2013-05-24 -- Actually, nevermind about the permanent discontinuation of SWPC processing, they are back.

2013-04-31 -- NGDC begins a new era of POES processing, a new directory structure will be phased in for all data.

2013-04-10 -- While the SWPC processing for all POES satellites was supposedly discontinued, NOAA-17 was decommissioned for real!

2013-01-01 -- POES/METOP Data Discontinued

NOAA's Space Weaher Prediction Center (SWPC) has discontinued processing of the POES data files.
SWPC says that production may resume at some later time but that nothing is definite at this time.
NGDC is working assuming the processing of these data in the future, the current projection is for the new processing to begin in March of 2013.

2012-03-12 -- POES/METOP Data Resumed

SWPC has resumed processing of the POES data files, NOAA data resume 3/8, MetOp data 3/9.

2012-02-29 POES/METOP Data Discontinued

NOAA's Space Weaher Prediction Center (SWPC) has discontinued processing of the POES data files.
SWPC says that production may resume at some later time but that nothing is definite at this time.

2011-07-11 POES/METOP files truncated

We have satellite data coming from 5 POES satellites and 1 MetOp satellite to NOAA/NWS/SWPC (Space Weather Prediction Center). There was a problem that the MetOp data was being truncated by as much as eight hours each day starting on 9 June 2011 as seen on the NGDC quality control plots. The problem was resolved and plots reflect this starting on July 5, 2011.

2011-04-25 POES files truncated

These POES files don't have an integral number of records.

PoesSummary says this file is incomplete poes_m02_20070308.bin (doy:067) size is 4956160 bytes, 448 leftOver bytes
PoesSummary says this file is incomplete poes_m02_20080221.bin (doy:052) size is 2686976 bytes, 512 leftOver bytes
PoesSummary says this file is incomplete poes_m02_20080912.bin (doy:256) size is 2678784 bytes, 2496 leftOver bytes
PoesSummary says this file is incomplete poes_m02_20081016.bin (doy:290) size is 4481024 bytes, 1040 leftOver bytes
PoesSummary says this file is incomplete poes_m02_20090204.bin (doy:035) size is 708608 bytes, 1376 leftOver bytes

NGDC has complete data files in RAW format, however, no access to the conversion software.

2009-04-24 Fortran Error Corrected in Unpacking Routine

A minor error has been uncovered in the FORTRAN version of the SEM-2 archive file unpacking routine (subroutine readArcSub ) that is detailed in Appendix E of the SEM-2 Instrument Descriptions and Archive Data Documentation. About 20 lines from the end of this FORTRAN subroutine code the statement: if(mdf(2,3).eq.1) tedback(2,3) = -999. should be replaced with if(mdf(26,j).eq.1) tedback(2,3) = -999. A revised Appendix E (Appendix E_rev1.pdf) is available and the code file has been revised (readArcSub_v2.f).

This error affected only the background readout from the 30-degree, 1-20 keV proton analyzer in the TED instrument compliment. This single data point is the tedback(2,3) entry in the tedback array returned from the unpacking routine. The error arises only in cases when that data point was lost in the processing of the data received from the satellite and a default pad value was inserted in its place. In those instances the unpacking routine did not access the missing data flag correctly and returned a value of 1998848.0 for that tedback entry rather than the correct value of –999.0 for missing data. This particular error occurred on the average of 1 or 2 times a day and did not affect any other data point in the SEM-2 archive data. Finally, this error was not present in the C version of the SEM-2 archive data unpacking routine.

2008-04-22 MepOp Orbital Maneuvers

It has come to our attention that during instances when the MetOp satellite exercises orbital maneuvers the entire SEM-2 instrument is turned off.  Unfortunately the data processing software does not recognize when the SEM-2 data processing unit (DPU) has been turned off and so assumes that the data received from the TED and MEPED detectors, that had been turned off, are valid in spite of the fact those data are zero-fill during these periods.  The TED and MEPED instrument ON-OFF flags in the daily archive files are not properly set to OFF at these times.

Those periods of time during which the SEM-2 instrument on MetOp were turned off are:

April 19, 2007 between 0819:58 and 1616:14 UTC
April 24, 2007 between 0000:00 and 1359:27 UTC
July  12, 2007 between 0922:22 and 1538:22 UTC
Sept. 18, 2007 between 1454:10 and 2400:00 UTC
Sept. 19, 2007 between 0000:00 and 1402:58 UTC
Jan.  31, 2008 between 0921:30 and 1538:33 UTC
April  8, 2008 between 0732:23 and 2400:00 UTC
April  9, 2008 between 0000:00 and 0437:58 UTC

The TED and MEPED ON-OFF flags in the MetOp archive files for these days have been corrected and the corrected files are available.

2008-01-07 Revised NOAA-16 file for day 2007-07-09

It was discovered that orbital data contained in the NOAA-16 archive file for July 9, 2007 (day 190) was incorrect for times after 2035:11 UT. A revised NOAA-16 archive file for that day has been produced that eliminated records with incorrect orbital information.

2007-03-23 Revised NOAA-18 archive files

Revisions to NOAA-18 archive data files for July 23-24, 2005 and for the entire period July 11, 2006 through December 26, 2006 were necessary because incorrect calibration data for the TED instrument were used.  These revisions impacted only the TED auroral particle energy flux values and many of the data quality flags.  All of the MEPED data in the original archive data files were correct.

Users of TED data for these periods are advised to download the revised archive data files for NOAA-18.

2005-08-16 ERRATA: Software errors in unpackSem2.c since year 2000.

It has been brought to our attention by a user of POES SEM-2 data that the C language archive unpacking routine returns anomalous sensor values during periods where there are data missing as flagged by 999. We have examined that code and uncovered two significant errors in the C version of the code that have existed since 2000.

Both errors involve the way the C unpacker implements missing data flags. The first error centered around the fact that missing data flags associated with sensor data from the second half of the 32-second archive record were being improperly unpacked from the binary archive record. The consequences of that error were that TED and MEPED sensor values during periods when data were missing were often improperly flagged with valid data being set to -999. indicating missing data, while entries that should have been flagged as missing data were given count values of 19988488, the highest count value possible.

The second error also involved the implementation of missing data flags and impacted P8 and P9 omni-directional detector data throughout an entire 32-second archive record. The P8 and P9 data share the same data word in the TIROS data format and readouts alternate between the two. In the archive record the data listing for P8 alternates between valid data readouts and values set to 0 in the archive format while the P9 data alternates in the same fashion but with valid readouts when the P8 value is zero. The existing documentation explains this and a user is expected to extract only the appropriate entries when using P8 and P9 data. During instances when data were missing, the C unpacker assigned the missing data flag (-999.), in both the P8 and P9 data, to the entry appropriate to 'no data' while the entries appropriate to valid data were given count values of 19988488. This count value is unphysical for P8 and P9 data.

Both these errors would have impact upon any analysis that had not recognized these unphysical sensor count.

The FORTRAN version of the archive record unpacker did not have either of these problems.

Corrected sofware (unpackSem2_v2.c) is now available for distribution.