A very preliminary, crude analysis of the effects of fringes on AIRS radiances is presented here. For two arrays, we selected by eye two SRFs that were ``reasonably'' different. Different means that the fringe peaks were offset different amounts from the SRF center. We made sure these two different SRFs had the same halfwidth. We then computed the array radiances twice, using these two different SRFs for each channel. This was performed for 6 widely varying profiles. Both the bias and rms differences are shown.
Although the rms error is small in the water region, this will be problematic since we cannot tune water channels with radiosondes at pressures lower than 400-500 mbar. Consequently, the bias errors in the water region are more indicative of the effects of fringe errors on retrievals.
We are presently modelling the effects of fringing by multiplying the Lockheed filter transmittance curves by ``perfect'' simulated SRFs. We are using these results to see how the fringing effects the Bruker SRF center/width/symmetry measurements, and how the movement of the fringes relative to the channel center frequencies will affect the measured radiances. Look back here tomorrow for preliminary results.
