CALIPSO, CIMSS

Methodology

The CALIPSO lidar CALIOP provides measurements of backscattered power, the calibrated data available in the level 1B data files is attenuated backscatter and is defined as:

where , and , are the backscattering coefficients for molecules and aerosols, transmittances do to molecules (m), aerosols (a), and ozone (O3), respectively. Note that the aerosol term can be a stand in for any particulate scattering, i.e. aerosols or hydrometeors. Operating the wavelet transform on this profile data is inconvenient because of the molecular attenuation of the lidar signal (i.e. the term). Instead we divide the attenuated backscatter by the molecular scattering coefficient and transmittances to compute a profile of attenuated scattering ratio as:

In regions where there is little or no aerosol (particulate) scattering , in the PBL region where the scattering from aerosols is significant, , see Figure 1 below 1.7 km in altitude. , and are computed for each profile using the GEOS (need version) temperature, pressure, and ozone number density data that is provided in the CALIPSO lidar level 1B files.

Figure 1 Attenuated scattering ratio computed CALIPSO lidar level 1B attenuated backscatter data from 9/10/2010 near Chicago, IL.
The wavelet covariance transform is computed for each lidar profile , as the following:

where , and are the bottom and top of the lidar measurement profile, and h is the Haar function:

where z is the altitude, b is the translation, and a is the dilation. Before the wavelet covariance transform is calculated it is critical that we mask out the lidar surface return signal, if left in the lidar signal it will completely overwhelm the transform signal, making the PBLH retrieval extremely difficult or impossible. Currently the covariance transform is computed over a range of dilations from 0.9 to 1.65 km in 30 m steps, and a mean value profile is generated across the range of dilation values as, . The profile of is then searched for local minima from the surface upward. From our case studies we have found that this method is superior to that used in [1] for the lower SNR CALIPSO data, as is more likely to present one unique minima at the correct altitude.