Abstract
Time series of hemispheric available potential (A) and kinetic (K) energies were used to examine the results of a series of observing system simulation experiments that were performed with the Goddard Laboratory for Atmospheres model to determine the impact of the proposed space-based wind profiler on forecast accuracy. The simulated data for the series of 5-day forecasts were produced from a 20-day integration using the ECMWF model, which was also used to produce the verification forecast for the 5-day period. The three simulated observational sets of data that represented conventional observations, satellite-retrieval temperatures, and wind profiles were produced by NMC.
The results in the Northern Hemisphere show that the magnitudes of A and K from the simulation forecasts are quite similar to each other and are uniformly higher than the verification forecast, reflecting systematic differences in the energy levels of the two models. In the Southern Hemisphere, differences in magnitude of A between simulation and verification forecasts are larger than in the Northern Hemisphere. The time series for K shows greater diversity in magnitude among the simulation forecasts, with all the simulation forecasts for K being higher than the verification forecast. The S1 skill scores and root-mean-square (rms) differences reveal little variation in the accuracy of the forecasts among the three simulation datasets in the Northern Hemisphere. In the Southern Hemisphere, however, forecasts using satellite temperature and wind-profiler data have much smaller rms differences and S1 scores, indicating an improvement in forecast accuracy over conventional observations. The addition of wind-profiler data provides the greatest improvement in forecast accuracy.
Geographical distributions of vertically integrated eddy A (Ae and K in the Northern Hemisphere reveal that these quantities in the three simulation forecasts are more similar to each other than with the verification forecast. In the Southern Hemisphere, the geographical distributions of Ae and K are more varied with the wind-profiler dataset producing a forecast closest to the verification forecast. In general, the impact of the addition of wind-profiler data on forecast accuracy of energy parameters is negligible in the Northern Hemisphere but distinctly positive in the Southern Hemisphere.
