SeaWiFS Lunar Calibration Time Series

Revised focal plane temperature corrections. Image generated March 17, 2009.

UPDATE ON the SeaWiFS CALIBRATION:

The performance of the SeaWiFS normalized water-leaving radiances have demonstrated remarkable consistency for most of the mission to date. Recently, analyses of the interannual repeatability have shown measurable deviations from the mission trends. Starting in late 2005, the global nLw averages have deviated from the previous results by typically 3% to 5%. The time series with the calibration drifts is shown here.

The cause of these deviations is a small drift in the SeaWiFS calibration, most notably for bands 7 and 8. This calibration drift is apparent as a periodic signal in the lunar calibration time series starting in mid-2005. The periodic nature of this drift has led the Cal/Val Team to examine thermal effects within the instrument to identify the source of the drift.

After extensive analysis of the lunar calibration time series, the Cal/Val Team has concluded that the sensitivity of the SeaWiFS detectors to changes in the focal plane temperatures has increased over time. This increase in sensitivity would give rise to a periodic calibration drift that corresponds to the varying Earth-Sun distance over the course of a year. The most likely cause of this increase in sensitivity is aging of the detectors on-orbit.

The Cal/Val Team has used the residuals from fits to the lunar calibration time series to compute a revised set of focal plane temperature corrections for each band. These revised corrections apply to data collected on or after 1 January 2006. The lunar time series with the revised temperature corrections shows that the periodic calibration drift in bands 7 and 8 has been reduced significantly. The time series with the revised focal plane temperature corrections is shown in the figure above.

GO HERE FOR A MORE DETAILED DISCUSSION OF THE REVISION TO THE SeaWiFS CALIBRATION.


CALIBRATION BACKGROUND

This plot of the current SeaWiFS lunar calibration time series shows how the radiometric response of the instrument changes with time. The vertical lines in the plot indicate January 1 of each year. The lunar observations for each band are fit by two simultaneous exponential functions of time. The solid lines are the fits contained in the current operational calibration table which are based on the 107 lunar observations obtained through October 2006.

The time series are fit by exponentials with a time constant of 200 days (the short period time constant) and with a time constant of 2500 days (the long period time constant). As an example, the short period and long period exponential fits for band 8 are shown here. The corrections to the SeaWiFS top-of-the-atmosphere radiances as a function of time, which are applied to the ocean data, are the inverses of the solid line fits shown above, computed for the times of the ocean retrievals.

Two different decay mechanisms are responsible for the changes in response of each of the bands. The first mechanism caused a rapid decrease in response that decayed away after approximately the first year of the mission. The second mechanism has been in effect over the entire mission. The long-period decay for the shorter wavelength bands (bands 1-4) most likely arises from yellowing of the instrument optics on orbit. The long-period decay for the longer wavelength bands (bands 5-8) most likely arises from charged-particle induced damage to the silicon photodiodes.

The SeaWiFS lunar calibration data have to be normalized to a common viewing geometry to produce the time series shown above. Corrections must be computed and applied for the Sun-Moon and instrument-Moon distances, the oversampling of the lunar images in the along-track direction, the phase angles of the observations, and the libration angles of the observations. These corrections are discussed in detail here.

References