Previous: Introduction Next: Results Up: Ext. Abst.

Method

To study the relationship between interannual climate variability and the interannual variability of atmospheric methane and ozone we use the University of Illinois at Urbana-Champaign (UIUC) Stratosphere/Troposphere Atmospheric Chemical Transport Model (ACTM) and the UIUC Coupled Chemistry/Climate Model (UIUC CC/CM). The photochemical routine of the ACTM and UIUC CC/CM includes all of the principal gas-phase and heterogeneous reactions involved in the production and loss of chemical species (Rozanov et al., 1999, 2000). The climate part of the UIUC CC/CM also reproduces the main features of the atmospheric circulation (Yang, 2000).

Both models have a 4 degree latitude by 5 degree longitude resolution and have 24 layers from the Earth's surface to 1 hPa.

We have performed two experiments with each of these models, a control run and an experiment run. In the first pair, each run was a 15-year equilibrium simulation with the UIUC CC/CM. The control run was made with the sea surface temperature, sea ice and solar insolation specified for the current climate. In the experiment run the spectral distribution of solar radiation at the top of the model atmosphere was changed to equal that observed at solar maximum (Lean et al, 1995). In the second pair, each a 6-year transient simulation with the ACTM, the control run was made with the current observed geographical distributions of the surface sources of CO₂, CH₄, N₂O, CFC-11 and CFC-12, taken from the NOAA/CMDL database (http://www.cmdl.noaa.gov/hats/index.html), and the surface fluxes of NO_x and CO, taken from MÃ?ller and Brasseur (1995). The experiment run was made with the surface source of CFCs increasing 3% per year (WMO, 1999). In these ACTM simulations, the three-dimensional winds and temperature fields were prescribed from the UKMO reanalysis dataset for 1993-1999 (Swinbank, and Oâ?™Neill, 1994).

To define the principal modes of climate variability in the stratosphere and troposphere over the high-latitude domain, 50°N-90°N, together with the principal modes of the variability of ozone and methane, we calculated Temporal Empirical Orthogonal Functions (TEOFs). TEOF analysis allows presentation of the original time series in terms of the sum of the product of a set of time-independent spatial modes, , and their corresponding spatially-independent temporal eigenvectors, ,

where k is an atmospheric level; j is a mode of variability;and are longitude and latitude; t is time; and is the time series of quantity being investigated. Each mode j with eigenvector has its unique eigenvalue , which after being normalized, defines the contribution of the corresponding mode to the total variability of the quantity. For each model variable we analyze the first mode of its variability, which is expressed by the first term in the summation in the above equation, . For each atmospheric pressure level, the pattern and time evolution of the first mode may be different. In our analysis they are sorted in increasing order of magnitude of their eigenvalues, thus the first mode defined by the largest value of the eigenvalue. We calculated the TEOFs for each of 24 model atmospheric pressure levels and each month of the n-year model runs.

We have mapped the magnitude of the normalized eigenvalues expressed in percent, for each quantity as a function of atmospheric pressure and month. Mapping the normalized eigenvalues of the first mode gives visual information about the atmospheric location where and month when the first mode dominates the variability. The closer the eigenvalues are to 100%, the closer the first mode is to the variabilty climatology over the corresponding time period. We have compared the climate variables' eigenvalue maps (EVMs) calculated from the model simulations and the UKMO and NCEP reanalysis datasets (http://www.sparc.sunysb.edu/html/ref_clim.html). Then we produced maps of the difference between the control and experimental runs to locate the changes in the dominance of the first mode of interannual variability for geopotential height (GPH), temperature and chemical species. Finally, we examined how the changes in solar radiation and the surface sources of CFC affected the interannual variability.