The seasonal cycle dependence of temperature fluctuations in the atmosphere

Abstract

The correlation statistics of meteorological fields have been of interest in weather forecasting for many years and are also of interest in climate studies. A better understanding of the seasonal variation of correlation statistics can be used to determine how the seasonal cycle of temperature fluctuations should be simulated in noise-forced energy balance models. It is shown that the length scale does have a seasonal dependence and will have to be handled through the seasonal modulation of other coefficients in noise-forced energy balance models. The temperature field variance and spatial correlation fluctuations exhibit seasonality with fluctuation amplitudes larger in the winter hemisphere and over land masses. Another factor contributing to seasonal differences is the larger solar heating gradient in the winter. 40 years of monthly mean surface data and 25 years of monthly mean 700mb and 500mb data is averaged over the seasons. The spatial correlation of four northern hemisphere midlatitude test sites, two ocean sites and two land sites, at the surface, at 700mb and at 500mb is examined for the winter, spring, summer and fall. The correlations between the different vertical levels and the variance of each level is also presented and examined.