Summer-Season Forecast Experiments with Upgrades in the Land Component of the NCEP Climate Forecast System (CFS)

To improve seasonal prediction skill and to examine the impact from land upgrades on the National Centers for Environmental Prediction (NCEP)'s Climate Forecast System (CFS), coupled T126 CFS experiments are carried out for 9 selected summers with 4 ensemble members using the latest version of Noah land model with new soil and vegetation datasets, and compared with the current operational CFS where an old version of the Noah land model is employed to compute surface fluxes. The nine years are composed of three ENSO-cold, three ENSO-warm, and three ENSO-neutral years. The initial land states are taken from the NCEP CFS Reanalysis (CFSR) and the initial dates are from May 1-4 with initial time of 00 UTC. Using anomaly correlation as a primary measure, the summer-season prediction skill of the CFS using the two versions of the Noah LSM is assessed for SST, precipitation, and 2-meter air temperature over the Contiguous United States (CONUS) on an ensemble basis and at the seasonal time scale. Results from the CFS experiments indicate that the upgrades in the Noah land model has a positive impact on the Sea Surface Temperature (SST) prediction skill over the Atlantic Ocean, the Indian Ocean, and the western Pacific Ocean in the Northern Hemisphere, but not over the eastern equatorial Pacific Ocean. The differences in June-July-August (JJA) mean precipitation and 2-meter air temperature prediction skill generally reflect the disagreement in predicting SST anomalies over the coastal states. The largest differences are found to be in the inland states where the CFS with the new version of Noah land model shows a better performance in predicting precipitation anomaly over the central Great Plains and 2-meter air temperature anomaly over the western CONUS, illustrating that ocean is still the main control and the land impact is only pronounced where the land-atmospheric coupling is strong.