Premium
Assessing the impact of pre‐ GPM microwave precipitation observations in the Goddard WRF ensemble data assimilation system
Author(s) -
Chambon Philippe,
Zhang Sara Q.,
Hou Arthur Y.,
Zupanski Milija,
Cheung Samson
Publication year - 2013
Publication title -
quarterly journal of the royal meteorological society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.744
H-Index - 143
eISSN - 1477-870X
pISSN - 0035-9009
DOI - 10.1002/qj.2215
Subject(s) - data assimilation , weather research and forecasting model , environmental science , meteorology , numerical weather prediction , precipitation , climatology , mesoscale meteorology , quantitative precipitation forecast , geography , geology
The forthcoming Global Precipitation Measurement ( GPM ) Mission will provide next‐generation precipitation observations from a constellation of satellites. Since precipitation by nature has large variability and low predictability at cloud‐resolving scales, the impact of precipitation data on the skills of mesoscale numerical weather prediction ( NWP ) is largely affected by the characterization of background and observation errors and the representation of nonlinear cloud/precipitation physics in an NWP data assimilation system. We present a data impact study on the assimilation of precipitation‐affected microwave ( MW ) radiances from a pre‐ GPM satellite constellation using the Goddard WRF Ensemble Data Assimilation System (Goddard WRF ‐ EDAS ). A series of assimilation experiments are carried out in a Weather Research Forecast ( WRF ) model domain of 9 km resolution in western Europe. Sensitivities to observation error specifications, background error covariance estimated from ensemble forecasts with different ensemble sizes, and MW channel selections are examined through single‐observation assimilation experiments. An empirical bias correction for precipitation‐affected MW radiances is developed based on the statistics of radiance innovations in rainy areas. The data impact is assessed by full data assimilation cycling experiments for a storm event that occurred in France in September 2010. Results show that the assimilation of MW precipitation observations from a satellite constellation mimicking GPM has a positive impact on the accumulated rain forecasts verified with surface radar rain estimates. The case‐study on a convective storm also reveals that the accuracy of ensemble‐based background error covariance is limited by sampling errors and model errors such as precipitation displacement and unresolved convective scale instability.