MEDS Analysis of Atlantic DAC flags
Introduction
MEDS has been looking closely at a file received from AOML to see what are the differences. To do this, we have written software that reads the incoming data from AOML, locates the same station in our own archives and compares the contents byte for byte. In the discussion below we have ignored changes such as the inclusion of a record of the tests applied by AOML, and history records created by them.
There were 1424 stations in the incoming file covering November 1990 to January 1991. There were 323, roughly 22%, which had differences in either date, position, record flag or profile flags. There were 28 stations that showed position differences, 1 that had a date change, 5 with record flag changes and 320 with profile flag changes. There were 46 stations that had other changes.
In the discussions that follow, flag values are referred to by the coded values present in the file. So a flag of '1' means that the data appear to be correct, '2' means that the data are inconsistent, '3' means the data are doubtful and '4' means the data are wrong. MEDS assigns a flag of '2' only if an observation fails a test against the Levitus climatology. Note also that where percentages are quoted later on, these refer to the 320 stations which showed differences between AOML and MEDS records.
Differences in the Header
We can explain some of the differences found with some general observations. First, contained in MEDS files are records from the same ship, but that came in after the data were sent to AOML. We know this because the time of receipt stored in the BUL_TIME field of the record. In many cases the first message arrived at MEDS, was processed and sent to AOML. Later, a second message arrived, sometimes a few days later, and because data can be retransmitted on the GTS with corrections, the later message took precedence over the earlier one. We have the later message in our archive, while the file from AOML had the earlier one. To some extent, this is a feature of a continuously managed database; it is always changing. We believe the number of differences in this comparison is a consequence of how NODC used to manage the data files from MEDS. Now that the data are entering their database immediately, such situations would only occur near the start and end of months. We do not expect it would be a major problem.
We found 8 stations for which the ship call sign in MEDS files had changed from that present in the AOML data file. This arises from the way MEDS carries out its QC work. So, it may take a few days to recognize a damaged callsign, and therefore make the correction. In that time, the data has been passed to NODC and on to the science centres. It points out once again that the concept of a continuously managed database will inevitably lead to these sorts of differences.
Changes in position can be accounted for by QC operations at MEDS. Because of the way data arrive at MEDS, it is not always possible to know if a station location is consistent with other stations until some days later when more data are received. We then are able to construct a sensible cruise track and certain stations received earlier stand out as having problems. In this case, we would make changes or flag positions or dates, or perhaps even delete stations. This type of occurrence accounted for all of the position differences between MEDS and AOML.
Profile Differences
The major difference between AOML and MEDS was in flags set on the profiles. Of the ones examined, about half of the stations had observations that had failed the climatology test at MEDS and the '2' flags were reset to '1' by AOML to indicate that the observations appeared to be okay. MEDS policy has been to use the flag of '2', which is only set by a failure of the Levitus climatology test, as an indicator of unusual but not necessarily suspect profiles. Generally, we would not reset the flag back to '1', but we would set the flag to '3' or '4' if the profile did appear to be more than just unusual. The Levitus statistics cannot be considered the final word in climatologies. However, since processing centres will be using different climatologies to judge data, it is important that these are well described. At some point, it would also be useful for an intercomparison to be made between these different climatologies.
The next major difference tended to be where AOML would change flags set by MEDS (usually set to '1') to '3' either at the top but more likely near the bottom of a profile. This occurred in about 30% of the stations. We have used software to display the target station on a background of other profiles from the same one degree square and in the same month or one month on either side of the month of the target station. Examples of this are shown in figures 1 and 2 and tables 1 and 2. We believe that the reason for the flag change in both cases is due to the increase in temperature at the bottom of the profile. For the second cruise cited, we would agree, but the first station was collected in waters which often have a cold intermediate layer. This is evident from some of the other profiles that do go deeper in the same region. When examining a single profile, it is necessary to exercise caution on the strict application of a temperature inversion test. We agree that a flag of 3 was cautious. This difference emphasizes that fact that users of the data must be careful in the interpretation of the flags that are set. In particular, a flag of '2' or '3' does not mean the data are wrong. Rather it is a statement that certain features are in question.
