Content uploaded by Thierry Delcroix
Author content
All content in this area was uploaded by Thierry Delcroix on Dec 15, 2015
Content may be subject to copyright.
The
Atlantic North Equatorial Countercurrent Inferred From Dynamic
Height
and
Thermocline
Depth
From a given temperature profile, three
basic oceanographic parameters are Usually
derived: the dynamic height
Dh
(through a
mean T-S curve), the heat content
Ht,
and
the thermocline depth
Th,
the latter being
frequently defined
as
the depth
of
a fixed
isotherm
or
of
the maximum temperature
gradient. Any relations among these para-
meters should be
useful
for
ocean-monitor-
ing
or
interpretations
of
the results of
numerical models. Therefore,
I
document
here the reltionships among these three pa-
rameters
in
the tropical Atlantic Ocean and
then analyze,
as
an example,
the
impact of
using specific parameters
(Th
and
Dh)
to
estimate the annual. cycle
of
the western
part
of
the Atlantic North Equatorial Coun-
tercurrent (NECC).
The
basic data used in
this
note have been gathered by- the FOCAL
program and
are
described by Merle and
Delcroix
(1984)
and
Merle
and Arnault
(1985).
The
three
parameters referred to above,
Dh,
Ht,
and Th, are first compared through
linear regression analysis. Correlation
coef-
ficients
r
were calculated from teniperature
profiles collected over
4"
longitude by 2
"
latitude rectangles, including at least 20 ob-
servations running from the surface
to
300
m.
The
correlations between the depth
of
the 20°C isotherm and
the
heat content, de-
fied
as
the mean temperature
of
the upper,
300
m.
first
appear to be very useful west
of
20%' and between
10"s
and 20°N (Fig-
ure
1).
In
fact,
in
the latter region, the cor-
relations are always
>
+
0.8,
i.
e.,
more than
65%
of
the heat-content variability can be
attributed to its linear relation with the
depth
of
the
20°C isotherm. Nevertheless,
the heat content
can
be determined &om
the depth
of
the 2OoC isotherm, with
a
standard
error
of about
0.2OC
(4"s-1O"N)
to
0.4"C
(IO"S-4"S
and lO"N-20"N). In the
4"s-
IOW
band,
this
is
an
order
of
magnitude
smaller
than
the
mean annual cycle.
of
heat..-
storage (Merle,
1983)
and
in
the
ratio
of
1
in
the
10"s-4"s
and 10"N-2OoN bands. The
correlations
are
thus
only
interesting in the
first
latitudinal band,
west
of
í!OW.
East
of
20W,
35%
of
the heat.content variability
(r=
+
0.6)
can
be associated
with
the
ver-
tical displacements
of
the 20°C isotherm,
moving from
30
to
80
m (Merle and Del-
croix,
1984).
The
thermal variability should
thus be considered in this area.
The correlations between depth
of
the
20°C isotherm and the -dynamic height
(O/
300
dbar) are presented (Figure 2).
A
glance at their spatial variations suggests
that the depth
of
the 20°C isotherm can be
used as a prosy
for
the dynamic height,
LONGITUDE
FIGURE
1
(Delcroix)
Correlation
coefficients,
between the
20°C
isotherm
ciepth
and
the heat content
(0-300
m).
Correlation coflickais betwe&. the depth
of
the 20°C isotherm and the.surface
ckvnamk
height relative to
300
db&\
mainly between
7s-
IOON
und webt of
15"W
and
in
a small part
of
the Gulf
of
<;iiinra
(r
+OB).
Nevertheless, in these regions,
the dynamic height can be inferred from
the
depth of the 20°C isotherm with a niean
standard error
of
0.35
ni's-'
(5.5
dyn
cni).
Such a value is half
the
amplitude
of
the
annual component
of
the
surface dynamic
height relative to
500
dhtr (Merle ;incl
Ar-
nault,
1985).
which is close to
the
SOO-dhxr
reference level. Therefore.
using
the depth
of
the 2OoC isotherm
to
deduce quxitita-
tively
the
surface dynamic height relative to
300
dbx
appears
to
be
quite
imprecise
for
use
in the whole basin.
The correlations between heat content
and dynamic height are not presented here
since they were always
>
+
0.9
in
the
\viiole
basin, with
no
significant spatial variations.
The second part
of
this note illubtrates the
differences obtained in
the
me;m annual
cycle
of
the western part
of
the Atlantic
NECC,
Le.,
the
region
of
4"N-1O"N;
j5W-
45"W,
inferred from the slope of the ther-
mocline (depth
of
the
20°C
isotherm and
maximum temperature gradient
)
and
.?om
dynamic height gradients. Thib latter region
was chosen because
it
is
an area
of
niaxi-
mum correlation among the three parame-
ters referred to above. Although some
previous investigations (Merle and l>elcroix,
1984;
Merle and Arnault,
1985:
Garzoli and
Katz,
1983)
already noted disagreements
when monitoring different paramiters in
the NECC, they
coli@
be attributed
to
their
own raw data management.
