Retrievals And Analysis Of Precipitable Water Vapor And Aerosol Optical Properties Over Semi-arid Area Of Northwestern China

Water vapor and aerosol are the major atmospheric components, both of them modulate the radiative energy balance of Earth-atmosphere system and influence human activeites. Water vapor, as the highest concentrations of greenhouse gases in the atmosphere, is the important factor in climate change research for its positive feedback. With the intensification of human activities, more and more aerosols are emitted into the atmosphere, which play a crucial role in the global and regional climate change by the direct, indirect and semi-direct effect, and cause severe damage to environment and human health. The region of northwest China located in the climate-sensitive band represents the typical arid and semi-arid climate ozone with most serious area of water scarcity, so the analysis of aerosol optical properties and water vapor in this region shows theoretical and practical significance to the research of climate change and human activities. We have developed the retrieval algorithm of pricipitable water vapor (PWV) and aerosol optical depth using the Multifilter Rotating Shadowband Radiometer (MFRSR), which has been deployed at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL), SACOL city station (Lanzhou university campus) and Zhangye mobile facility during the China-US joint dust field campaign from April to June2008. PWV from MFRSR is analyzed and compared with other ground-based remote sensing instruments. To evaluate and validate the accuracy of satellite PWV retrieval, the comparisons between PWV retrievals from satellite and MFRSR are implemented, basing on which a corrected method for satellite products is proposed. Aerosol optical properties as well as its dependence on air mass long-distance transport are analyzed using aerosol optical depth (AOD) retrieved by MFRSR combined with Mie theory and HYSPLIT model.PWV is retrieved from water vapor absorption channel from MFRSR based on the law of radiation extinction, LBLRTM for simulating the relationship between water vapor transmission and water vapor content and Gaussian extrapolation of daily calibration constant. PWV retrieved from MFRSR at SACOL shows a seasonal variation, with the average value of0.72±0.53cm and75%of PWV values smaller than1.07cm during the observation period. The results of comparisons show a remarkable agreement between MFRSR and sunphotometer with correlation coefficient of0.99and root mean square difference (RMSD) of0.09cm, and correlation coefficient of0.97and RMSD of0.14cm between MFRSR and microwave radiometer, which verifies the accuracy and reliability of MFRSR PWV retrieval. To evaluate PWV products from MODIS and AIRS, we use the PWV from MFRSR to compare with MODIS near-IR and AIRS PWV over northwest China. The results suggest that an overestimation of MODIS increases with PWV value, and a systematic underestimation of AIRS PWV retrieval relative to MFRSR. The errors of MODIS PWV retrieval are likely caused by the uncertainty of surface spectral reflectance and the neglect of multiple scattering effects of aerosol. AIRS PWV retrieval is very sensitive to location and the instruments chosen to compare. According to the results of comparison, a method based on MFRSR PWV as the reference value to correct the MODIS PWV products for Terra and Aqua is proposed. Note that the evaluation and application of satellite products over semi-arid area of northwest China need to seriously consider the applicability of sensor characteristics and the retrieval algorithm.Since MFRSR retrieves the AOD and Angstrom exponent at the same time, the size of fine aerosols and the contribution of fine aerosol to the total AOD over SACOL are presented using straight-forward graphical framework for discriminating different aerosol types. The results show that the fine mode radii is about0.05-0.15μm, especially in the0.13μm, and fraction contribution of fine mode to total AOD is mostly less than70%. The larger AOD is due to well-know presence of mineral dust aerosol, which contribute the largest proportion of61%in spring, followed by37%in winter. This research about understanding the characteristic of fine aerosol in the whole atmospheric column can complement the existing researches.Based on AOD retrieved by MFRSR, the HYSPLIT model and black carbon mass concentration by aethalometer, the influence of long-distance transport of air mass on AOD and black carbon mass concentration over SACOL is analyzed. The backward trajectories are classified into15clusters, in which the average AOD in870nm channel could be up to0.29±0.12under the control of air mass from Taklimakan Desert where dust aerosols are transported significantly, the average AOD could be down to0.14±0.02under the control of air mass from Bangladesh where fine aerosols are transported. According to the classification of different air mass sources, the results show that the greatest contribution to the average AOD, accounting for almost41.1%, come from local and regional sources, followed by about28.4%from dust areas, and the remaining17.9%from central Asia source and12.6%from Europe and Middle East. The black carbon mass concentration under the influence of air mass from western European is the largest with the value of2886.39±1580.34ng/m3, followed by air mass from Bangladesh with the value of2704.40±1407.46ng/m3, with the strong sources of back carbon emissions, all of which is the important air mass sources for affecting the back carbon concentration at SACOL, while the black carbon mass concentration under the influence of air mass from Mongolian Gobi is the smallest.Finally, to further analyze aerosol properties over semi-arid area, comparative experiments from different regions are implemented. The analysis over urban and rural region shows that AOD in Lanzhou city with diurnal variation by single-valley type and with the minimum at14:00is higher systematically than at SACOL in winter. AOD at Zhangye with unobvious daily variation is similar to SACOL. The PM10mass concentration trend of diurnal variation at Zhangye shows a good agreement with SACOL, with two peaks at08:00and22:00respectively, and90%of mass concentration is less than320μg/m3.