The two main problems encountered in the validation of space-born Earth observation measurements of atmospheric state variables are:
- Natural variability along with less than perfect coincidences of measurement locations lead to unquantified discrepancies which can not be explained by the error budgets of the instruments.
- Even if the problem of natural variability is under control and the discrepancies encountered are not explainable by the combined error budget of the instruments, precision validation gives usually no clue which of two instruments under comparison underestimates its error budget.
In this project, a scheme is developped to use high resolved data from a chemistry climate model to estimate what fraction of discrepancies between different satellite data sets can be attributed to natural atmospheric variability of trace gases. A parametrization, allowing an estimation the variability-induced uncertainty as a function of the mean spatial and temporal distance of the measurements, being developped. A scheme which intercompares three instead of two data sets and will infer corrections applicable to the random uncertainty budgets of each instrument using a linear system of equations. These methods are applied to exemplary sets of measurements; the schemes will be made public.
This project contributes to the SPARC Activity TUNER