The ratio of temporally adjacent lidar pulse returns is commonly used
in differential absorption Lidar (DIAL) to reduce correlated noise. Th
ese pulses typically are generated at different wavelengths with the a
ssumption that the dominant noise is common to hath. This is not the c
ase when the mean number of laser speckle integrated per pulse by the
lidar receiver is small (namely, less than 10 speckles at each wavelen
gth). In this case a large increase in the standard deviation of the r
atio data results. We demonstrate this effect both theoretically and e
xperimentally. The theoretical value for the expected standard deviati
on of the pulse-pair ratio data compares well with the measured values
that used a dual CO2 laser-based lidar with a hard target. Pulse aver
aging statistics of the pulse-pair data obey the expected sigma(1)/roo
t N reduction in the standard deviation, sigma(N), for N-pulse average
s. We consider the ratio before average, average before ratio, and log
of the ratio before average methods for noise reduction in the lidar
equation. The implications of our results are discussed in the context
of dual-laser versus single-laser lidar configurations. (C) 1997 Opti
cal Society of America.