Interannual variability in tropospheric ozone and its precursors: a modeling study constrained by long-term ground-based and satellite observations

Tropospheric ozone (O3) is a harmful pollutant and an increasingly important green-house gas. Tropospheric O3 distributions are determined by complex and non-linear interactions between photochemistry and transport processes. The first objective of this thesis is to examine the factors that contribute to the interannual variation (IAV) of tropospheric O3 and related compounds, such as carbon monoxide (CO), for a period of 19 years (1987-2005), with a focus over the northern mid-latitudes. A series of long-term measurements, including ozonesondes, groud-based sites and satellites, are interpreted with a global chemical-transport model (GEOS-Chem) and a number of sensitivity simulations are performed to quantitatively assess the relative contribution of individual processes to the IAV of O3 and CO. We first assess the capability of the model to reproduce O3 and CO IAV. We find that the model is able to reproduce reasonably well many of the observed features of the O3 and CO IAV. The model however, fails to reproduce the observed variation in O3 in the northern mid-latitudes during the first half of the 1990s (negative anomaly in 1991-1993 followed by an upward trend in 1993-1996) which is likely due to an inadequate representation of stratospheric dynamics and stratospheric O3 concentrations in the model. This study has demonstrated that the principal drivers of O3 and CO IAV are meteorology and emissions. We examine in particular the 1998-1999 period. A large anomaly in O3 was observed at several sites over Europe and is fairly well reproduced by the model. Different processes are found to contribute to this long-lasting anomaly. The first period (spring 1998) was largely influenced by the strong preceding El Niño episode of 1997. El Niño Southern Oscillation (ENSO) affects the stratosphere-troposphere exchange (STE) and the Asian pollution export, through a change in convective activity and a strengthening of the subtropical jet. North American pollution export was also enhanced through an increase in temperature and in photochemical activity. During the second period (summer-fall 1998), the O3 budget over Europe was influenced by both boreal wildfires and Asian pollution. Finally, the O3 anomaly during the third period (spring 1999) was affected by enhanced biomass burning emissions in Southeast Asia. We further noticed that a significant correlation exists between the Southern Oscillation Index (SO1) in fall and the European tropospheric O3 column in the following spring (R= -0.65, N=19), throughout the period from 1987 to 2005. In other words, increasing tropospheric O3 column is seen over Europe the following spring of an El Niño year. In the second part of this thesis, we used inversion methods to derive monthly NOx, emissions from GOME and SCIAMACHY space-based retrievals of nitrogen dioxide (NO2) columns for a period of 10 years (1996-2005). We calculated emission trends over the highly industrialized regions and large cities around the globe. Decreasing emissions were found in the U.S. and Europe. However, negligible or even positive trends were seen in urban areas in northeastern U.S. (e.g. Boston) and southern Europe (e.g. Lisbon, Istanbul) which are most probably due to increasing motor vehicle NOx, emissions. Furthermore, a slowdown in the decreasing trend in NOx, combustion fuel emissions over North Europe has been noticed after 2000 and especially in 2003-2004. This is probably related to a slowdown in the decrease in road transport emissions. Increasing trends, on the other hand, have been found in the combustion fuel NOx, emissions over east China (8.3% year-1), due to the fast industrial and economical development of China. In addition, we found an upward trend in soil NOx, emissions which may be related to the increasing use of fertilizers in the northeastern part of the country. Finally, we examined the influence of using two different retrievals of NO2 columns on the resulting a posteriori inventories. The inversion was carried out twice for the year 2000, using two different retrievals, one from Dalhouse/SAO and one from BIRA/KNMI. The two resulting inventories were found to correspond well globally, but show important differences (more than 60%) at specific times and regions.


Related material