Study On Aerosol And Water Vapor In Tibet Plateau By Multi-wavelength Lidar

The Tibet Plateau(TP)is the highest,largest,and most complex plateau on the earth.It plays an important role in East Asian atmospheric circulation,Asian monsoon,and ocean-atmosphere interaction.The strong convective activity in the TP in summer is an important channel for the transport of water vapor and pollutants from the troposphere to the stratosphere.Aerosols are carried from lower troposphere to tropopause,resulting in an increase in aerosol concentration near the tropopause.The total amount of aerosols,water vapor and other substances transported to the stratosphere in TP is even stronger than that of the entire tropics.It is important to obtain accurate distribution of water vapor and aerosol for studying the climate radiation effect of the TP.Due to the sparseness of meteorological stations in this area,the distribution of aerosols and water vapor is unknown in most places.It has important scientific significance and application value for the climate research to obtain the distribution of water vapor and aerosol in the TP using the Atmosphere Profiling Synthetic Observation System(APSOS),the newly deployed in YangbaJain(YBJ,4300 m Above Sea Level,ASL)about 80 km northwest of the city of Lhasa.As the highest altitude lidar site in China,the APSOS was deployed in YBJ in October 2017,and got the first valuble data.Before that,it has been tested in Huainan(36 m ASL)for two years.In this paper,we will use the multi-wavelength lidar,the sub-system of APSOS,to derive the water vapor and aerosol optical parameters of YBJ.The effects of atmospheric model that are often neglected during the inversion are evaluated.The neural network method is used to improve the model,and then to improve the inversion results.The inversion method of the aerosol micro-physical parameters derived by the optical parameters obtained from lidar is simulated.They are as follows:(1)Comparison of water vapor and aerosol distribution in Huainan and YBJ.Lidar data in both Huainan and YBJ were used to derive water vapor mixing ratio and aerosol optical parameters.In Huainan,the microwave radiometer synchronously observation was used to obtain the calibration constant,while in YBJ we used radiosonde in Lhasa.The water vapor mixing ratios of the two sites show that water vapor is richer near ground and quickly declines along with height.At about 4 km,the water vapor mixing radio is reduced to 1/10 of the ground.The difference of the two sites is that water vapor mixing ratio in Huainan is bigger than in YBJ,which is 25 g/kg in maximum in Huainan,while it is 9 g/kg in YBJ.The aerosols in the two sites are similar.The aerosol extinction coefficient is in the same order,such as they are range from 0.02 to 0.03 km-1 below 6 km.The backscattering ratio in YBJ is slightly larger than that in Huainan.The average lidar ratios during 4 km to 8 km are 14 Sr for Huainan and 8 Sr for YBJ.(2)Evaluation of the atmospheric models.Atmospheric temperature and density are needed in the inversion process.But,in the most time,we don't have the synchronous device.Usually,the U.S.Standard Atmosphere(USSA-1976)is used.But the error is noticeable in Lhasa.Compared with radiosonde data,its root mean square errors(RMSEs)can up to 40 K for temperature and 5×1024 m-3 for density,which may be cause 94% error for the aerosol extinction coefficient.Another atmosphere model,Naval Research Laboratory Mass Spectrometer and Incoherent Scatter Radar Exosphere(NRLMSISE-00 or MSIS-00),can be used.The RMSEs are 8 K for temperature and 0.5×1024 m-3 for density compared with radiosonde data.So we choose MSIS-00 model to obtain temperature and density profiles in water vapor and aerosol inversion process.(3)Improvement of the MSIS-00 model.In order to minimize the errors in the inversion of aerosols,the neural network method combined with historical radiosonde data was used to improve the temperature and density of the MSIS-00 model.Radiosonde temperature profiles of about three years were used as training database for this neural network.The radiosonde temperatures in year 2016 were then used to evaluate this method.Our results showed that the deviations between the MSIS-00 model and the in situ radiosonde data have been significantly reduced.The RMSE of the corrected MSIS-00 model temperature,compared with the radiosonde data,was reduced from 9.2 K to 3.2 K at 00 UT,and from 12.5 K to 3.6 K at 12 UT.This study also demonstrates the feasibility of neural network for improving the atmospheric model accuracy by using radiosonde data as training samples.The corrected MSIS-00 model can improve the inversion of aerosol nearly about 7%.(4)Simulation of lidar ratio iterative method.The aerosol optical parameters obtained by remote sensing can derive the microphysical parameters.The Tikhonov regularization method is commonly used.This method requires as many optical parameters as possible.The backscattering coefficient at 1064 nm obtained from lidar is not exact,due to the hypothetical lidar ratio(LR).But we can use LR iterative method.The simulation test shows that this method is feasible.That is the error of the derived microphysical parameters will get smaller and smaller along with the LR approaching to truth.In this paper,multi-wavelength lidar located in YBJ has been used to derive the water vapor mixing ratio and aerosol optical parameters.The aerosol microphysical parameters inversion algorithm has been simulated.In addition,the MSIS-00 model of the TP has been modified.Although these efforts have made good progress,more deeply research is needed.In order to obtain more accurate climatic effects of water vapor and aerosol,long-term lidar observation should be carried out,and joint comparison and verification can be carried out in combination with satellite and in-situ detection instruments.