CHARMS--Calibration of the 1064 nm HSRL channel

Description

The CHARMS study seeks to derive aerosol properties using a combination of HSRL and Raman lidar data. The SGP Raman lidar provides elastic backscatter profiles at 355 nm and nitrogen Raman profiles at 387. While the UW HSRL provides calibrated 532 nm extinction and backscatter profiles along with an attenuated backscatter profile at 1064 nm. The 1064 nm channel is a new addition to the UW HSRL and is the subject of this paper. The HSRL isolates molecular scattering from particulate scattering at 532 nm. The separate molecular profile allows robust calibration of both backscatter and extinction profiles. However, a separated molecular profile is not available to calibrate the 1064 nm measurements and other methods must be used to provide the 1064 nm channel gain and to correct for extinction. The 1064 nm return does not contain sufficient information to correct for extinction without making assumptions about the aerosol properties that we are attempting to derive. Incorporating a forward model of the 1064 nm lidar return in the inversion model could mitigate this problem. However, the current model requires a calibrated, attenuation corrected backscatter profiles as an input. Fortunately, attenuation is relatively small in most of the CHARM data. Calibration of channel gain is more difficult. Traditionally, both the gain calibration and attenuation correction are provided with a Klett solution. A clean layer is selected in each profile and the backscatter cross-section at that level is assumed equal to the molecular backscatter backscatter cross-section. A Klett solution using an assumed value of the aerosol extinction to backscatter ratio then provides the backscatter profile. Unfortunately, it is difficult to determine if the selected reference level is truly devoid of aerosol scattering. When the attenuation is small, as it is in most of this data, the error in the derived backscatter is directly proportional to the ratio of the total backscatter to the molecular backscatter at the reference level. An alternate approach uses measurements of the relative gains of the 1064 nm and 532 nm channels made over a longer period of time. This has the advantage of providing a greater likelihood of finding a clean layer for the calibration, but is then sensitive to calibration changes with time. In the absence of a forward model to account for attenuation, an angstrom coefficient can be assumed to translate the 532 nm extinction coefficient to 1064 nm.