Statistical parameters derived from ozonesonde data of importance for passive remote sensing observations of ozone

A reliable climatology of vertical ozone and temperature variability is essential for our understanding of climate, modelling work and profile retrieval from passive remote sensing observations. This paper describes a statistical analysis of ozonesonde data collected from 1970 to 1996 at 17 stations in the northern hemisphere. The aim of the analysis is to provide a number of parameters describing vertical ozone and temperature variance, justified primarily by the importance of such information for remote sensing applications. Deviations between local means, probability density functions, standard deviations, and correlation between different pressure levels are reported for altitudes below 10 hPa and three latitude bands (around 35°N, 50°N and 75°N). The analysis includes also cross-correlation between ozone and temperature. Differences are found between local mean values up to 40% and 15 K for ozone and temperature, respectively. Standard deviations are 10–80% for ozone and 2–15 K for temperature. For ozone the exponential correlation length is normally 2–5 km, while for temperature the vertical correlation is more inhomogenous. The implications for inversion of remote sensing measurements are discussed.