Forecasting Thermosphere Density: an Overview

Abstract:

Our knowledge of the thermosphere has improved considerably since 2000 thanks to the availability of high-resolution accelerometer inferred densities. Consequently, precision and shortcomings of thermosphere models are better known. Thermosphere density forecast accuracy is limited by: 1) the accuracy of the thermosphere model 2) the solar and geomagnetic activity forecast 3) the quality of the data assimilation system. The precision of semi-empirical thermosphere models is 10-25%. Solar activity forecasts can be accurate up to 5 days. They become less accurate with time, but some proxies are more forecastable than others. Geomagnetic activity forecasting is more problematic, since in most cases storm events cannot be predicted on any time scale.

The forecast accuracy is ultimately bounded by the thermosphere model precision and the (varying) degree to which mainly the solar proxy represents EUV heating of the atmosphere. Both errors can be corrected for by means of near real time (nrt) assimilation of satellite drag data, provided that the data is of high quality. At present, only the classified High Accuracy Satellite Drag Model of the Air Force has that capability operationally, even if other prototype nrt models have been developed. Data assimilation significantly improves density forecasts up to 72-hours out; there is no gain for longer periods due to the short memory of the thermosphere system. Only physical models, e.g. TIMEGCM and CTIPe, can in principle reproduce the dynamic changes in density for example during geomagnetic storms. However, accurate information on atmospheric heating is often missing, or not used. When it is, observed and modeled Traveling Atmospheric Disturbances are very similar. Nonmigrating tides and waves propagating from the lower atmosphere cause longitudinal density variations; sources of geophysical noise for semi-empirical models, they can be predicted qualitatively and sometimes quantitatively with physical models.

This presentation will give a summary of density modeling and forecasting by addressing all the above listed items.