Quality control (QC) procedures developed and implemented at AOML for the examination of real time XBT data are presented. The steps required to QC XBT data is outlined in a "cook book" format. The methods employed are primarily subjective as many of the stages involve interactive input from the user.
AOML is one of several Research Science Centers participating in a global
effort to quality control oceanographic data collected from either research
vessels or volunteer observing ships (VOS). AOML efforts are currently focused
on XBT data collected in the Atlantic Ocean. The data being examined consists
of "real-time", delayed, historical and Navy declassified modes.
The "real-time" data represents data obtained by the MEDS offices in Canada
from various ships. This data has already been quality controlled by the MEDS
office; nevertheless, AOML examines the data, without prejudice to the Canadian f
lags, and then compares its quality control flags against the Canadian flags.
The delayed mode data represents both updated profiles submitted as part of the
"real-time" data set and data that was not transmitted as "real-time" for 1990
. The historical data represents all of the XBT data that has been archived by
NODC between 1966 and 1989. The Navy declassified data is the data recently
released by the Navy covering the period 1985 through 1990.
The discussion that follows gives an overview of the procedures that are used
to quality control any of the XBT data sets. For exemplary purposes, the "real-ti
me" data set is discussed. NODC first obtains a data set from the MEDS office
. AOML then receives an electronic mail message from NODC indicating that a
particular month of data is available. AOML then copies the MEDS data over
the SPAN network. The MEDS file then rewritten into AOML's Indexed Sequential
Access Method (ISAM) Database. This is a direct access "keyed" database for
which XBTs are retrievable according to various keys, including: Latitude,
Longitude, Time, NODC_Unique_ID_Number, NODC_Cruise_Consec_Number,
MEDS_Station_Number, Ship_Radio_Call_sign.
Once the database has been updated, several programs are run to examine the
XBT profiles. The quality control procedures presently employed consists of
the following:
In general, the AOML procedures can be characterized as more subjective than
objective. There are no automatic flags established by any of these tests.
Profiles that have been identified as suspicious are examined individually and
flagged as appropriate. A flowchart and description of the programs used at
AOML, to insure quality control of the XBT database in the Atlantic, follows.
This program is used to get a detailed listing of a MEDS format file received
from NODC. Each field of every entry is listed in the output file. It counts
the different types of entries so that we may compare the number with the numbe
r expected.
It also does some error checking to assure that the
MEDS2AOML program will be able to add the new profiles to the existing
database.
MEDS - MEDS format file copied from NODC over the network.
PM - A listing file showing every field in the MEDS file from NODC.
MEDS2AOML reads an NODC MEDS format file and adds the XBT profiles to the
appropriate AOML ISAM database. TESAC messages are not stored.
XBTs not in AOML's defined Atlantic Region are not stored.
XBTs are stored in Indexed Sequential Access Method (ISAM) files which allow
access by various keys including NODC and AOML profile numbers, time,
position, or ship radio call sign.
The CRUISE file is appended to add the cruises in this new set of data.
A "relationship" (RELAT) file is written which shows the relationship between
the incoming XBTs and where they are stored in the database. This relationship
is important to reconstruct datasets that need to be returned (with Quality
Control information) to NODC.
MEDS - MEDS format file copied from NODC over the network.
AXBT - AOML XBT database (ISAM file).
FIND_DUPS reads the XBTs stored in Indexed Sequential Access Method (ISAM)
files to find potential duplicate profiles. They are read according to the
time key field. Profiles within a specified position and time space are
listed in an output data file as possible duplicates.
Typically, the XBTs within plus or minus 25 hours, and plus or minus .125
degrees (centered on a quarter of a degree) in latitude of longitude are
checked.
AXBT - AOML XBT database (ISAM file).
FD - ASCII file containing profile numbers of possible duplicate
XBT profiles. This file will be input to the PL_DUPS
program.
This program is used to determine graphically if the profiles identified by
the program FIND_DUPS are in fact duplicate
profiles. The operator has the ability to accept or reject a profile as a
duplicate.
Graphical output includes a plot of temperature vs depth and header
information for all profiles identified as possible duplicates.
AXBT - AOML XBT database (ISAM file).
PD - ASCII file containing profile numbers and associated profile
number for which the listed profile is considered a duplicate. That
is, the first profile number will be considered the actual profile and all other profiles are considered
as duplicates.
PUT_DUPFLAG reads a PD (ASCII) file, which contains the information about
duplicate profiles. The XBT (stored in Indexed Sequential Access Method,
ISAM file) is read in using its database profile number. The profile is
marked as a duplicate and a pointer to the primary XBT (of which this XBT is
considered a duplicate) is added and the XBT is rewritten back to the ISAM
database.
AXBT - AOML XBT database (ISAM file).
AXBT - AOML XBT database (ISAM file).
This program compares individual profiles to statistics derived from the
historical NODC database. Means and standard deviations have been estimated
for those 5 degree longitude/latitude boxes with sufficient data to make these
computations. The center of the 5 degree boxes begin at 97.W;37.5S and end at
17.5E;57.5N. The operator has the option of selecting which profiles to review
by specifying an acceptable range (sigma tolerance) within which the real time
data can fall.
