Stochastic Modeling of Stratospheric Temperature

This study suggests a stochastic model for time series of daily-zonal (circumpolar) mean stratospheric temperature at given pressure level. It can be seen as an extension previous studies which have developed models surface temperatures. The proposed is sum deterministic seasonality function and L\'evy-driven multidimensional Ornstein-Uhlenbeck process, mean-reverting process. More specifically, the deseasonalized order 4 continuous autoregressive model, meaning that modeled to directly dependent on over four preceding days, while model's longer-range memory stems from its recursive nature. based data European Centre Medium-Range Weather Forecasts ERA-Interim reanalysis product. residuals are well-represented by normal inverse Gaussian distributed random variables scaled with time-dependent volatility function. A monthly variability in speed reversion found, hence suggesting generalization 4th model. this paper, used geophysical analyses improve understanding dynamics. In particular, such characterizations may step towards greater insight modeling prediction large-scale middle atmospheric events, example sudden warmings. Through stratosphere-troposphere coupling, stratosphere source extended tropospheric predictability weekly timescales, great importance several societal industry sectors.