Estimation of Tropopause Height Using Linear Regression Equations and Empirical Formula with Large-Scale Meteorological Variables for Northern Hemispheric Aviation Weather Service

Abstract

An analysis and a forecast of tropopause height using a linear regression equation and empirical formulas with large scale meteorological variables obtained by objective analysis and a numerical prediction model for the northern hemispherical aviation weather service is discussed.
Selected variables in the regression equations for the analysis and the forecast of the tropopause height were the geopotential heights, the stabilities of upper troposphere and the 500mb vorticity. The multiple correlation coefficients of analysis regression equations were about 0.87 and those of forecasting regression equations were about 0.85 except winter cases. The multiple correlation coefficients of the winter forecast remained around 0.65. It is found that observed tropopause heights considerably deviated from the analyzed and forecast values in the tropical and cold cutoff cyclone regions. In those regions the heights derived from the regression equations were modified by the empirical formulas. By this modification, resulting forecast error decreased to less than 20mb except in winter.