The Retrieval Of Atmospheric Boundary Layer Height Based On Lidar Data And An EnSRF OSSE For The Boundary Layer Height

As the lowest layer of troposphere,atmospheric boundary layer is directly affected by surface conditions and has a significant impact on human's activities.Boundary layer height(BLH)acts as a fundamental parameter for describing the lowest layer of troposphere,retrieval of it is thus essential in boundary layer study.However,BLH can't be observed directly,and it is estimated mainly by analyzing vertical profiles of different meteorological variables.Besides,the BLH retrieved by different instruments may vary.Thus,in our study,we first reviewed the existing retrieval methods for boundary layer height determination based on different instruments,concluded the advantages and disadvantages of each method,and put forward problems may raise in boundary layer height retrieval.Micro pulse lidar(MPL)is widely used to retrieve BLH for its capability to operate unattended for long periods.In this paper,four methods(i.e.,the gradient method(GM),the idealized backscatter method(fitting),Haar wavelet covariance transform method(HM),and Mexican Hat wavelet covariance transform method(MHM))are used to estimate the convective BLH(CBLH)based on MPL normalized relative backscatter(NRB)data collected from the Semi-arid Climate Observatory and Laboratory of Lanzhou University(SACOL).The retrieved CBLH is also compared to the results from the L-band radiosoundings at Yuzhong using 1.5-theta-increase method,and a good agreement is found between them.Further sensitivity analysis using real profiles under different orders of magnitude of background counts show that,the orders of magnitude of background counts can't represent the noise of signal enough since NRB profiles in high orders of magnitude of background counts are possible to have less noise in signals.Besides,when different initial input values are set,the idealized backscatter method always obtains consistent CBLH for profiles with only single shape similar to the idealized curve.For two wavelet methods,it is the noise of profile that really makes sense in BLH retrieval,and the different CBLHs are always obtained with the increase in the wavelet amplitude((35)h)when noise is significant.From a study of real profiles,we are able to estimate the empirical values of(35)h for most profiles for our site and found that(35)h value of 300 m is suitable for HM and MHM.A severe problem can't be ignored in CBLH retrieval based on lidar data is that its application has been hindered by sharp extinction of the signal in high humidity conditions,e.g.,clouds.To remedy this,we developed an effective and simple limiter to obtain more accurate estimates of the CBLH.The limiter is based on the algorithm for the convective condensation level(CCL),and is aimed at limiting the vertical extent of the lidar backscatter profile used in lidar methods to search for the CBLH.Studying of four cloudy cases proves that,with the limiter applied,more accurate retrieval of the CBLH is carried out compared with the CBLHs estimated by the parcel method based on microwave radiometer data.With the development of detection technology,Cloud-Aerosol lidar and Infrared Pathfinder Satellite Observation(CALIPSO)Observations provide broader data in space,which can also be used to estimate BLH.In this study,BLHs estimated from MPL and microwave radiometer data at SACOL are used to evaluate the accuracy of BLHs from CALIPSO.Maximum variance method and Mexican hat wavelet method are utilized to CALIPSO Level 1 backscatter product to eatimate BLHs,and Level 2 aerosol layer product is also used to calculate BLHs simultaneously for comparison.The result shows that the retrieval result from CALIPSO Level 2 aerosol layer product is more reasonable.With the considerable number of ground-based lidar systems and lidar networks,there are large amount of unconventional observation data.A simple way to use these lidar data in numerical model is assimilating BLHs estimated from these instruments.Thus,in the last portion of the study,an Observing System Simulation Experiment(OSSE)bases on the Ensemble Square Root Filter(En SRF)technique for assimilating boundary layer height is made.The result shows that the initial meteorological fields can be improved after absorbing BLHs,however,this improvement dose not last long.