Another difference had AOML flagging values as '4' usually near the bottom of a profile where MEDS may have only flagged them as '3'. A variant of this was if the deepest observations in a profile were very much deeper than the previous observation, AOML tended to flag these deep observations with a '4' (see figure 3 and table 3). This happened at about 12% of the stations. MEDS would like to improve its handling of these situations to be more consistent with judgments of science centres.
After the above, there were various other conditions where flags were reset. Sometimes, a major portion of a profile was judged suspect and flags set (see figure 4 and table 4). This happened at about 5% of the stations. Other examples had AOML resetting MEDS flags from '1' to '2' or changing flags of '3' or '4' as set by MEDS to '1' (see figure 5 and table 5).
There were a few profiles where MEDS missed flagging observations which certainly would appear to have been suspect at least. This can happen where the feature in the profile which appears suspect is not caught by any test, and MEDS operators also do not set a flag. Fortunately, this does not appear to happen very often.
In some cases there are small features in a profile that AOML has judged suspect or worse, but which MEDS has left as flagged (or not) by the test procedures. One such example is shown in figure 6 and table 6). We prefer to continue this practice, relying on the science centres to make the determinations of data quality on the more subtle features of profiles.
AOML has reset some flags from '1', as set by MEDS to '2'. It was not obvious in the
stations examined what was the feature of the profile that caused the flag to be set
to '2'. An example of this was already shown in figure 4.
Conclusions
There were a significant number of stations where there were differences between MEDS files and the records returned from AOML. The largest proportion of differences is explained as climatology failures flagged by MEDS but reset to '1' by AOML. The next largest proportion of the differences are those where observations have not failed the objective tests and usually MEDS operators have not changed flags. In these cases, AOML has deemed the observations to be doubtful. The above two differences account for about 80% of the differences found. Smaller proportions of the differences were due to a variety of things including resetting MEDS flags of doubtful to okay by AOML.
Figure 1
Table 1
ONE_DEG_SQ 233135 CR_NUMBER CGDV 90 OBS_DATE 19901129 OBS_TIME 0006
DATA_TYPE BA STN_NUMBER 0 LATITUDE 44.4833 LONGITUDE 52.9000
Q_POS 1 Q_DATE_TIME 1 Q_RECORD 1 UP_DATE 19920426
BUL_TIME 199101172025 BUL_HEADER SOVD02 SOURCE_ID CWOW STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS -
NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 102.0
VECTOR OF 0 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARMVECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 00400000 0 QCF$ 00000000 0 QAO$ 0000001 0
VECTOR OF 13 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME IG02 1.0 19910128 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910204 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910205 DC RCRD 9999.999 0.000
