Strategy for early detection of climatic change

A general strategy for detection of climatic changes of stratospheric temperatures is discussed in the light of expected short-term and long-term natural and spurious (instrumentally-induced) variability. The effects of the brevity of both the radiosonde record and the satellite sounding record are stressed, and the need is emphasised for homogeneous time-series from satellites as well as from sondes.The general principles are illustrated by means of an attempt to detect climatic changes in the extratropical upper troposphere and lower stratosphere using radiosonde data for approximately the past 25 years. It is shown that, even in the absence of instrumental shortcomings, a further 35 years of radiosonde data, or 20–25 years of satellite sounding data, are needed for detection of temperature trends of the size indicated by global circulation models for the present rate of increase of concentration of atmospheric carbon dioxide.