Thermospheric density impacts satellite position and lifetime through atmospheric drag. More accurate specification of thermospheric temperature, a key input to current models such as the High Accuracy Satellite Drag Model, can decrease model density errors. This paper improves the model of Burke et al. (2009) to model thermospheric temperatures using the magnetospheric convective electric field as a driver. In better alignment with Air Force satellite tracking operations, we model the arithmetic mean temperature, T 1/2, defined by the Jacchia (1977) model as the mean of the daytime maximum and nighttime minimum exospheric temperatures occurring in opposite hemispheres at a given time, instead of the exospheric temperature used by Burke et al. (2009). Two methods of treating the solar ultraviolet (UV) contribution to T 1/2 are tested. Two model parameters, the coupling and relaxation constants, are optimized for 38 storms from 2002 to 2008. Observed T 1/2 values are derived from densities and heights measured by the Gravity Recovery and Climate Experiment satellite. The coupling and relaxation constants were found to vary over the solar cycle and are fit as functions of F 10.7a, the 162 day average of the F 10.7 index. Model results show that allowing temporal UV variation decreased model T 1/2 errors for storms with decreasing UV over the storm period but increased T 1/2 errors for storms with increasing UV. Model accuracy was found to be improved by separating storms by type (coronal mass ejection or co‐rotating interaction region). The model parameter fits established will be useful for improving satellite drag forecasts.
Frey, W. R., Lin, C. S., Garvin, M. B., & Acebal, A. O. (2014). Modeling the thermosphere as a driven-dissipative thermodynamic system. Space Weather, 12(3), 132–142. https://doi.org/10.1002/2013SW001014