Temperature and water vapour soundings from satellites have been available for more than 20 years. Because of the global nature of these data, satellite soundings were expected to improve significantly global numerical weather prediction (NWP). Indeed, experiments with the satellite soundings provided during the Global Weather Experiment (GWE) in 1979 revealed the positive impact of these data on extended-range forecasts, particularly in the southern hemisphere. Unfortunately, since the GWE, the value of satellite soundings has apparently diminished. I believe this to be due largely to improvements in the models which have surpassed the inherently poor vertical resolution of these data relative to that of the contemporary NWP models. The satellite data can even degrade the analysis/forecast operation through the aliasing of small-scale features which may be properly represented by the vertical scale resolved by the model.

This lecture addresses, from a historical perspective, the dilemma of the decreasing value of satellite soundings in the weather-forecast application (i.e., the reasons for the unfulfilled expectations of these data). Attention is then turned to the improvements in the satellite technology and the application techniques needed to reverse this trend.