Geoscience Reference
In-Depth Information
(a)
ARGO profile on 11/8 at 20h
(b)
ARGO profile on 22/08 6:52 UTC
0
0
50
50
SSSsmos on 11/8 at 13h39
(6h30 before ARGO profile)
SSSsmos on 21/8 at 13h49
(17h before ARGO profile
100
100
SSSsmos on 16/8 at 13h44
(5 days after ARGO profile)
150
SSSsmos on 24/8 at 13h33
(2 days after ARGO profile)
150
200
200
32.5
32.5
33.0
33.5
34.0
34.5
35.0
33.0
33.5
34.0
34.5
35.0
SSS
SSS
(c)
Fig. 13 Two successive Argo profiles taken by float 4900325 (blue curve) in the eastern tropical Pacific on
a 11 August 20:00 UTC (latitude = 12.4N; longitude = 117.6W) and b 22 August 6:52 UTC (latitude:
12.2N; longitude: 117.8W). Mean SMOS SSS collocated within a 5-day window and a radii of 50 km with
these profiles are indicated by red dashed point. In each case, two SMOS passes have participated to these
collocations: mean SMOS SSS corresponding to each pass is indicated as red filled point. The corresponding
ISAS SSS in August is indicated by the green point. The time series of the 3-hourly satellite rain rate from
TRMM 3B42 and averaged over (11-13N; 116-118W) is provided in (c). The time at which SMOS and
Argo acquired SSS data is indicated by red and blue dots, respectively
stratification, a triple collocation between Argo, SMOS level 2 products (at *40-km
resolution, non-averaged in time) and SSMI satellite rain rate (RR) data was conducted.
SMOS and SSMI RR data were colocated within a temporal window of -40 min and
?80 min, while a ±5-day windows was considered to colocate SMOS and Argo data.
The theoretical error on the SMOS SSS retrieved level 2 data used in this colocation
exercise is *0.5. Without any RR sorting, the statistical distribution of the differences DSSS
is skewed toward negative values (Fig.
15
, Table
3
); when only SMOS non-rainy events are
considered, the negative skewness disappears, and statistics of the SMOS-Argo differences
in the tropical Pacific Ocean become close to the ones in the subtropical Atlantic Ocean
(Tables
2
,
3
). Largest skewness toward negatives differences is obtained when only SMOS
SSS close to rain events are considered. For these rainy SMOS cases, we find a negative
dependency of the SMOS-Argo SSS differences with respect to SSMIs RR of -0.17
pss/mm
-1
h, i.e., a freshening of 1.7 for a SSM/I RR of 10 mm h
-1
(Boutin et al.
2012a
,
b
).
The non-sorting of SMOS measurements close in time with rain events in SMOS-Argo
collocated data sets (within 10 days and 100 km) is responsible for (1) a mean -0.1 negative
bias over 3 months between 5 and 15N in the tropical Pacific region with respect to non-
rainy conditions and with respect to the subtropical Atlantic region and (2) a negative
skewness of the statistical distribution of SMOS minus Argo SSS difference (Fig.
15
). Given
that the whole set of SMOS-Argo collocations also includes the situations with rainy Argo
measurements collocated with non-rainy SMOS measurements, these results indicate a
