MAX-DOAS Measurements Of Tropospheric NO₂ And SO₂ Vertical Column Densities Over A Typical Polluted Area In North China

Tropospheric NO2 and SO2 are both key pollutants, which are important precursors of photochemical pollution and haze-fog and are extremely harmful to ecology, environment and human health. North China(NC for short) is one of the most severely polluted regions worldwide, being recognized as “a pollution and oxidation pool”. Large mounts of NO2 and SO2 in the atmosphere can play important roles in pollution formation and photochemical oxidization processes. Therefore, measurement of tropospheric NO2 and SO2 in typical polluted areas in NC are of the most importance to understand the variation characteristics, emission origins and fomation mechanisms of pollution in this region.In this thesis, the tropospheric vertical column densities(VCD) of NO2 and SO2 are retrieved, based on DOAS method and the so-called geometric approximation, from the scattered sun light spectra measured by a Multi-Axis Differential Optical Absorption Spectroscopy(MAX-DOAS) at the Gucheng Eco-meteorological Observation Experiment Station(GCH for short, 115o48′E,39o08′N, 15.2 m asl, located 110 km southwest of Beijing Megacity) in Hebei province from September 2008 to September 2010. Then, the seasonal and diurnal variations of the tropospheric NO2 and SO2 are systematically analyzed in combination of the in situ surface mixing ratios and NASA OMI satellite tropospheric column density products of these two species. The vertical distributions of the tropospheric NO2 and SO2 are studied and effects of the geographical distribution of respective emission sources and regional transport from different directions are investigated. Good agreement in the seasonal variation patterns are found for both tropospheric VCDs and surface mixing ratios of NO2 and SO2, which are high in winter and low in summer. The diurnal variation patterns of the tropospheric NO2 VCD are shown to be similar to surface measurements for most months with a vally at noon. A noon-peak diurnal variation pattern of tropospheric SO2 VCD is identified for winter. The diurnal vatiation exhibits a vally pattern in summer, which is opposite to the noon-peak pattern of diurnal vatiations of surface mixing ratio of SO2 of the same period.The pollution layer height(defined as the tropospheric NO2(or SO2) VCD divided by its surface concentration) along with the ratio of DSCD3°,trop and DSCD30°,trop from our MAX-DOAS measurements are used to quantatively characterize the vertical distribution of NO2 and SO2 over GCH. In comparison with the sepheric Mc Artim radiative transfer model simulations, it is found that the layer heights of SO2 are much higher than those of NO2 in general, and elevated layers occur more frequently for SO2 than for NO2.Finally, the tropospheric NO2 and SO2 VCDs retrieved from our MAX-DOAS measurements are compared with 5 types of OMI satellite tropospheric VCD products to evaluate the reliability of these products for typical polluted rural areas in NC. In addition, a radiative transfer model(Mc Artim) is used to study the effects of various combinations of aerosol and trace gas profiles on the AMFs and thus on the tropospheric VCDs retrieved from both MAX-DOAS and satellite observations. Very good agreement is found for monthly mean tropospheric NO2 VCDs in summer between MAX-DOAS and OMI level 2 products(OMNO2 from NASA and DOMINO from TEMIS). However, the OMI products systematically underestimate the tropospheric NO2 VCDs with respect to MAX-DOAS measurements at GCH during wintertime. From the AMF simulations for MAX-DOAS, we conclude that the errors of the geometric approximation are typically smaller than 20%. Thus the most probable reason for the observed differences is an underestimation of the true NO2 VCDs by the satellite. The analysis of satellite data with filtering criterion of center and outermost pixels of OMI indicates weak spatial gradients of NO2 for typical polluted rural regions like GCH, therefore the gradient smoothing effect cannot affect the satellite retrieval so much as for regions with strong gradients(like e.g. Beijing or Shanghai). Consequently this underestimation by OMI can be mainly attributed to the aerosol shielding effect.Comparisons between monthly mean tropospheric SO2 VCDs retrieved from MAX-DOAS and OMI level 3 gridded product(OMSO2e) exhibit fair consistency, with the same seasonal variation of high values in winter and low in summer. The ratios of monthly averaged tropospheric VCDs of MAX-DOAS and OMI range from 0.82 to 2.16, approximately around 1. The correlation coefficient of individual SO2 VCDs of MAX-DOAS and OMI is relatively lower than that between MAX-DOAS and level 3 tropospheric NO2 VCD product of OMI(OMNO2d). Unlike NO2, no clear conclusion can be obtained for whether the OMI satellite products(here referred to as OMSO2 e and level 2 grided product OMSO2G_PBL) underestimate or overestimate the tropospheric SO2 VCD compared to grounded MAX-DOAS observations. It is demonstrated by radiative transfer model simulation that the satellite retrieval of SO2 has larger uncertainties under typical aerosol polluted conditions(like haze), but the corresponding uncertainties are much smaller for MAX-DOAS. This indicates that improvements of the retrieval algorithm for MAX-DOAS and off-line corrections of satellite measurements for the tropospheric SO2 VCDs should be given more emphasis in the future.