SIMULATION OF COHERENT DOPPLER LIDAR PERFORMANCE IN THE WEAK-SIGNAL REGIME

The performance of coherent Doppler lidar in the weak-signal regime is investigated by computer simulations of velocity estimators that accu mulate the signal from N pulses of zero-mean complex Gaussian stationa ry lidar data described by a Gaussian covariance function. The probabi lity density function of the resulting estimates is modeled as a fract ion b of uniformly distributed bad estimates or random outliers and a localized distribution of good estimates with standard deviation g. Re sults are presented for various velocity estimators and for typical bo undary layer measurements of 2-mu m coherent lidars and also for propo sed space-based measurements with 2- and 10-mu m lidars. For weak sign als and insufficient pulse accumulation, the fraction of bad estimates is high and g approximate to w(v), the spectral width of the signal i n velocity space. For a large number of accumulated pulses N. there ar e few bad estimates and g proportional to w(v)N(-1/2). The threshold s ignal energy or average number of coherent photoelectrons per pulse wi th accumulation is defined by a given fraction of random outliers and is proportional to N--1/2 for large N and decreases faster than N--1/2 for small N. Ar the threshold level, the standard deviation g of the good estimates is approximately constant fur large N. For space-based measurements and with the signal statistics determined by the wind flu ctuations over the: range gate, the 2- and 10-mu m lidars have similar performance when referenced to the average number of photoelectrons d etected per velocity estimate. The threshold signal level for large N can be described by simple empirical functions.