The CO₂ seasonal cycle as a tracer of transport

CO 2 time series in the lower stratosphere have been constructed as a function of latitude and altitude using ER‐2 aircraft observations. A seasonally varying signal derived from tropical surface CO 2 data was fit to these data in a manner similar to that of Boering et al. [1996]. This analysis shows how the tropospheric CO 2 seasonal cycle in air entering the tropical lower stratosphere ascends and gradually loses amplitude. The midlatitude results show that in the 370‐ to 400‐K range the tropical signal is transported poleward rapidly with only a small loss in amplitude. The signal also reaches the midlatitudes rapidly at higher altitudes but with markedly less amplitude. Above ∼440 K the midlatitude CO 2 data best fit a line whose slope is the secular increase in CO 2 . This observational analysis is used as a basis for diagnosing transport in a CO 2 simulation produced using GEOS‐1 Data Assimilation System (DAS) winds in an off‐line chemistry and transport model. The model is forced at the lowest levels by a 5‐year zonal mean CO 2 data set derived from surface observations. The analysis of model results as a function of height and latitude determines whether (1) the seasonally varying CO 2 cycle forced in the troposphere survives in the tropical lower stratosphere and whether (2) the CO 2 cycle in the tropical lower stratosphere propagates realistically to the midlatitudes. We find that the model transports a CO 2 seasonal cycle from the surface to the tropical upper troposphere that agrees with the phase and amplitude of tropospheric aircraft measurements. The cycle input into the model tropical lower stratosphere attenuates with height in a reasonable way compared with observations. The phase and amplitude of the cycle transported from the tropics to the midlatitudes agree well with the data up to about 440 K, but above that the model transports more of the tropical signal to the midlatitudes than is suggested by the ER‐2 data. Overall, there is remarkable consistency between the model and observations in the upper troposphere/lower stratosphere, which supports the use of assimilated winds in assessments of the effects of aircraft exhaust.