MSIS‐UQ: Calibrated and Enhanced NRLMSIS 2.0 Model With Uncertainty Quantification

The Mass Spectrometer and Incoherent Scatter radar (MSIS) model family has been developed improved since the early 1970's. most recent version of MSIS is Naval Research Laboratory (NRL) 2.0 empirical atmospheric model. NRLMSIS provides species density, mass temperature estimates as function location space weather conditions. models have long a popular choice thermosphere in research operations community alike, but—like many models—does not provide uncertainty estimates. In this work, we develop an exospheric based machine learning that can be used with to calibrate it relative high-fidelity satellite density directly through parameter. Instead providing point estimates, our (called MSIS-UQ) outputs distribution which assessed using metric called calibration error score. We show MSIS-UQ debiases resulting reduced differences between 25% 11% closer than Space Force's High Accuracy Satellite Drag Model. also model's estimation capabilities by generating altitude profiles for temperature. This explicitly demonstrates how probabilities affect within 2.0. Another study displays post-storm overcooling alone, enhancing phenomena capture.