Solar Backscatter UltraViolet (SBUV and SBUV/2)

 Algorithm and Calibration Details

Algorithm

Data from the SBUV series of instruments have been processed using the SBUV Version 8 algorithm. This algorithm has been optimized to produce a trend quality data set and maintain long-term calibration as best as possible. All the SBUV instruments preformed well over their 3-year planned lifetimes. As the satellites aged beyond their planned lifetimes, they continued to provide invaluable data, but various hardware issues have made long-term calibration difficult. We address some of these difficulties in the individual instrument sections below.

The SBUV instrument measures incoming solar irradiance and radiance backscattered by the atmosphere at 12 wavelengths in the ultraviolet range of the spectrum. Radiation as these wavelengths is absorbed by ozone, such that the difference between the incoming and outgoing radiation can be related to the amount of ozone in the atmosphere. Radiation at the eight shortest wavelengths is absorbed by atmospheric ozone before reaching the surface, implying that radiation scattered back to the instrument came from a particular altitude region. Radiation at successively longer wavelengths penetrates deeper into the atmosphere before being completely absorbed by ozone, allowing for a measure of the ozone profile. In the Version 8 algorithm, the total ozone is calculated as the sum of the retrieved profile ozone, rather than from measurements at the four longest wavelengths, which do penetrate to the surface. This makes the total ozone less sensitive to variations in surface reflectivity and scattering processes in the troposphere.

Improvements in the Version 8 algorithm (related to profile ozone) include:

• Lower total ozone sensitivity to atmospheric temperature, aerosols, clouds, and surface reflectivity
• A Priori derived from a time-independent climatology, eliminating a priori information as a source of trend
• Updated a priori profile ozone climatology, including tropospheric climatology
• Improved multiple scattering and cloud/reflectivity modeling
• Improved terrain height information
• Use only necessary wavelengths to derive ozone profile, thus limiting longer wavelength measurements which are more likely to be affected by other scattering processes.
  
• Included latest calibration corrections for instrument hardware issues

For more information on the SBUV Version 8 algorithm, please see the SBUV V8 Algorithm Description, also available from the Version 8 SBUV DVD.

Calibration

Incoming solar irradiance measurements and backscattered terrestrial radiance measurements at wavelengths absorbed by atmospheric ozone are made by the SBUV instrument, and the ratio of these is used to determine the amount of ozone in the atmosphere. The solar irradiance and backscattered radiance are measured using the same optical components,  so any calibration error in the optics will cancel in the ratio of radiance to irradiance measurement. However, the solar diffuser mirror, used to reflect diffuse sunlight into the instrument optics, is only used in the solar irradiance measurement, and is thus the primary source of time-dependent changes in instrument calibration. The optical properties of the diffuser mirror change with time as a function of cumulative exposure to the sun. The SBUV series of instruments includes an on-board calibration system to monitor changes in the diffuser mirror over time. As the instruments aged beyond their planned lifetimes, these systems failed to work adequately for various instruments/times, requiring the development of other more complicated calibration methods.

In Version 8, the calibration of each SBUV instrument is first determined using pre-launch and inflight calibration data. Then a final absolute calibration is derived based on intercomparison of data from the different instruments. The NOAA 11 SBUV/2 absolute calibration is derived from overlap comparisons with measurements from the Shuttle SBUV (SSBUV) instrument. The calibrated NOAA 11 data are then used to determine the absolute calibration of the Nimbus 7 SBUV and NOAA 9 SBUV/2 instruments according to their respective overpass comparisons. The NOAA 16 SBUV/2, which do not have sufficient overlap with NOAA 11, are processed based on inflight calibration data. The most recent calibration update is based on data collected through August 2007. By adjusting the absolute calibration based on instrument intercomparisons within the processing, the Version 8 algorithm internally makes adjustments similar to those made when creating the MOD data set. This result is reflected in the generally lower external adjustments derived for Version 8 data in the MOD data set. However, we also include TOMS data in the MOD analysis, so further adjustment is possible based on the addition of more data.

NOAA 9 SBUV/2:
From launch in 1985 through 1989, the NOAA 9 satellite drifted from an early afternoon orbit to a near-terminator orbit. From 1990 through 1992 the orbit drifted through the terminator and into a morning orbit, and by 1993 reliable measurements had resumed.  Subsequent analysis indicates a unexplained calibration shift in the NOAA 9 data as the orbit processed through the terminator. The calibration during the morning portion of the orbit, from 1993-1997, is established based on overlap comparisons with the NOAA 11 SBUV/2 data set. Therefore we only use NOAA 9 data after 1993 in this analysis. After June, 1997 the NOAA 9 instrument lost significant longitudinal coverage and the data are not used.

NOAA 11 SBUV/2:
From launch in 1989 through late 1994, the NOAA 11 satellite drifted from an early afternoon orbit to a near-terminator orbit. From 1995 through 1997 the orbit drifted through the terminator and into a morning orbit, and by August 1997 reliable measurements had resumed. Grating drive errors began affecting the NOAA 11 instrument in 1993, and worsened in the late 1990s. This period of data is difficult to calibrate, but is necessary to cover the period after N9 SBUV/2 (profile) and EP TOMS (total) until the launch of NOAA-16 SBUV/2.

NOAA 16 SBUV/2:
The NOAA 16 satellite was launched on September 21, 2000. NOAA 16 SBUV/2 became operational in March, 2001. However, test data from October 2000 through February 2001 are available. These data are used in our analysis to optimize the length of the overlap period between NOAA 16 and NOAA 11 SBUV/2.  NOAA 16 began experiencing electronic interference leading to enhanced noise in measurements used to calculate total ozone, but total ozone derived as the sum of the profile ozone is not sensitive to these errors. These errors have diminished substantially since February 2004. The NOAA-16 data from the beginning of the record have been reprocessed using the best calibration information available through August 2007. Unlike the previous NOAA SBUV/2 satellite series, NOAA 16 was placed into a more stable orbit. The orbit remained stable through 2003, then began to slowly drift, but now that drift is accelerating rapidly. We do not use NOAA 16 data after June 2007 as the measurements are affected by higher solar zenith angle conditions, particularly in the Southern Hemisphere.

NOAA 17 SBUV/2:
The NOAA 16 satellite was launched on June 24, 2002 and measurements begin in August. NOAA-17 was also launched into a more stable orbit, and the orbit remains optimal. 

Responsible NASA official & Website Design Dr. Richard Stolarski and Stacey Frith