ENSEMBLES (EN3) quality controlled T and S profiles - changes from EN2
======================================================================
Updated input data sources
==========================
WOD05 (World Ocean Database 2005) is used in place of WOD01
(main copy taken on 4 July 2006, 1990 onwards updated 4 October 2006)
GTSPP is used to supplement WOD05 for 1990 onwards
(copy of the "best" datasets 6 October 2006)
ARGO data from US-GODAE is used for recent years - this is
mainly to pick up salinity bias corrections, where available,
from the delayed mode processing (copy taken 7 September 2006)
Revised climatology
===================
The climatology is used to provide the initial conditions,
and in data voids the "objective analysis" and background
fields relax towards the monthly climatology.
For EN2 we used the World Ocean Atlas 1998 (WOA98) climatology -
interpolated as necessary.
For EN3 a monthly climatology was constructed from the
1971-2000 EN2 objective analyses. Both WOA98 and this
30-year climatology showed bulls-eyes at some levels/months.
The bulls-eyes were reduced by horizontal smoothing and
also a 1-2-1 time smoothing
(eg Tjan,final = 0.25*Tdec + 0.5*Tjan + 0.25Tfeb).
Constant value check
====================
The check for salinity was changed to match that for temperature
(so if 90% of salinity values covering at least 100 m are
identical then all the salinities are rejected).
Reject-lists
============
XBT values more than 950 m deep and less than 4 m deep are
rejected (in EN2 XBT values were rejected below 1000 m).
All reports in the Black Sea are rejected (due to limitations
in our current system).
There are also more specific rejections.
Timeseries of T300 and S300 (temperature and salinity averaged
over the top 300m) for various regions of the globe have been
examined for spikes. In some cases these can be related to
clear or probable biases/errors in individual cruises
(usually the worst values have been flagged, but sufficient
bad data remains to compromise the analysis fields).
Table 1a. Suspect cruises rejected 1950-1997
YYYY.MM Ident. Type No Area Rejected
1963.06 606734 MBT 15 Atl: Gulf of Guinea T
1964.01-12 712?00 OSD 183 Eq. East Pacific S
1964.12 294 OSD 12 Eq. East Pacific S
1965.08-09 14788 OSD 88 Indian Ocean S
1969.05-10 1533168 OSD 148 Central Pacific T&S
1974.05 4574435 XBT 168 Indian + Med. + Atl. T
1975.02-03 1396335 OSD 36 Eq. Pacific S
1978.04-07 1567399 OSD 75 Atlantic Ocean S
1980.02-03 71001 OSD 27 Eq. Pacific S
1980.02-05 556 OSD 114 Indian Ocean S
1980.03 1399350 OSD 27 SW Pacific S
1982.01 33696 OSD 4 Indian Ocean S
1982.01-03 604 OSD 112 Indian Ocean S
1984.06 6618302 XBT 6 N Atlantic T
1984.07-08 6618402 XBT 19 N Atlantic T
1984.09-12 12389093 XBT 245 Indian Ocean T
1984.10-11 1401411 OSD 36 S Pacific S
1985.05* 976777 OSD 28 N Atlantic S (* lat > 20)
1985.07-09 1356733 CTD 172 NW Pacific S
1985.10 1356733 CTD 22 NW Pacific T&S
1985.11-12 938613 CTD 8 SE Pacific S
1987.01 991865 CTD 17 Eq. Atlantic S
1987.02-04 1367033 CTD 79 Indian Ocean S
1987.10 1427916 CTD 44 S Pacific S
1988.06 791444 OSD 16 Indian Ocean S
1988.09-10 71412809 CTD 92 NW Atlantic S
1988.12 12653882 XBT 17 S Atlantic T
1989.01-02 12653882 XBT 17 S Atlantic T
1989.02-05 1543962 OSD 161 NW Pacific S