Profiles can be also tested against the mean Levitus 1 degree climatology by
specifying acceptable temperature tolerances around this means. This is
useful in areas where there are insufficient XBT data to compute the
historical means and standard deviations.
This program is typically run using a 3 sigma tolerance about the historical
mean and 6 C tolerance about the Levitus 1 degree climatology. A profile is
displayed if the temperature at any depth in a profile differs from the mean
profile or Levitus climatology by more that the specified tolerance. The
operator has the option to mark questionable profiles and save them to a
"warning" file which is then used as input to other quality control programs
such as PL_CRUISE and
XBTEDIT .
Graphical output for each 5 degree box includes the
following :
AXBT - AOML XBT database (ISAM file).
(*) Two means are available. The first is the mean
calculated directly from the AOML historical XBT data base (1966-1988). The
second is a mean (clean mean) which is recalculated after
first discarding all profiles that are more than 2 standard
deviations from the historical mean.
SWARN - SIGMA warning files to identify suspicious
profiles.
This program examines all of the XBTs associated with a specific cruise.
Profiles, as well as time and position locations, are checked for consistency
with surrounding XBTs. In addition to identifying individual problem profiles,
this program also provides for a useful test for quality of data received
from entire VOS merchant ship tracks.
Questionable profiles identified from the SIGMA
program serve as input to aid in the decision making process. All profiles
identified as questionable are saved to "warning" file and later input into
the XBTEDIT program. Questionable time and
positions are saved to a "position" file for updating the flags in the data
base.
The program can be run either an interactive or hard copy mode. Typically,
hard copy output is generated for all cruises and examined. Then the program
is run interactively to further examine and identify the profiles for specific
cruises that are suspicious.
Graphical output includes the following:
AXBT - AOML XBT database (ISAM file).
PWARN - PL_CRUISE warning files to identify suspicious
profiles.
PUT_POSFLAG reads a POS (ASCII) file, which contains the information about
position or time corrections for an XBT profile. This file was created by the
PL_CRUISE quality control plotting program.
The XBT (stored in Indexed Sequential Access Method, ISAM file) is read in
using its database profile number. The profile is updated with the corrected
position and/or time and the XBT is rewritten back to the ISAM database.
AXBT - AOML XBT database (ISAM file).
AXBT - AOML XBT database (ISAM file).
This program interactively examines and edits any XBT profile. Suspicious
profiles from previous tests are highlighted so that the operator knows what
specific test a profile has failed. The historical mean climatology (5 degree
box) and the Levitus climatology (1 degree box) are plotted to the screen as
an additional aid in determining the quality of a profile. The operator can
also request information about surrounding profiles from the cruise that
containing the profile or historical data.
XBTEDIT can be run in four editing modes. The first is the initial
editing stage where flags can be set for any profile that contains bad data.
Data can be flagged as inconsistent (2), doubtful (3) or bad (4).
One of the following 12 reasons (Bailey et al. (1994); Figures 8a-8l) is
given for each flag value of 3 or 4:
There is obviously considerable subjectivity between flags 2 and 3. The other
editing modes allow for re-examination of previously flagged profiles,
comparison of the AOML and Canadian flags and comparison of the flags
generated by any 2 different operators.
Graphical options in the program give the operator full flexibility when
examining a profile. These include the ability to rescale, zoom in and out of
features and compare the profile with neighboring profiles from the same
cruise.
AXBT - AOML XBT database (ISAM file).
(*) Two means are available. The first is the mean calculated
directly from the AOML historical XBT data base (1966-
1988).
The second is a mean (clean mean) which is recalculated
after
first discarding all profiles that are more than 2
standard
deviations from the historical mean.
XED - ASCII file containing profile numbers and quality control
flags associated with depth-temperature pairs in the
profiles.
This program generates values and associated flags for the sea surface
temperature (SST), the temperature at 150m (T150), the temperature at 400m
(T400), the temperature at 600m (T600) and the average temperature of the
upper 400m layer (T400L) from either the AOML or Canadian database.
AXBT - AOML XBT database (ISAM file).
SST - File containing values of SST, T150, T400, T600 and
T400L and related XBT header information.
This program interpolates the sea surface temperature (SST), the temperature
at 150m (T150), the temperature at 400m (T400), the temperature at 600m (T600)
and the average temperature of the upper 400m layer (T400L) onto a 2 degree x
2 degree grid within the region (100W-20E; 40S-60N) and contours the resulting
fields (Figures 10,11,12 for example) in order to identify suspicious profiles
(i.e. those producing "bulls-eyes"). The location of each data value is
indicated by an x. Data values with a flag of 2 are double size. At each
location a '+' indicates a positive anomaly and 'o' a negative anomaly from
the Levitus 1 degree climatology. Only data values that have received a flag
of acceptable (1) or inconsistent (2) are used as input for the interpolation
to define additional erroneous profiles.
SST - File containing values of SST, T150,T400, T600 and
T400L and related XBT header information.
