Global Observing System Simulation Experiments at National Centers for Environmental Prediction

OSSE at NCEP has been focused on evaluating data impacts on a global scale. A 31-day run (February 5th to March 7th 1993) by ECMWF has been used as the Nature Run (NR). The NCEP operational data assimilation system is used to assimilate the data. Adjustment of the NR and calibration using existing instruments were performed (Masutani et al. 1999). The results were satisfactory and justified conducting more OSSEs to evaluate future instruments. During the calibration process, a larger impact from wind as compared to temperature data was confirmed.

Forecast impacts from space based Doppler wind lidar (DWL) wind data were assessed. The line-of-sight (LOS) components of wind were directly assimilated, instead of using the processed horizontal component of the wind. In the case of RAOB and TOVS radiance, the impacts are mainly observed in the mesoscale (Masutani et al. 2004). Forecast improvements over large scales are achieved when correlated observational error is added to simulated conventional data. Non-scan lidar data showed a similar amount of forecast impact on TOVS radiance data, which is much smaller than the impact from RAOB wind. Magnitudes of impact similar to that of the RAOB wind is achieved only by DWL with scanning (Masutani et al. 2002).

AIRS and AMSU data are simulated by NOAA/NESDIS, and the simulation of CrIs data is nearly ready (Kleespies et al. 2003). Strategies and software for the simulation of cloud motion based on winds have been developed by Simpson Weather Associates (O'Handley et al. 2001). The OSSE system is being transferred to NCEP's IBM SP system in order to accommodate high resolution data, such as AIRS radiance. The IBM system is also capable of performing higher resolution data assimilation. The combined impacts from high resolution radiance data, cloud motion wind, and DWL will be pursued.

In order to pursue a more detailed assessment of the future data, a NR with higher resolution over a longer period is required. The experience with current NCEP OSSEs will be valuable in designing the next generation of NR.