Error Analysis Of Water Vapor Observed By Raman Lidar

Water vapour is an important element in atmosphere,it concerns with the process of cloud and precipitation,atmospheric chemistry,energy cycle,and et al.Water vapour is the only atmospheric component which has phase transformation.The unique characteristic and special effect on human being of water vapour make it a hot issue compared with other gas.As one of the active remote sensing equipment,Lidar can observe spatio-temporal variation of water vapour continuously and instantly with high accuracy,sensitivity and resolution.So it is widely used in fields of atmospheric sounding,environmental monitoring,climate research,etc.According to the equation of Raman lidar and the theory of error analysis,the water vapour mixing ratio and its error are studied from the experimental data observed by our developed Raman lidar system.By theory of error analyzing,We find out that the error sources on water vapour observed by Raman lidar includes the calibration constant,the transmission correction and the Raman scattering signals.The calibration constant error is related to the parameter of system and calibrated method.It is about 4-6%and constant with height,and it is the major contribution to the total error under 1.5km height.The transmission correction error is originated from the optical characteristics of atmosphere.It increases slowly along with the height and less than 4%in the condition of clean air and less than 5%in the condition of polluted air.The Raman scattering signals error is associated with the parameter of system and the optical characteristics of atmosphere.It less than 20%under 3km height and becomes the main source above 3km height when the air is clean,but it less than 30%under 2km height and becomes the main source above 2km height when the air is polluted.As for Raman scattering signals is the most uncertain part in water vapour mixing ratio observed by Raman Lidar,the random error method,Poisson method and background method are compared to estimate relative error of Raman scattering signal to get the relative error of water vapour mixing ratio.Meanwhile,In order to validate the errors of water vapour derived by Raman lidar data,the calculated relative errors are compared with that calculated relative errors from Raman lidar and radiosonde data.It turns out that each method both has its own merits and drawbacks.Random error method agrees well with the comparing relative error,but when there is weather system,this method regards the signal as noise,the relative error is bigger than the compared relative error;Poisson method agrees well with the comparing relative error only in condition that the signal to noise ratio is big enough;Background signal method is under the influence of standard deviation of background signal,the error of background signal method is less than the comparing relative error in low layer,but it is useful when the signals decline rapidly.The comparison shows that the relative error derived from Raman Lidar agrees well with the calculated relative error from the Raman lidar and radiosonde data,which proves that the retrieval of relative error from the observed data is accurate to estimate the quality of water vapour measured by Raman lidar.In summary,the analysis results above are helpful to improve the credibility of water vapour observed by Raman Lidar and to promote the application of Raman lidar in weather forecast.