To test the impact of high-resolution Nimbus-6 sounding data on Australian region forecasts, two parallel analysis/forecast cycling experiments were carried out, using data for 14 days during August and September 1975. In one of these cycles, only conventional data and manual interpretation of satellite imagery were used as input, while the other cycle used conventional and Nimbus-6 sounding data. A manual mean sea level pressure analysis was used in each cycle to provide reference level information over the oceans.
Two series of 24 h limited area prognoses were prepared from these two sets of analyses, using the primitive equations prognosis model developed at the Australian Numerical Meteorology Research Centre. An average improvement in geopotential forecasts of more than 5 skill score points was achieved at all levels over the Australian continent when the Nimbus-6 data were included in the base analyses. Also, significant reductions were obtained in 24 h forecast root-mean-square (rms) temperature errors.
1 On assignment to the ANMRC during 1977 from National Environmental Satellite Service, NOAA, Washington, D.C. 20233.
