INVESTIGATION OF TRACER GAS VARIABILITY FROM OBSERVATIONS AND MODELS

The variability of the tracer gases CO2 and ozone was investigated utilizing in-situ measurements, satellite retrievals, and model simulations. In the tropical region, using Atmospheric Infrared Sounder (AIRS) mid-tropospheric CO2 retrievals led to the discovery of a signal with a periodicity of around two years, which was found to be related to the Tropospheric Biennial Oscillation (TBO). During strong (weak) monsoon years, the Western Walker Circulation was strong (weak), resulting in more (less) CO2 in the mid-troposphere. The MOZART-2 model results were consistent with observations with a smaller amplitude. In the North Hemisphere polar region, the influence of sudden stratospheric warming (SSW) on AIRS CO2 was investigated. The Eliassen-Palm flux divergence was negative before the SSW; as a result, the westerly wind in the stratosphere decreased. During the SSW, the polar zonal mean wind switched to easterly and the temperature increased. Mid-latitude CO2 was transported to the high latitudes, leading to an increase of mid-tropospheric CO2 concentrations in the polar region. Over the Central Pacific Ocean, the influence of El Niño on the mid-tropospheric CO2 was investigated using the MOZART-2 model. Model simulation results were consistent with the observations. There was more (less) mid-tropospheric model CO2 in the central Pacific and less (more) mid-tropospheric model CO2 in the western Pacific during El Niño (La Niña) events. In addition to exploring these CO2 variations, the impacts of ENSO on the tropical total column ozone, the tropical tropopause pressure, and the 3.5-yr ozone signal in the mid-latitude were also investigated. Both the observations and Goddard Earth Observing System Chemistry-Climate Model (GEOS CCM) show tropical tropopause pressure to be related to the ENSO signal in the total column ozone. The GEOS CCM was also used to investigate a possible mechanism for the 3.5-yr signal observed in the mid-latitude total column ozone. Results suggested that a model with realistic ENSO could reproduce the 3.5-yr signal. Hence, it is likely that the 3.5-yr signal was caused by ENSO.