The atmospheric boundary layer (ABL) is the lowest layer in the troposphere, which is relatively more uniformly mixed under the influence of the ground-air interaction. This layer is directly affected by surface forcing during the interaction between the surface and the atmosphere, and its response time is less than one hour, usually in the range of 100-3000 m. Meteorological elements in ABLs have obvious diurnal changes, and there are multiple scales of motion, with turbulence as the main form. ABL greatly affects peoples production, life, and other activities. ABL height (ABLH) has high research significance in meteorology and other related fields. A high concentration of aerosol and water vapor in ABL contrasts significantly with a free atmosphere that is above the ABL. From the perspective of substance distribution, the atmospheric material boundary layer (AMBL) and atmospheric material boundary layer height (AMBLH) can be determined. AMBL is a part of ABL, and the characteristics of AMBLH are similar to that of ABLH but they have differences. AMBLH mainly reflects the spread of atmospheric substances and pollution in ABLs, but it is not capable of representing all turbulent motion in ABLs. ABL is closely related to air pollution, in which turbulence has a diffusion and mixing effect on air pollutants. ABLH is negatively correlated with the concentration of pollutants near the ground, and heavy pollution is often accompanied by a lower ABLH. Traditional observation techniques include radiosondes, meteorological towers, airplanes, etc., which can realize direct observation of the profile of meteorological elements and ensure the accuracy and credibility of the inversion. However, limited by cost and observation methods, such means fail to achieve long-term and continuous observation. With the development of remote sensing technology and its application in the field of meteorology, ground-based remote sensing equipment such as lidar, microwave radiometer, ceilometer, and sodar has shown their advantages in ABL observation. As an efficient ground-based remote sensing instrument, coherent Doppler lidar (CDL) can realize continuous and high-precision observation of ABLs and has been used widely in recent years.

A large number of retrieval methods based on lidar data have been applied, and each of them focuses on different characteristics of ABLH and has various advantages. Thus, different results may be obtained. In order to increase the accuracy of AMBLH retrieval, we proposed a novel comprehensive AMBLH retrieval method based on several mature methods, including the gradient method, wavelet covariance transform method, and signal-to-noise weight-fraction method. The comprehensive AMBLH retrieval method contains three modules for increasing retrieval accuracy. Firstly, by summarizing various characteristics of different weather, a complex weather distinguishing method was developed. Next, based on the wavelet covariance method, a cloud-deleting module was applied. Thus, complex weather conditions and clouds would be eliminated. Finally, by inputting sunrise and sunset time information, different methods can be applied for the convective boundary layer and aerosol contamination layer between the stable boundary layer and the residual layer, respectively. With the means mentioned above, the comprehensive AMBLH retrieval method can obtain AMBLH with high accuracy and efficiency.

By using the CDL observation experiment data in Qingdao from November 2020 to November 2021, the AMBLH in Qingdao within one year is retrieved and analyzed. All the months of the year except June and July show a clear diurnal trend of AMBLH, indicating that solar radiation heating has a great impact on the development of AMBL. The diurnal variation trend of AMBLH is not obvious for two months in summer, and it is considered that the temperate monsoon climate in Qingdao is characterized by the prevailing southeast wind from the ocean in summer. The development of AMBLH is affected by the ocean surface temperature front cold center, which increases the stability of the lower AMBL. At the same time, atmospheric mixing and sea surface wind speed are reduced, which limits the development of AMBL in both thermal and dynamic factors. Correspondingly, the median AMBLH in June and July is the lowest throughout the year.

The statistical analysis of AMBLH in each season shows that the overall development degree of the seasonal diurnal variation of AMBLH is similar in spring and winter but more obvious in autumn, and the degree in autumn is higher than that in summer. A higher degree indicates a more obvious diurnal rise and fall of AMBLH. The daily mean AMBLH fluctuates less in each season, and the median of AMBLH in each season in descending order is spring, winter, autumn, and summer. In addition, the median of AMBLH in spring and winter is similar. It is found that the diurnal development and variation characteristics of AMBLH in spring, autumn, and winter are similar. Since the development of AMBL is significantly affected by solar radiation heating, it is believed that this phenomenon is caused by the small diurnal variation of temperature in Qingdao throughout the year. The timing of the uplifting trend of AMBL development in the diurnal variation of each season is different, which may be caused by the combined effect of sunrise time and seasonal temperature differences. The order from early to late is summer, spring, autumn, and winter.

Based on a variety of well-developed AMBLH retrieval methods, we propose a comprehensive AMBLH retrieval method, which performs targeted retrieval for the difference between the mixed layer and aerosol contamination layer during the day and night and effectively removes the influence of clouds and complex weather such as rain and snow. Therefore, the accuracy of AMBLH retrieval is improved. Compared with the synchronous observation results of radiosondes, this method shows high consistency and good agreement (determination coefficient R² =0.93). AMBLH in Qingdao in the time period of 13 months is retrieved by the comprehensive method, and the results are statistically analyzed.