ME IG05 1.0 19910208 QC RCRD 0.000 0.000
ME IG05 1.0 19910208 CF TEMP 100.000 1.000
ME IG05 1.0 19910208 CF TEMP 101.000 1.000
ME IG05 1.0 19910208 CF TEMP 102.000 1.000
ME IG05 1.0 19910208 QC RCRD 0.000 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF TEMP 99.000 1.000
AO XED1 1.0 19920426 CF TEMP 100.000 1.000
AO XED1 1.0 19920426 CF TEMP 101.000 1.000
AO XED1 1.0 19920426 CF TEMP 102.000 1.000
PROFILE_SEG 01 TYPE TEMP NO DEPTHS 11 D-P-CODE D
- DEPTH PRESS, DP FLAG, PROF PARM, PROF Q PARM
3.0 1 6.9000 1 51.0 1 6.9000 1 54.0 1 6.2000 1
57.0 1 4.3000 1 59.0 1 0.7000 1 64.0 1 -0.3000 1
93.0 1 -0.2000 1 99.0 1 0.3000 3 100.0 1 0.9000 3
101.0 1 1.4000 3 102.0 1 1.7000 3
Figure 2
Table 2
ONE_DEG_SQ 253130 CR_NUMBER PJJU 90 OBS_DATE 19901213 OBS_TIME 0946
DATA_TYPE BA STN_NUMBER 0 LATITUDE 39.5500 LONGITUDE 72.6000
Q_POS 1 Q_DATE_TIME 1 Q_RECORD 1 UP_DATE 19920426
BUL_TIME 199101131422 BUL_HEADER SOVD5 SOURCE_ID KWBC STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS
- NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 206.0
VECTOR OF 1 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARM
BATH 83.0000 1
VECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 0021FFDE 0 QCF$ 00000000 0 QAO$ 0000001 0VECTOR OF 8 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME IG02 1.0 19910128 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910204 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910205 DC RCRD 9999.999 0.000
ME IG05 1.0 19910208 QC RCRD 0.000 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF RCRD 192.000 1.000
AO XED1 1.0 19920426 CF RCRD 201.000 1.000
AO XED1 1.0 19920426 CF RCRD 206.000 1.000PROFILE_SEG 01 TYPE TEMP NO DEPTHS 10 D-P-CODE D
- DEPTH_PRESS, DP_FLAG, PROF_PARM, PROF_Q_PARM
3.0 1 12.2000 1 9.0 1 12.0000 1 72.0 1 12.1000 1
73.0 1 12.5000 1 75.0 1 13.4000 1 78.0 1 13.8000 1
115.0 1 13.2000 1 192.0 1 13.2000 3 201.0 1 14.8000 3
206.0 1 13.6000 3
Figure 3
Table 3
ONE_DEG_SQ 253130 CR_NUMBER PJJU 90 OBS_DATE 19901129 OBS_TIME 0041
DATA_TYPE BA STN_NUMBER 0 LATITUDE 39.3500 LONGITUDE 72.3000
Q_POS 1 Q_DATE_TIME 1 Q_RECORD 1 UP_DATE 19920426
BUL_TIME 199011291510 BUL_HEADER SOVD5 SOURCE_ID KWBC STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS
- NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 538.0
VECTOR OF 1 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARM
BATH 550.0000 1
VECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 00400000 0 QCF$ 00000000 0 QAO$ 0000001 0VECTOR OF 9 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME IG02 1.0 19901213 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910109 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910114 DC RCRD 9999.999 0.000
ME IG05 1.0 19910117 QC RCRD 0.000 0.000
ME OCUP 1.0 19910118 UP RCRD 9999.999 9999.999
ME IG04 1.0 19910114 DC RCRD 9999.999 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF TEMP 484.000 1.000
AO XED1 1.0 19920426 CF TEMP 538.000 1.000
PROFILE: SEG 01 TYPE TEMP NO DEPTHS 20 D-P-CODE D
- DEPTH_PRESS, DP_FLAG, PROF_PARM, PROF_Q_PARM
3.0 1 13.6000 1 4.0 1 13.2000 1 21.0 1 13.1000 1
27.0 1 13.9000 1 30.0 1 14.4000 1 39.0 1 14.7000 1
