| description abstract | Two-week experimental forecasts were carried for 12 July Cases with a nine vertical level, 270 km grid-size hemispheric model, and the results were examined statistically. The solutions studied were the stationary (10-day average) and transient components of the general circulation; in particular, the ensemble mean of temperature, zonal wind and eddy kinetic energy; and the hemispheric maps of ensemble mean height fields. The predictive ability of this model was examined by comparing the results with observation and calculating statistical scores such as standard deviation, correlation coefficient and horizontal gradient (S1) score for 1000, 500 and 50 mb geopotential height verified against the NMC (National Meteorological Center) analysis. The results were also analyzed in terms of zonal wavenumbers of geopotential. So far as this model (1967 version) is concerned, the predictability in the lower atmosphere seems to decay more rapidly in July than in January. However, the predictability of the middle and upper troposphere is greater in July than in January; the planetary-scale waves were better predicted. The simulation of the summer stratosphere is very poor with the nine-level vertical resolution. The prediction was also compared with that of a spectral model of comparable horizontal and vertical resolution. | |
