The cooling of the atmosphere is generally dominated by the infrared emission of nitric oxide (NO) and carbon dioxide (CO 2) in the lower thermosphere and mesosphere. During extreme geomagnetic storms, the NO infrared emission is significantly enhanced because both the temperature increase by Joule heating and the impact ionization and excitation associated with particle precipitation can greatly raise the NO production rate, whereas the CO 2 infrared emission shows a more modest increase (Lu et al., 2010). Consequently, NO emission becomes the major thermospheric cooling source during and after storms (Mlynczak et al., 2005, 2018). An accurate determination of NO emission is important for the prediction of stormtime thermospheric temperature and density. The effects of NO emission during storm times are usually estimated by using theoretical models (Chen & Lei, 2018; Sheng et al., 2017). In most of the numerical models, however, there are considerable uncertainties in NO emission reaction rates which are critical to NO emission simulations. The reaction rates of NO production and emission have been investigated both experimentally and theoretically (Duff et al., 2003; Herron, 1999), but they have not fully been validated. A series of studies have suggested that the simulated NO emission has obvious discrepancies with data when compared with the Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) observations (