| | |
|
| |
|
| |
|
| |
|
| |
|
| | PROFILE OZONE DERIVED ADJUSTMENTS |
| |
 |
Assumptions |
The technique used to merge the data sets into an internally-consistent record rely on two primary assumptions:
We use the Earth Probe TOMS calibration from launch through summer 1999 as a reference, and adjust all the other instruments externally to match that calibration. For trend calculations, an error in the Earth Probe TOMS calibration is of lesser importance, because the trend is independent of an absolute calibration error. It is more important for all the instruments to have a consistent time-dependent calibration. However, beginning in mid-1999, the Earth Probe TOMS data begin to drift low relative to NOAA 11 SBUV/2 data. This drift continued through the NOAA 11 and NOAA 16 time period, and is consistent against both instruments. We therefore interpret this as a drift in Earth Probe TOMS. We use Earth Probe TOMS data through August 1999 to complete two years of overlap with the NOAA 11 instrument.
We use the time-dependent SBUV and SBUV/2 instrument calibrations to establish the relative calibration of the merged data record. Because of limited overlap periods between instruments, we cannot use direct instrument comparisons to establish time-dependent changes (with the exception of N7 TOMS/N7 SBUV). Therefore we must rely on the internal instrument calibration. Establishing the long-term calibration is complicated by the failure of the primary onboard calibration systems for NOAA 9 and NOAA 11 as the instruments aged beyond their anticipated lifetimes. We do not use the early portion of the NOAA 9 SBUV/2 data because of a known calibration offset between the afternoon and morning portions of the orbit. For more details on the calibration of each instrument, see SBUV Algorithm & Calibration.
| |
Methodology |
For each instrument, we construct 5° zonal mean, and 5° by 10° and 10° by 30° gridded monthly-mean data sets. Details of the averaging procedure and filtering requirements can be found under Averaging Details. In this version we do not use coincident matches to compare the data, but instead compare the zonal and gridded monthly averages computed from all available data for each instrument.
| N7 TOMS-N7
SBUV |
N7
TOMS-N11 SBUV/2 |
|---|
| EP
TOMS-N9 SBUV/2 |
EP
TOMS-N11 SBUV/2 |
EP TOMS-N16
SBUV/2* |
|---|
| N16 SBUV/2 - OMI
TO3 |
N16
SBUV/2 - N17 SBUV/2 |
|---|
In general, differences in the TOMS/SBUV gridded averages have a distinct longitudinal pattern, which would incorrectly alias into a simple latitude-dependent adjustment. Therefore we use the 50°S-50°N average difference in lieu of a latitude-dependent adjustment. The 50°S-50°N mean is calculated each month by averaging the 5° zonal mean differences.
To determine the external calibration for each instrument, we use the difference of the TOMS/SBUV 5° zonal monthly means during the overlap period averaged between 50° South and 50° North latitude. That is, the external calibration consists of a single additive adjustment that is applied at all times and latitudes.
Time series of TOMS-SBUV Mean Differences Before External Calibration
Specifically, we make the following adjustments:
Earth Probe TOMS measurements overlap with the morning portion of the NOAA 9 and NOAA 11 data records, while the afternoon portion of each satellite's orbit overlaps the Nimbus 7 TOMS data. In the Version 8 MOD data set we do not include the afternoon portion of the NOAA 9 data, so we use the NOAA 11 data to connect the Nimbus 7 and Earth Probe TOMS data to a common calibration.
We use the adjusted N16 SBUV/2 data as the calibration basis
for the instruments originating after 2000. We use the mean overlap
bewteen OMI TO3 and NOAA 16 over the period August 2004 through June
2007 to determine the offset for OMI TO3 data.
We
adjust the NOAA 17 SBUV/2 data directly to the N16 data using the
instrument overlap from August 2002 through June 2007. As the NOAA-16
satellite approaches the terminator, we will use N17 data to set the
calibration for future instruments, include NOAA-18 SBUV/2.
In the previous MOD data set, we found that the NOAA 9 and NOAA 11 gave inconsistent results as to the offset between Nimbus 7 TOMS and Earth Probe TOMS, adding an element of uncertainty to the combined data set. More recent studies have shown an unexplained calibration shift in the NOAA 9 data between the afternoon and morning portions of the orbit, and the decision was made to not use the early portion of the NOAA 9 data record. Therefore in this version of the MOD data set we only use the NOAA 11 data to determine the offset of Nimbus 7 TOMS relative to Earth Probe TOMS. Nimbus 7 TOMS overlaps the NOAA 11 SBUV/2 from 1989 through early 1994. The Nimbus 7 TOMS are adjusted to match on average the NOAA 11 SBUV/2 data (which are already calibrated to Earth Probe TOMS).
The Nimbus 7 SBUV data are adjusted to match the Nimbus 7 TOMS data using the long period of overlap between these instruments. In the previous MOD data set, a time-dependent offset was applied to the TOMS data to best match the SBUV, but advances in the Version 8 algorithm have eliminated the need for a time-dependent correction in this version.
Time series of Avg. TOMS-SBUV Differences After External Calibration
Once all the data sets have been adjusted to be consistent
with
the calibration of the Earth Probe TOMS instrument, we average the data
together
to construct a single merged data set.
If you would like more information about the construction of the MOD Total Ozone Data Set, Production Notes are available by request. Please contact Stacey Frith (Stacey.M.Frith@nasa.gov).
| | |
|
| |
|
| |
|
| |
|
| |
|
| | PROFILE OZONE DERIVED
ADJUSTMENTS |
| |
 |
Responsible NASA
official
& Website Design Dr. Richard Stolarski and Stacey Frith |
