SENSITIVITY OF ESTIMATED TRENDS OF TROPOSPHERIC AND LOWER-STRATOSPHERIC TEMPERATURE FROM RADIOSONDE DATA. PART 2. DETECTION OF INHOMOGENEITIES IN MONTHLY-RESOLUTION SERIES

A. M. Sterin

This paper is the second of two papers devoted to the analysis of the sensitivity of upper-air temperature trend estimates derived from radiosonde data. The paper contains a description of a method for detecting temporal inhomogeneities in upper-air (U/A) temperature anomaly series (for the troposphere and the lower stratosphere). The method is based on the ARIMA (AutoRegression—Integrated Moving Average) statistical procedure. The method makes it possible to consider potential inhomogeneities detected by both purely statistical processing techniques and by analysis of historical station upper-air metadata. The detected inhomogeneities affect the values of linear trends. The method is applied to temperature anomaly series for the globe and latitude zones and for selected stations.