This technical note summarizes and classifies the various approaches to validation of remote measurements of atmospheric state variables, and tries to recommend a clear and unambiguous terminology. The following approaches have been identified: Intercomparison of single profiles for accuracy validation; statistical comparison of matched 5 pairs of measurements with respect to bias determination and precision validation; statistical intercomparison of randomly sampled measurements by two

more »

... nts, and comparison of a single measurement to an ensemble of measurements. Applicable statistics are shortly reviewed, and recipes for evaluation of the co-incidence error due to less than perfect co-incidences are presented. A rigorous approach is suggested 10 to quantitatively validate profile measurements when full covariance matrices are unavailable. We distinguish between "necessary validation" which is rejection of the null hypothesis that a difference between two measurements is significant, and "sufficient validation" which means to provide evidence that the probability that there is a significant difference is definitely small. 15 25 4974 ACPD 6, 2006