1989.07 776844 OSD 26 S Atlantic S
1990.01 1046031 OSD 105 Indian S
1990.06 884209 CTD 22 NW Atlantic T&S
1990.06 852812 OSD 10 Eq Atlantic T&S
1990.08 14270697 XBT 62 S Atlantic T
1991.03 1323 OSD 28 S China Sea S
1991.06 14283997 XBT 54 Mid Atlantic T aka HOQT
1991.07 14285097 XBT 53 Mid Atlantic T aka HOQT
1991.09 14286897 XBT 48 Mid Atlantic T aka HOQT
1991.10 14288097 XBT 16 Mid Atlantic T aka HOQT
1991.10 FTCS CTD 45 S Pacific S
1992.01 453 OSD 92 W Pacific S
1992.01 003 XBT 24 Eq Atlantic T 0-30N 20W-10E
1992.02 ELHL6 CTD 1 C Atlantic T&S (XBTs OK)
1992-1993 0059 PFL 67 N Atlantic T (2 in 1994 missed)
1992.03-05 WTER CTD 76 S Atlantic T&S
1992.03-04 SHIP CTD 24 Eq Atlantic T&S 10S-10N, 15-0W
1993.03 11437316 XBT 79 N Atlantic T aka DAKE
1993.09 44109 CTD 9 Arabian Sea S
1993.12 1002634 CTD 23 N Atlantic S
1994.06-08 3077112 OSD 156 NW Pacific T&S
1994.07-08 3077232 OSD 33 NW Pacific T&S
1994.09 1282177 CTD 12 NE Pacific S W of 140W
1995.02 1687244 CTD 12 N Pacific S
1995.03 14332515 XBT 21 C/N Atlantic T aka 9HVO3
1995.04 14334615 XBT 13 C/N Atlantic T aka 9HVO3?
1995.05 14335815 XBT 2 C/N Atlantic T aka 9HVO3?
1995.05 14155613 XBT 11 Indian Ocean T E of 30E
1995.04 LOPD CTD 36 S Atlantic S
1995.06 0250 PFL 2 S Atlantic T
1996.04 17103 PFL 1 Indian Ocean T
1996.04,07 17067 PFL 3 Indian Ocean T
1996.08 14362921 XBT 48 C/N Atlantic T aka 3ETQ5
1997.06 980213 CTD 53 Indian Ocean S
1997.08 17124 PFL 1 Indian Ocean T
1997.08 17133 PFL 1 Indian Ocean T
1997.08-12 17220 PFL 9+ N Atlantic S
1997.08-12 17221 PFL 9+ N Atlantic S
Table 1b. Summary of suspect cruises rejected 1990-2005
No rej PFL XBT CTD Temp Salt Num
/OSD ARGO
1990 4 0 1 3 3 1
1991 6 0 4 2 4 2
1992 6 1 1 4 5 1
1993 4 1 1 2 2 2
1994 3 0 0 3 2 1
1995 7 1 4 2 5 2
1996 3 2 1 0 3 0
1997 5 4 0 1 2 3
1998 18 18 0 0 6 12
1999 16 16 0 0 2 14 311
2000 20 19 0 1 3 17 371
2001 14 13 0 1 0 14 572
2002 8 8 0 0 1 7 888
2003 49 47 2 0 24 25 1264
2004 115 114 1 0 78 37 1855
2005 165 164 1 0 111 54 2555
Notes
1. The number of distinct buoys in the ARGO GDAC archive
(Num ARGO) is included for comparison in table 1b.
2. If temperature is rejected, the whole report, including
salinity is rejected.
3. The 1950s weren't examined manually as being of lower priority.
4. In general I have given reports the benefit of the doubt if they
look "marginal" to me. There is of course subjective judgement in
these decisions. There is also a risk that a few identifiers,
dates etc have been typed in incorrectly so that rejections aren't
as intended.
5. Through most of the period suspect reports/cruises were identified
by looking for large and "unphysical" anomalies from climatology
in fields of temperature and salinity averaged over the top 300 m
(and in some cases 3000 m). Both plotted monthly fields and
regional timeseries were used.
6. From about 1998 increasing numbers of profilers came to
dominate the rejection list (and precluded listing individual
identifiers in this table). From about 2003 profilers listed
in the Argo greylist were examined and most of them were rejected
- plus other profilers which were found to be erroneous or highly
suspicious, see the argo_notes file for more details.
The volume of profiler data in recent years probably masks some
errors from ship-based reports.
Bruce Ingleby 26 January 2007