This program will find all XBTs that fall inside of a circle of specific
radius, around a specific location. It is used to identify those XBTs that
may have caused "bulls-eyes" to appear on the contours maps generated by
INT2X2. Those XBTs that are identified as
suspicious are re-examined by running XBTEDIT
and are flagged if appropriate.
SST - File containing values of SST, T150, T400, T600 and
T400L and related XBT header information.
PUT_FLAG reads a XED (ASCII) file, which contains the flags to be added to the
database as permanent Quality Control flags.
The XBT (stored in Indexed Sequential Access Method, ISAM file) is read in
using its database profile number, which is listed in the XED file. The QC
flags for that profile, from the XED file, are added and the XBT is rewritten,
with the added quality control flags, back to the ISAM database.
AXBT - AOML XBT database (ISAM file).
AXBT - AOML XBT database (ISAM file).
This program will generate a cruise report for any cruise in the database.
The report is based on a specific month and year and contains the following
information :
AXBT - AOML XBT database (ISAM file).
RPT - Cruise report file for the year and month
selected.
QC_RETURN creates a file which will be used to return RSC Quality Controlled
data to NODC.
QC_RETURN reads an original NODC MEDS format file looking at each XBT profile
in the order of the original MEDS file.
The appropriate "relationship" (RELAT) file is read which shows the
relationship between the incoming MEDS XBT and where it was stored in the AOML
database. This cross-reference is used to retrieve the related XBT (stored in
Indexed Sequential Access Method file) from the AOML database.
The QC flags and other appropriate QC fields are added to the original data
and the MEDS file is rewritten for return to NODC.
(NODC has requested that the output of this program be changed to be an ASCII
file which contains just the Quality Control recommendations from the AOML_RSC
QC procedures and personnel. Therefore the output of QC_RETURN is likely to
change.)
AXBT - AOML XBT database (ISAM file).
RSC - MEDS formatted file containing Quality Control
information to be returned to NODC.
II. INTRODUCTION
III. AOML QUALITY CONTROL PROCEDURES
Quality control flowchart
1) PRINT_MEDS
2) MEDS2AOML
AXBT - AOML XBT database (ISAM file).
CRUISE - ISAM data file containing a list of cruises and profiles
in each cruise.
CRUISE - ISAM data file containing a list of cruises and profiles in each
cruise.
RELAT - ISAM data file which contains a cross-reference
between the incoming NODC designators for each XBT
profile and where the XBT was stored in the database.
YYMM - A listing file showing every field in the MEDS file from
NODC. (YYMM is the designator for the data set,
usually Year and Month.)3) FIND_DUPS
4) PL_DUPS
(Figure 1, Figure 2)
FD - ASCII file containing profile numbers of possible duplicate
XBT profiles.5) PUT_DUPFLAG
PD - ASCII file containing profile numbers and associated
profile number for which the listed profile is considered a duplicate. That is,
the first profile number will be considered the actual profile and all other
profiles are considered as duplicates.6) SIGMA
5DEG - Five degree historical mean temperature
profiles (*).
LEV1 - Levitus one degree climatology.
LEV5 - Levitus five degree climatology.
5DEGBOX -Identifier for all profiles in a 5 degree box.7) PL_CRUISE
CRUISE - ISAM data file containing a list of cruises and
profiles in each cruise.
SWARN - SIGMA warning file
identifying suspicious profiles.
POS - PL_CRUISE flag file identifying a position or
time field.8) PUT_POSFLAG
POS - ASCII file containing profile numbers and associated
corrected position and/or time.9) XBTEDIT
(Figure 9)
5DEG - Five degree historical mean temperature
profiles (*).
LEV1 - Levitus one degree climatology.
SWARN - SIGMA warning file
identifying suspicious profiles.
PWARN - PL_CRUISE
warning files identifying suspicious profiles.10) WR_SST
XED - ASCII file containing profile numbers and quality
control
flags associated with depth-temperature pairs in the
profiles.11) INT2X2
12) FIND_XBT
13) PUT_FLAG
XED - ASCII file containing profile numbers and quality
control flags associated with
depth-temperature pairs in the profile.14) CRUISE_RPT
This is particularily useful to identify problem VOS ships and to notify them
so that problems with the "real time" data can be corrected as soon as
possible.
CRUISE - ISAM data file containing a list of cruises and
profiles in each cruise.15) QC_RETURN
MEDS - MEDS format file copied from NODC over the network.
RELAT - ISAM data file which contains a cross-reference
between the incoming NODC designators for each
XBT profile and where the XBT was stored in the database.Additional plots:
Sea-surface temperature contour
150m temperature contour
Upper 400m average temperature contour
IV. ACKNOWLEDGEMENTS
The authors wish to thank Dr. Robert L. Molinari for his contribution to the
completion of this report and Rick Bailey and Gary Meyers for permission to
use figures from their CSIRO report. This work was supported in part by a
grant from NOAA's Long Term Ocean Observing Program.V. REFERENCES
Bailey. R.,A. Gronell, H. Phillips, G. Meyers, and E. Tanner (1994).
CSIRO cookbook for Quality Control of Expendable Bathythermograph (XBT) Data.
CSIRO Marine Laboratories Report No. 220, 75 pp. (In press).