The importance of simulated errors in observing system simulation experiments

Abstract

Observing System Simulation Experiments (OSSEs) for numerical weather prediction rely on simulated observations that should include simulated observation errors in order to realistically represent the behaviour of real data. Real observations include many types of error, such as instrument error, representativeness error, and observation operator error, with some portion of this error being correlated in time and space or possibly between data types. Data assimilation systems are designed to account for random, uncorrelated errors, but are not yet adept at handling correlated errors; as a result, the correlated errors are more readily incorporated into the analysis increment by the data assimilation system than uncorrelated errors. In this work, the role of correlated observation errors in modifying the behaviour of the National Aeronautics and Space Administration Global Modeling and Assimilation Office (NASA/GMAO) OSSE framework is investigated. The effects on analysis increment, analysis error, forecast errors and observation impacts of including or neglecting correlated simulated errors is explored. The use of correlated observations for calibration and validation of the OSSE is also discussed.

Keywords:

• data assimilation
• observing system simulation experiment
• observation error

Acknowledgements

Support for this project was encouraged by Steven Pawson and funded by NASA/GMAO core provided under the Modeling, Analysis, and Prediction Program. The software for simulating GPSRO observations was provided by the Radio Occultation Processing Package (ROPP) of the Radio Occultation Meteorology (ROM) Satellite Applications Facility (SAF) of EUMETSAT, with the assistance of Sean Healy at ECMWF. Resources supporting this work were provided by the NASA High-End Computing (HEC) Program through the NASA Center for Climate Simulation (NCCS) at Goddard Space Flight Center. Helpful comments by three anonymous reviewers lead to significant improvements in the manuscript.

Data availability statement

The dataset on which this paper is based is too large to be retained or publicly archived with available resources. Documentation and methods used to support this study are available from Nikki Privé at NASA/GMAO.

Disclosure statement

No potential conflict of interest was reported by the authors.