| Temperature and wind,as important environmental parameters,characterize the state of the atmosphere.In the region of upper mesosphere and lower thermosphere(UMLT,75-115 km),due to a lack of effective tools,there is a relative lack of observation data of the vertical distribution of temperature and wind.The resonant fluorescence lidar uses metal atoms and ions in the UMLT region as tracers,and it is possible to detect temperature and wind by stimulating resonant fluorescence signals of the tracer.Among many in-situ and remote sensing detection methods,the resonant fluorescence lidar,with its high spatial and temporal resolution,high accuracy and continuous observation,has become a powerful tool to detect the temperature and wind.At present,sodium resonance fluorescence lidar is mainly used in the world,while iron resonance fluorescence Doppler lidar(Fe lidar)has the advantages of whole day detection and is also an effective means to detect the temperature and wind profile.The narrow-band and frequency-stabilized pulsed output laser is required indispensably during wind measurement.If the laser frequency is locked and calibrated using the Doppler-free saturated absorption spectrum of iron atoms,the temperature of the iron vapor cell needs to be heated to above a thousand degrees Celsius to produce a high density of iron atomic vapor,which is extremely difficult to implement.The difficulties in laser technology have limited the development of Fe lidar.Till now,only a few of Fe lidars were operated by the research groups in University of Illinois,University of Colorado and DLR in Germany.In order to achieve high accuracy detection of temperature and radial wind velocity in the UMLT region,a series of key technical studies have been carried out in this paper,including the derivation of random errors,design of system parameters,simulation of detection performance,development of laser sources,study of filtering techniques.In particular,a new way of generating frequency-stabilized,frequency-switched laser was proposed,and the first narrow-band,high-spectral-resolution Fe lidar operating at 372 nm wavelength has been successfully developed in China.The specific research content can be divided into four points as follows.First,the influence of random errors on the temperature and wind measurements are deduced,based on the basic theory of Fe lidar and the error propagation method.The key technical parameters affecting the measurement performance are analyzed and confirmed,which provide a valuable guidance for Fe lidar design.The system errors caused by frequency jitter and drift are simulated by Monte Carlo method,and the frequency stability indexes are given that need to be focused on when designing the laser source.Second,a new solution of generating frequency-stabilized,frequency-switched laser has been identified after analyzing the advantages and disadvantages of the few 372 nm lasers used in the United States and Germany.In detail,seed laser achieves stable frequency output through the 558 nm iodine saturation absorption spectrum technique;Using optical phase-locked loop technology instead of traditional acoustooptic modulators,laser central frequency and two wing frequencies outputs can be achieved depending on the dwell time;the frequency selection technique suppresses the oscillation of the high gain spectral line and the adjacent spectral line at 1116 nm in the Nd:YAG crystal.The use of a Ramp-Fire seed injection technique with feedback bias,this technique guarantees the frequency stability within 3 MHz@rms of the 372 nm output laser;The dual-range amplification combined with the non-linear frequency conversion module achieves a pulse energy more than 10 mJ.Thirdly,a spatial correlation filtering technique was proposed for filtering the high frequency noise in the lidar backscattered signal,since the detail coefficients of the wavelet transformed signal have a strong correlation between adjacent decomposition orders,while the detail coefficients of the noise show weak or no correlation.The technique can enhance the signal and suppress the noise by multiplying the detail coefficients between adjacent decomposition orders directly.This process is to achieve smoothing of the full-band signal,effective extraction of the signal in the low signalto-noise region and preservation of the detail features of the high-frequency signal.After detailed analysis of the necessary processes,such as wavelet basis functions,decomposition order and threshold selection,taking Rayleigh scattering and resonant fluorescence scattering signal,Mie scattering signal with aerosol layers and thin clouds as research subject,and wavelet analysis and power spectral density as evaluation criteria.The spatial correlation filtering technique shows better performance than sliding average and median filtering methods.This is an effective filtering method.Fourth,Fe lidar with narrow-band,frequency-stabilized,high spectral resolution was first developed in China.The lidar can measure the atmospheric temperature,wind and iron atomic number density in the altitude range of 80 km to 100 km.Through experimental analysis and comparative validation studies,the correctness of detection data and inversion results were ensured. |
