Analysis of Trace Gas Heterogeneity using In-Situ and Remote Sensing Measurement Techniques

Abstract

Differential Optical Absorption Spectroscopy (DOAS) is a remote sensing spectroscopic technique capable of retrieving vertical column densities of trace gases in the atmosphere. While the technique can be used to retrieve vertical profiles of trace gases this retrieval is not applicable to Mobile-DOAS measurements due to the lengthy measurement times. In addition, the requirement for rapid measurements for Mobile-DOAS application prevents a single instrument from retrieving information about the horizontal trace gas distributions around the instrument, in addition to the lack of vertical distribution information. However, with the addition of collocated in-situ measurements of surface concentration, as well as the use of Multi Axis DOAS (MAX-DOAS) measurements taken from multiple azimuth viewing directions, it is possible to estimate these distributions using a minimal amount of measurements, allowing this gap in mobile DOAS measurements to be filled. The relationship between the vertical column density measured by Mobile Zenith-DOAS and in-situ surface concentration is first explored using measurements made during the first TRACER-AQ field campaign. This relationship, expressed as the column to surface ratio, is then used to identify trace gas plumes and other transport patterns that can be separated from local, vehicular emission based on this ratio during two case days, in order to determine the origin of the high Ozone event that was observed on both days. In addition, the effect of the trace gas vertical distribution on satellite-based DOAS measurements is also determined, where it is found that the column to surface ratio is proportional to worsening agreement between Mobile Zenith-DOAS and TROPOMI Satellite-Based DOAS, however the relationship is only somewhat correlated (r2=0.3), indicating that other factors are a greater influence on the ability of Ground-Based DOAS and Satellite-Based DOAS to agree. During the second TRACER-AQ field campaign, the column to surface ratio was combined with analysis of MAX-DOAS measurements in order to determine both the vertical and horizontal distributions of trace gas around the instruments. This analysis of the horizontal distribution is expressed with the Horizontal Heterogeneity Index, which is a comparison between two MAX-DOAS azimuth viewing directions, behind and to the right of the moving vehicle, with the Zenith DOAS measurements each MAX-DOAS instrument is making. The analysis of both distributions is then used to pinpoint the potential sources giving rise to two more high Ozone events during two more case days during the second field campaign. In addition, the horizontal trace gas distribution was also compared to TROPOMI validation, and was found to correlate more strongly than the vertical distribution (r2=0.69), showing that while there are many issues with satellite validation, the horizontal distribution is most likely responsible for deviations between Ground-Based and Satellite-Based DOAS measurements. Finally, analysis of trace gas distributions requires low uncertainty retrievals of Zenith DOAS VCD's. This is typically performed using radiative transfer simulations, but which require information of the aerosol scattering properties that are not retrievable from a mobile platform. In order to perform these simulations, a novel technique of averaging the aerosol conditions as measured by a network of six AERONET aerosol measuring instruments was investigated through a series of sensitivity studies. These studies analyzed the effect of potential errors in the aerosol properties on the resulting Air Mass Factor estimations, in order to determine how these errors effect Air Mass Factor uncertainty. The studies showed that the total error introduced into the model through potential errors in aerosol properties, as well as through errors in the assumption of an a priori trace gas profile used to retrieve the Air Mass Factors, was less than 20% for AMF retrievals in Visible light and 25% for UV retrievals, approximately the same level of uncertainty used for Zenith AMF retrievals in other studies, demonstrating that this averaging technique is capable of retrieving sufficiently low uncertainty AMF estimations.