Abstract

The deconvolution of lidar profiles is an effective software way to improve the lidar resolution and to extract small-scale aerosol inhomogeneities, especially in coherent CO₂ (Doppler and DIAL) lidars, where long laser pulses are emitted as a rule by TE (TEA) CO₂ laser transmitters (see e.g.[1]). Different deconvolution techniques, based on the Fourier transformation or Volterra integral equation or a recurrence relation have been developed in [2,3]. As shown in [4], the uncertainty in the measurement of the laser pulse shape (e.g. a spike cut) causes a shift up or down as a whole of the restored lidar profile (with respect to the true one) and distortions in the spectrum of the retrieved inhomogeneities. As a rule the spike duration does not exceed several sampling intervals Δt. On the other hand, the increase of the sampling rate to reduce the uncertainty causes hardware problems such as higher data rate, lower dynamic range etc.

© 1995 Optical Society of America