Using Combined 532 NM HSRL and 1064 NM Elastic-Scatter Lidar Observations to Verify and Update CRAM Dual-Wavelength Aerosol Retrieval Models

The widely employed Fernald lidar equation solution relation retrieves aerosol backscatter versus height for an assumed aerosol extinction-to-backscatter ratio, S_a, and known system calibration factor, C, at a given wavelength, subject to the constraint/assumption that S_a is spatially constant over the solution layer (height range). At 532 nm, the calibration factor may be estimated fairly accurately by Rayleigh (molecular) normalization in high-altitude, clean-air regions. For elevated aerosol layers embedded in clean air, the direct transmittance approach may be used to estimate the optical depth of the layer, thereby giving an auxiliary input that permits S_a to be determined as a part of the solution. Otherwise, S_a must be specified based on models or determined through use of some other constraining information. The retrieval at 1064 nm is more problematic because the molecular scatter contribution is much weaker, typically making it difficult to either calibrate by molecular normalization or to accurately estimate the transmittance of an elevated aerosol layer, particularly for the weak signals from satellite lidar or eye-safe airborne lidar. As such, aerosol retrievals from dual-wavelength satellite lidars such as those onboard ICESat and CALIPSO have largely been limited to 532 nm retrievals using a table look-up approach to select assumed S_a values based on climatological/geographically determined models. Dual-wavelength retrievals using a Constrained Ratio Aerosol Model-fit (CRAM) retrieval method, employing AERONET based model ratio parameters, have also been obtained with some success for a few satellite lidar data sets.