To
prevent such
possibilities, we made similar treatments
for
the parameters (average over
2"
latitude by
10"
longitude boxes, weighted interpolation,
and estrapolation b,y- imposing
an
annual
sinusoidal cycle).
Two estimates
of
the annual cycle
of
the
western Atlantic NECC, from the meridional
slopes
of
the depths
of
the
20°C iostherm
the maximum temperature gradient be-
tween the 2"N-4"N/3
jW-45"W
and
1O"N-
Ií!"N/35W-45"W rectangles, are presented
(Figure
3).
They
are
quite consistent with
what has been observed (Merle and
Del-
croix,
1984;
Garzoli and Katz,
1983)
using
different data processing. The tonal velocity
components
of
the NECC, deduced
from
each dope, agree well
in
phase, whereas
n).
L
6
they are in the ratio
of
2
throughout
the year.
This
is because the
20°C
hothelm
is generally deeper
(
+
10
m) than the depth
of
the maximum temperature gradient in
the northem rectangle and shallower
(-5
m)
in
the southern rectangle. Calculation
of
the
meridional slopes increases these differ-
ences.
This
illustrates quite well the impact
of using a given definition of thermocline
depth.
In
a two-layer
ocean
model, the slope
of
the thermocline
is
proportional to the geos-
trophic velocity
U
in the upper layer:
where
g'
is
the
reduced gravity and
f
the
Coriolis parameter, whereas the meridional
gradient of the dynamic height at a given
level is proportional to the geostrophic ve-
locity
u
at this level:
assuming a level
of
no
motion. Therefore, a
third estimate of the annual cycle in the
NECC
was
performed
looking
at the meri-
dional gradients
of
dynamic height at differ-
ent levels. This
is
presented (Figure
4)
for
the
o,
30,
60,
and
90
dbar relative
to
the
300-dbar reference level, As previously de-
duced from the thermocline depth gra-
dients, this third estimate clearly shcws that
the average zonal velocity over a
100-m
depth
(io,
the mean annual depth of the
maximum temperature gradient in the area)
changes sign from March/ lpril
to
June.
Zonal geostrophic velocities
u,
deduced
from dynamic height gradients, reverse only
in
May in the first
30
m,
in
contrast to the
3-month period of the westward current oc-
curring between
60
and
90
m.
Using
slopes
of
the thermocline has masked this vertical
heterogeneity during the period of reversal.
In
the investigated area, variations in the
meridional gradients
of
the surface dynamic
height can not account for those
of
the
zonal surface velocity deduced from ship-
drifts (Richardson and McKee,
1984).
Ek-
man
flow
should thus be considered.
In
conclusion, a given oceanographic pa-
rameter
(Dh,
Th,
or
Ht),
calculated
in
the
tropical Atlantic Ocean, can not be reliably
-4
u=
-
fg'/
f)
.
f
d
Th
1
du),
U'
-(¡/f).(dDh/du),
-
from
max.
temp.
grad.
I
FIGURE
3
(Delero&)
Themzocline
depth differences defined
as
the 20°C
isothem
depth
ad
tbe depth
of
the maximum temperature gradient
between the rec&ng&s 2"N-4"N/35"W-45"W
and
lO"N-12"N/35"W-45"u!
used to deduce quantitatively the remaining
ones
through linear relations, except
for
the
depth
of
the
20°C
isotherm
and
the
heat
content
(0-300
m) west
of
20W
and
be-
tween
4"SlO"N.
It was then
shown,
using
3
different estimates of the annual cycle
of
the
western part
of
the NECC, that
the
choice
of specific parameters
(Oh
and
Th),
even
in
regions where they are strongly correlated,
is
crucial.
References
Merle
J.
and
T..
Delcroix
(1984).
Seasonnt
variability
of
the thermocline topography
in
the tropical
Atlantic
Ocean
(Submitted
to theJ.
phys
Oceatwgz).
Merle
J.
and
S.
Arnault
(1985).
Seasonal var-
iability
of
the surface dynamic topogra-
phy
in
the tropical Atlantic 0cean.J.
MI:
Res,
43,
267-288.
1.3
5
7
9
II
I
Monlhs
FIGURE
4
(DelCroix)
Dynamic
height differences at
O,
30,
GC
and
90
dbar relative to the 3OO.dba
m&ence
level,
between the rectangles 2'1%
4"Nl.35"W-45"W
and
lO0N-12"Nl35"W
45"W
Merle
J.
1983.
Seasonal
variability
of
subsur
face
thermal structure in the tropical
At
lantic Ocean.
Hydrodynamics
of
th
Equatorial
Ocean
J.
Nihoul, ed., Elsevie
Ganoli
S.
and
E.
Katz
(1983).
The force4
annual reversal of the Atlantic Nort
Equatorial Countercurrent.
J.
phy:
Richardson,
I!
and
T.
McKee
(
1984).
Averag
seasonal variations
of
the Atlantic Equ:
torial currents
from
historical shipdrift
pp.
31-50.
&~w?KJ~,
13,
2082-2090.
J.
phys Oc~~gI:, 14,
1226-1238.
Thierry Delcroi
ORsTon
B.
R35
29273
Brest Cede.
Franc
