sea surface salinity as detected by smos, aquarius and by...
TRANSCRIPT
a
c
b
ISAS SMOS OI
SATELLITE SSS (level 3, monthly maps)
● SMOS LOCEAN derived from ESA L2OS v5 (wind-model 1) SSS:
- SMOS LOCEAN 0.25°: SMOS SSS maps averaged over 0.25x0.25°
- SMOS OI: SMOS LOCEAN 0.25° interpolated with similiar method as the one
applied to ISAS maps.
- SMOS CEC LOCEAN: SMOS SSS (originally at ~43km resolution) maps are
averaged over 100x100km2 and oversampled every 0.25° (Boutin et al. 2013).
● SMOS IFREMER derived from ESA Level 1 product
- SMOS CEC IFREMER: SMOS-CATDS CEC IFREMER SSS maps, over a regular
grid of 0.25° x 0.25°, with a daily 5°x 5° adjustment with respect to World Ocean
Atlas 2001 climatology.
● AQUARIUS level 3 maps: 2 products V2.0 and V2 CAP versions 1x1°resolution
Sea surface salinity as detected by SMOS, AQUARIUS and by in situ sensors: In and around the SPURS / STRASSE experiment
In the frame of the Salinity Processes in the Upper Ocean Regional Study (SPURS) experiments, the STRASSE (Subtropical Atlantic surface salinity experiment,
PI G. Reverdin) campaign took place in August-September 2012, providing a very high resolution monitoring of the salinity in the high salinity region of the
subtropical North Atlantic. In addition to this survey, Sea Surface Salinity (SSS) monitoring across the high salinity region is performed every month by ships of
opportunity (Toucan and Colibri). Here, we take advantage of this wide in situ SSS monitoring for validating the SMOS (ESA version 5) and AQUARIUS (level 3
version 2 ) SSS and evidence the gain of coverage and resolution of satellite product over in situ interpolated products (Argo, ISAS, WOA,...).
SSS seasonal variability in the SPURS
region from SMOS and Aquarius is not
well consistent with observations (Fig 2).
Satellites do not retrieve properly absolute
SSS values: on SMOS SSS (except CEC
IFREMER adjusted to WOA), year to year
consistent seasonal bias are observed.
In order to study the spatial variability
of SSS, all datasets (SMOS and
Aquarius) are corrected from a bias B1
calculated for each month on the SPURS
region, as following:
Fig. 2: Mean SSS variability over SPURS region from
various product.
B1 = <SSS>SPURS - < SSS ISAS >SPURS
.
.
3 - Large scale variability
Fig 10: Scatter plot of SSS anomaly for SMOS OI and ISAS versus TSG SSS. TSG SSS was filtered at 300km. All products are gridded to 1°.
● SSS variability at scales of several hundreds of km is at first order consistent between SMOS, AQUARIUS and ISAS, but seasonal biases on satellite SSS first need to be
corrected. After correction, the calculated RMSE of SMOS CEC LOCEAN, AQUARIUS V2CAP and ISAS in respect to 14 transects of SSS TSG from 09-2011 to 12-2012 was
estimated to be 0.14. Global correlation coefficient r was estimated to be 0.93.
● Satellite products better reproduce SSS mesoscale variability (around ̴ 100 km) than interpolated in situ data. SMOS performs better than AQUARIUS.
References: Boutin, J., Martin, N., Reverdin, G., Yin, X., and Gaillard, F. (2013). Sea surface freshening inferred from SMOS and ARGO salinity: impact of rain. Ocean
Sci. 9, 183-192.
Gaillard, F., Autret, E., Thierry, V., Galaup, P., Coatanoan, C., and Loubrieu, T. (2009).Quality control of large argo datasets. Journal of Atmospheric and
Oceanic Technology.
Reul, N., Fournier, S., Boutin, J., Hernandez, O., Maes, C., Chapron, B., Alory, G., Quilfen, Tenerelli, J., Morisset, S., Kerr, Y., and Mecklenburg, S. (2013).
Sea surface salinity observations from space with SMOS satellite: a new tool to better monitor the marine branch of the water cycle. Surveys in Geophysics.
SMOS SSS CATDS CEC are available at www.catds.fr
Acknowledgements:
This study is supported by the ESA SMOS+SOS ('OSMOSIX') projet. It is part of the ESA SMOS/GLOSCAL Cal/Val
project funded by CNES It is also partly performed and funded in the frame of the ESA level 2 Ocean Salinity expert
support laboratory (ESL) projects. Sea surface salinity data derived from thermosalinograph instruments installed
onboard Toucan and Colibri voluntary observing ships were collected, validated, archived, and made freely available by
the French Sea Surface Salinity Observation Service (http://www.legos.obs-mip.fr/observations/sss/). AQUARIUS V2.0
version and AQUARIUS V2.0 CAP version were obtained from the Physical Oceanography Distributed Active Archive
Center (PO.DAAC) at the NASA Jet Propulsion Laboratory, Pasadena, CA. http://podaac.jpl.nasa.gov."
Collocation SSS products with ship data :
Level 3 SSS products at 0.25° (bilinear
interpolation) are compared to TSG SSS, binned
at 0.25°, except in Fig 10.
Statistics are done during the overlapping period
of SMOS and AQUARIUS.
