Calibration and Validation of COSMIC-2 Radio Occultation Data: Error Statistics Estimated through Comparison with Other Datasets

The COSMIC-2 mission aims to provide a wealth of high-quality profiles of the tropical atmosphere. With its constellation of six low-earth orbiting satellites, this mission is expected to produce on the order of 5000 daily, sub-kilometer vertically-resolved profiles of refractivity and bending angle, and with ancillary data, temperature and water vapor. The COSMIC-2 program office is currently undertaking an effort to calibrate and validate both the quality and quantity of these new data.

Here we present initial results of part of that effort: estimation of error statistics through comparisons between COSMIC-2 data and co-located datasets, including other radio occultation missions, radiosondes, model analyses and forecasts, and reanalyses. We use a suite of methods, including traditional metrics (e.g., biases, RMS, and standard deviations of differences) and the three-cornered hat (3CH) method (see Anthes et al. at this conference for details), to produce these estimates. The 3CH method provides us with robust error estimates of not only the current COSMIC-2 data, but also of the error statistics for the co-located datasets.