077–11085, 1999.
Vinayachandran, P. N., Chauhan, P., Mohan, M., and Nayak,
S.: Biological response of the sea around Sri Lanka
to summer monsoon, Geophys. Res. Lett., 31, L01302,
doi:10.1029/2003GL018533, 2004.
Wijeratne, E. M. S.: Tidal characteristics and modelling of tidal
wave propagation in shallow lagoons of Sri Lanka, Ph.D.,
Gothenburg University, Gothenburg, 2003.
Wijeratne, E. M. S., Woodworth, P. L., and Stepanov, V. N.: The
seasonal cycle of sea level in Sri Lanka and southern India, West
Indian Ocean J. Mar. Sci., 7, 29–43, 2008.
Wijeyananda, N. P.: Maritime zones, in: Arjuna’s atlas of Sri Lanka,
edited by: Somasekaram, T., Perera, M., de Silva, M. B. G.,
and Godellawatta, H., Arjuna Consulting Company Limited,
Colombo, 5–7, 1997.
Wolanski, E., Imberger, J., and Heron, M. L.: Island wakes in shal-
low coastal waters, J. Geophys. Res., 89, 553–569, 1984.
Wyrtiki, K.: Physical oceanography of the Indian Ocean, in: The Bi-
ology of the Indian Ocean, edited by: Zeitzschel, B. and Gerlach,
S. A., Springer-Verlag, Heidelberg, 1973.
Xu, J., Lowe, R. J., Ivey, G. N., Pattiaratchi, C. B., Jones, N. L.,
and Brinkman, R.: Dynamics of the summer shelf circulation
and transient upwelling off Ningaloo Reef, Western Australia,
J. Geophys. Res.-Oceans, 118, 1099–1125, 2013.
Yapa, K. K. A. S.: Upwelling phenomena in the southern coastal
waters of Sri Lanka during southwest monsoon period as seen
from MODIS, Sri Lanka J. Phys., 10, 7–15, 2009.
Biogeosciences, 11, 5909–5930, 2014 www.biogeosciences.net/11/5909/2014/