The region not affected by coast proximity was
selected for statistics (green square):
[50°W:27°W, 15°N:35°N]
Table 1: Statistics at 0.25°, n = 1723,
standard deviation of TSG =0.36. MBE=
mean bias estimate, RMSE= root mean
square error, r= correlation at 99% interval
confidence, σP: standard deviation of each
SSS product. Results of SMOS and ISAS
statistics remains unchanged by using 16
more transects (from 07-2010 to 08-2011).
● SMOS LOCEAN, LOCEAN OI, ISAS
OI and AQUARIUS V2CAP are the best
products to represent the observed
large scale structures (best RMSE,
best r2, Table 1).
● At 0.25º grid resolution, noise in SMOS
is reduced by averaging to 100x100km2:
SMOS OI shows good correlation but
less variability (lower σP)
Tradeoff between noise effect and
natural variability.
Fig 5: Scatter plot of SSS SMOS CEC
LOCEAN versus ship SSS.
4 – Mesoscale features
Colocation of all SSS products with TSG SSS
Fig 4: Spatial SSS maps from TSG ship (a), SMOS CEC LOCEAN (b), AQUARIUS
V2CAP (c), SMOS OI (d) and ISAS map (+ argo floats) (f). SSS from ships are
superimposed in color over SSS maps. The applied bias correction is B1 = -0.15. (f)
Isohaline contour for ISAS (red), AQUARIUS V2 CAP (cyan) and SMOS (black).
Fig 6: Position of the longitude (top) and latitude
(bottom) of the barycenter of the sea surface salinity
maximum for SMOS, ISAS and AQUARIUS V2CAP.
2- Salinity Data 3 - Methods
Fig 7: SSS anomaly maps from TSG ship (a), SMOS CEC LOCEAN (b) and AQUARIUS V2CAP (c). TSG SSS anomaly are superimposed in color over SSS maps.
● SMOS SSS and AQUARIUS SSS anomaly are significantly correlated with TSG SSS anomaly: r2 SMOS = 0.39, r2 AQUARIUS = 0.23
Fig 3 : Mean of monthly standard deviation of SSS
on 0.25º pixel between 2010 and 2012. Between
27°W and 20°W, high values are observed due to
a coastal effect on SMOS.
(Gaillard et al. 2009, D7CA2S0 re-analysis, at ~5m
depth) : sampled at 0.5° but nominal resolved
resolution of Argo area ~3°
● SSS TSG (at ~5m depth) from Ships of
Opportunity (Toucan, Colibri ship) from 08-2011 to
12- 2012 (data from french ORE-SSS)
● SSS thermosalinograph (TSG) from STRASSE
campaign (08-09 2012)
IN SITU SSS
● Optimal Interpolation of ARGO SSS: IFREMER In Situ Analysis System (ISAS)
● At 1º resolution grid, after filtering TSG SSS at 300km (as interpolated products), SMOS OI have better correlation than ISAS:
r2 SMOS = 0.34, r2 ISAS = 0.28.
O. Hernandez (1), J. Boutin (1) N. Kolodziejczyk(1) , G. Reverdin (1) , N. Martin(1), F. Gaillard(2) , N. Reul(3) , , J.L. Vergely(4)
(1)LOCEAN/IPSL, Paris, France, (2)LPO/IFREMER, Brest, France, (3)LOS/IFREMER, Toulon, France, (4)Acri-st, Paris, France
● SSS maximum present a seasonal
cycle, moving north to south from
summer to spring (Fig 6) from about
0.8º. In longitude, no seasonal cycle
is observed.
● SSS spatial anomalies in each product have been computed removing the monthly climatological lSAS
SSS (Fig 7).
● SMOS and AQUARIUS show higher SSS signal and smaller scales than in ISAS and SMOS OI product.
Despite a high noise level, it suggests that mesoscale features present in TSG ships are retrieved in satellite product.
Example: In south western zone of SPURS area
(red square, Fig. 7 and 9) SMOS and AQUARIUS
retrieved better SSS because lack of ARGO data
in this area.
Fig 1.:
14 transects
of TSG SSS
a
37.5
37.0
a b
d
f
c
e
SMOS also more accurate than ISAS at 300km
resolution, likely due to a better coverage of the
SMOS SSS measurements.
● Illustration of these SSS maps is presented Fig. 3. The
spatial structure of SSS shows difference (Fig. 4), which are
not evidenced in global statistics. Statistics are dominated by
the signal at low frequency.
● On satellite products, the maximum SSS isohaline
contours (Fig. 4f), show mesoscale features as observed by
looking isohaline contours (Fig 4f).
In order to study the ability of satellite products to retrieve
mesoscale feature, spatial SSS anomalies were studied by
removing large scale SSS signal.
SMOS CEC LOCEAN AQUARIUS V2CAP ISAS
SSS TSG ANOMALY SSS TSG ANOMALY SSS TSG ANOMALY
SS
S S
MO
S A
NO
MA
LY
SS
S A
QU
AR
IUS
AN
OM
ALY
SS
S IS
AS
AN
OM
ALY
Fig 8: Scatter plot of SSS anomaly sampled at 0.25° for SMOS CEC LOCEAN, AQUARIUS V2CAP and ISAS versus TSG SSS anomaly.
Fig 9: SMOS OI (a) and ISAS map (+ argo floats) (b). TSG SSS anomaly filtered at 300km are superimposed in color over SSS maps.
Satellite products better reproduce SSS
mesoscale variability than interpolated in situ
data.
b