Observations of trace gases from ground-based and satellite instruments: Retrieval development and analysis of vertical and horizontal distributions

Abstract

Air pollution affects human health, ecosystems, and the climate. For a better understanding of air pollution, continuous monitoring, measurement validation, and measurement improvement is needed. Therefore, ground-based and satellite-based measurements as well as model data are analysed in this thesis with the main goal to improve the knowledge about the trace gas vertical distribution. The data are compared for three different environments: rural, biomass burning, and industrialised areas. Two different nitrogen dioxide (NO2) vertical column density retrievals for satellite measurements are compared with ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) data above Xianghe, China. The influence of satellite retrievals' a priori assumption is investigated on their agreement with MAX-DOAS data. The a priori profiles are an important difference between the retrievals' input parameters. For this purpose, the MACC2 interim reanalysis data and the MOZART-2 model are used. Generally, the retrieval version with MACC2 interim reanalysis a priori profiles agrees better with the ground-based MAX-DOAS measurements. Furthermore, the NO2 columns are often higher for cloudy scenes than for cloud-free scenes in this area in winter, which might be related to aerosols. Thus, the influence of the cloud fraction and the aerosol content of the atmosphere on the differences between the two retrievals are discussed which provides also the opportunity to understand and reduce the differences between measurements from satellite and ground-based platforms. Not only above industrialised areas but also above rural areas validation measurements are needed. Due to the low pollutant content in these areas, the relative uncertainties of satellite measurements are high. There, satellite data regularly show enhanced levels of short-lived pollutants, which are unexpected to be present in the atmosphere at high levels so far away from any sources. Therefore, MAX-DOAS measurements were taken in the remote Atlantic Ocean to validate satellite measurements. For formaldehyde (HCHO) and glyoxal (CHOCHO), the latitudinal dependency of the MAX-DOAS measurements is similar to satellite observations and model simulations. The trace gas origin could be the African continent. Because of long transports of up to four days, the enhanced values might be related to the export of precursors from the continent which are transformed to HCHO and CHOCHO. Furthermore, the detected outflow is in elevated layers which is in agreement with FLEXPART backward simulations leading to an altitude between 1000m to 5000m. Thus, these results improve the confident in satellite measurements above remote ocean areas. The knowledge of the trace gas vertical distribution on a global scale is generally limited due to the low vertical information content of spectroscopic measurements for the absorbers. However, retrievals in different spectral ranges from satellite measurements can globally provide insight into the vertical distribution of NO2. Therefore, a new NO2 retrieval in the ultra violet spectral range was developed and compared with the standard retrieval in the blue spectral range. Due to the differences in sensitivity to the lower atmosphere, a rough estimation of the NO2 vertical distribution is possible from these two satellite retrievals. In anthropogenically polluted areas, the differences are large between the two slant column densities (approximately 60%), which indicates that the air pollution is located close to the ground. In contrast, the NO2 is uplifted into elevated layers in biomass burning areas which is reflected by reduced differences of approximately 36% between the two spectral ranges. Additionally, models can be validated by converting slant column densities into vertical column densities for both spectral ranges. In theory, if the a priori profiles are correct, the vertical column densities should be similar. Two models, TM5-MP and the MACC2 interim reanalysis data, are used for the conversion into vertical column densities. In total the TM5-MP leads to a better agreement between the ultra violet and blue spectral range. However, in this model the NO2 seems to be located too high in the atmosphere.