56.0 1 14.6000 1 59.0 1 13.9000 1 61.0 1 13.9000 1
67.0 1 14.1000 1 72.0 1 14.2000 1 91.0 1 13.3000 1
111.0 1 13.4000 1 129.0 1 13.1000 1 132.0 1 13.0000 1
136.0 1 12.4000 1 146.0 1 12.0000 1 184.0 1 10.8000 1
484.0 1 11.1000 4 538.0 1 11.1000 4
Figure 4
Table 4
ONE_DEG_SQ 174041 CR_NUMBER DBLK 90 OBS_DATE 19901225 OBS_TIME 1258
DATA_TYPE BA STN_NUMBER 0 LATITUDE -49.7500 LONGITUDE -6.0167
Q_POS 1 Q_DATE_TIME 1 Q_RECORD 1 UP_DATE 19920426
BUL_TIME 199101021054 BUL_HEADER SOVF01 SOURCE_ID EDZW STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS
- NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 753.0
VECTOR OF 0 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARMVECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 0021FFDE 0 QCF$ 00000000 0 QAO$ 0000001 0VECTOR OF 34 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME IG02 1.0 19910128 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910204 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910205 DC RCRD 9999.999 0.000
ME IG05 1.0 19910211 QC RCRD 0.000 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF TEMP 14.000 1.000
AO XED1 1.0 19920426 CF TEMP 16.000 1.000
AO XED1 1.0 19920426 CF TEMP 18.000 1.000
AO XED1 1.0 19920426 CF TEMP 55.000 1.000
AO XED1 1.0 19920426 CF TEMP 61.000 1.000
AO XED1 1.0 19920426 CF TEMP 62.000 1.000
AO XED1 1.0 19920426 CF TEMP 65.000 1.000
AO XED1 1.0 19920426 CF TEMP 153.000 1.000
AO XED1 1.0 19920426 CF TEMP 213.000 1.000
AO XED1 1.0 19920426 CF TEMP 214.000 1.000
AO XED1 1.0 19920426 CF TEMP 216.000 1.000
AO XED1 1.0 19920426 CF TEMP 219.000 1.000
AO XED1 1.0 19920426 CF TEMP 221.000 1.000
AO XED1 1.0 19920426 CF TEMP 324.000 1.000
AO XED1 1.0 19920426 CF TEMP 341.000 1.000
AO XED1 1.0 19920426 CF TEMP 395.000 1.000
AO XED1 1.0 19920426 CF TEMP 397.000 1.000
AO XED1 1.0 19920426 CF TEMP 398.000 1.000
AO XED1 1.0 19920426 CF TEMP 456.000 1.000
AO XED1 1.0 19920426 CF TEMP 461.000 1.000
AO XED1 1.0 19920426 CF TEMP 464.000 1.000
AO XED1 1.0 19920426 CF TEMP 466.000 1.000
AO XED1 1.0 19920426 CF TEMP 469.000 1.000
AO XED1 1.0 19920426 CF TEMP 470.000 1.000
AO XED1 1.0 19920426 CF TEMP 472.000 1.000
AO XED1 1.0 19920426 CF TEMP 495.000 1.000
AO XED1 1.0 19920426 CF TEMP 696.000 1.000
AO XED1 1.0 19920426 CF TEMP 704.000 1.000
AO XED1 1.0 19920426 CF TEMP 753.000 1.000
PROFILE: SEG 01 TYPE TEMP NO DEPTHS 29 D-P-CODE D
- DEPTH_PRESS, DP_FLAG, PROF_PARM, PROF_Q_PARM
14.0 1 5.0000 2 16.0 1 5.1000 2 18.0 1 5.4000 2
55.0 1 4.4000 2 61.0 1 4.7000 2 62.0 1 4.1000 2
65.0 1 3.9000 2 153.0 1 3.0000 2 213.0 1 2.9000 2
214.0 1 3.2000 2 216.0 1 3.2000 2 219.0 1 3.3000 2
221.0 1 2.9000 2 324.0 1 3.4000 2 341.0 1 3.3000 2
395.0 1 3.4000 2 397.0 1 3.2000 2 398.0 1 2.7000 2
456.0 1 3.7000 2 461.0 1 3.5000 2 464.0 1 3.5000 2
466.0 1 3.4000 2 469.0 1 3.7000 2 470.0 1 3.2000 2
472.0 1 3.5000 2 495.0 1 3.2000 2 696.0 1 4.0000 2
704.0 1 3.8000 2 753.0 1 3.4000 2
Figure 5
Table 5
ONE_DEG_SQ 232141 CR_NUMBER VC9450 90 OBS_DATE 19901209 OBS_TIME 1540
DATA_TYPE BA STN_NUMBER 0 LATITUDE 50.9500 LONGITUDE 51.5333
Q_POS 1 Q_DATE_TIME 1 Q_RECORD 2 UP_DATE 19920426
BUL_TIME 199101081942 BUL_HEADER SOVD02 SOURCE_ID CWOW STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS
- NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 251.0
VECTOR OF 0 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARMVECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 0021FFDE 0 QCF$ 00010000 0 QAO$ 0000001 0
VECTOR OF 16 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME IG02 1.0 19910128 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910204 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910205 DC RCRD 9999.999 0.000
ME IG05 1.0 19910208 QC RCRD 0.000 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF TEMP 0.000 2.000
AO XED1 1.0 19920426 CF TEMP 142.000 2.000
AO XED1 1.0 19920426 CF TEMP 173.000 2.000
AO XED1 1.0 19920426 CF TEMP 215.000 2.000
AO XED1 1.0 19920426 CF TEMP 222.000 2.000
AO XED1 1.0 19920426 CF TEMP 227.000 2.000
AO XED1 1.0 19920426 CF TEMP 235.000 2.000
AO XED1 1.0 19920426 CF TEMP 236.000 2.000
AO XED1 1.0 19920426 CF TEMP 237.000 2.000
AO XED1 1.0 19920426 CF TEMP 242.000 2.000
AO XED1 1.0 19920426 CF TEMP 243.000 2.000
AO XED1 1.0 19920426 CF TEMP 251.000 2.000PROFILE: SEG 01 TYPE TEMP NO DEPTHS 18 D-P-CODE D
- DEPTH_PRESS, DP_FLAG, PROF_PARM, PROF_Q_PARM
3.0 1 1.2000 1 50.0 1 1.2000 1 54.0 1 1.0000 1
66.0 1 0.8000 1 70.0 1 0.5000 1 77.0 1 -0.6000 1
96.0 1 -0.9000 1 142.0 1 -1.0000 1 173.0 1 -0.5000 1
215.0 1 -0.2000 1 222.0 1 -0.2000 1 227.0 1 0.2000 1
235.0 1 0.3000 1 236.0 1 0.4000 1 237.0 1 0.5000 1
242.0 1 0.6000 1 243.0 1 0.8000 1 251.0 1 0.8000 1
Figure 6
Table 6
ONE_DEG_SQ 174066 CR_NUMBER J8FO 90 OBS_DATE 19901225 OBS_TIME 0146
DATA_TYPE BA STN_NUMBER 0 LATITUDE -24.6300 LONGITUDE -6.5500
Q_POS 1 Q_DATE_TIME 5 Q_RECORD 5 UP_DATE 19920426
BUL_TIME 1991220210 BUL_HEADER SOVX21 SOURCE_ID EESA STREAM_IDENT AOBA
QC_VERSION DATA_AVAIL A
VECTOR OF 1 PROFILE DESCRIPTORS
- NO_SEG, PROF_TYPE, DUP_FLAG, DIGIT_CODE, STANDARD, DEEP_DEPTH
1 TEMP N 8 2 760.0
VECTOR OF 6 STATION LEVEL PARAMETERS - PCODE, PARM, Q_PARM
WDIR 160.0000 1 WSPD 8.0000 1 DRYT 21.0000 1
BATH 1000.0000 1 SCDT 40.0000 1 SCSP 0.5150 1
VECTOR OF 3 STATION CHARACTER FIELDS - SRFC_CODE, SRFC_PARM, SRFC_Q_PARM
QCP$ 0021FFDE 0 QCF$ 00010000 0 QAO$ 0000001 0VECTOR OF 11 HISTORY RECORDS
- IDENT_CODE, PRC_CODE, VERSION, PRC_DATE, ACT_CODE, ACT_PARM, AUX_ID, ORIG_VAL
ME BU01 1.0 19910125 CR RCRD 9999.999 9999.999
ME QCA1 1.0 19910130 ED DATE 9999.999 1224.999
ME QCA1 1.0 19910204 CR RCRD 9999.999 9999.999
ME IGO3 1.0 19910205 DC RCRD 9999.999 0.000
ME IG05 1.0 19910211 QC RCRD 0.000 0.000
AO QCR1 1.0 19920426 QC RCRD 0.000 0.000
AO XED1 1.0 19920426 CF TEMP 517.000 1.000
AO XED1 1.0 19920426 CF TEMP 569.000 1.000
AO XED1 1.0 19920426 CF TEMP 620.000 1.000
AO XED1 1.0 19920426 CF TEMP 701.000 1.000
AO XED1 1.0 19920426 CF TEMP 760.000 1.000
PROFILE: SEG 01 TYPE TEMP NO DEPTHS 25 D-P-CODE D
- DEPTH_PRESS, DP_FLAG, PROF_PARM, PROF_Q_PARM
3.0 1 20.6000 1 27.0 1 20.2000 1 31.0 1 19.6000 1
36.0 1 19.0000 1 43.0 1 17.6000 1 49.0 1 16.8000 1
52.0 1 16.3000 1 62.0 1 15.7000 1 71.0 1 15.2000 1
125.0 1 14.0000 1 167.0 1 13.2000 1 202.0 1 12.2000 1
270.0 1 11.3000 1 295.0 1 10.5000 1 340.0 1 9.4000 1
431.0 1 7.3000 1 455.0 1 6.9000 1 517.0 1 7.2000 3
569.0 1 7.1000 3 620.0 1 6.8000 3 701.0 1 6.1000 3
760.0 1 5.6000 3