Date: November 10, 2022

Dataset Title: Improving Forecasting Ability of GITM using Data-driven Model Refinement

Dataset Creators: B. M. Ponder & A. J. Ridley & A. Goel & D. S. Bernstein

Dataset Contact: Brandon Ponder bponder@umich.edu

Funding: 2028125 (NSF), 80NSSC20K1581 (NASA), FA9550-16-1-0071 (USAF)

Key Points:
 - Inaccurate approximations to physics terms and incorrect drivers within GITM can be corrected for using data-driven model refinement.

- Dynamic adjustments to the parameterized thermal conductivity coefficients can compensate for errors in model predicted mass densities.

- Comparative statistics were computed when GITM was configured in a biased version, an out-of-the-box version and the refined version.


Research Overview:
Physics-based models have a difficult time accurately estimating the upper atmosphere density. These densities are needed to compute satellite orbit trajectories to monitor for potential collisions. Inaccurate density estimation can be due to variety of factors and so methods of correcting the model-predicted density are needed. We are presenting a method to correct the densities using available satellite measurements from the CHAMP and GRACE satellites and the commonly used empirical model NRLMSISE-00. Upon reducing the model error, we show the improved ability of a physics-based model to capture a geomagnetic storm.


Methodology:
These data were generated by running the Global Ionosphere-Thermosphere Model at various time periods (shown below) and extracting the modeled thermospheric mass densities along the those corresponding to the orbit of the CHAMP or GRACE satellite and the subsolar point from the empirical model MSIS. The data is then post-processed and visualized in a combination of Python and MATLAB scripts. 

Files contained:
Please refer to specific README.txt's in each Fig/ directory.

Related publication(s):
Ponder, Brandon & Ridley, Aaron & Goel, Ankit & Bernstein, Dennis. (2022). Improving Forecasting Ability of GITM using Data-driven Model Refinement. 10.1002/essoar.10512265.1.  In Press.


Use and Access: 
This data set is made available under a Creative Commons Public Domain license (CC0 1.0).


To Cite Data: 
Ponder, Brandon & Ridley, Aaron & Goel, Ankit & Bernstein, Dennis. (2022). Improving Forecasting Ability of GITM using Data-driven Model Refinement. 10.1002/essoar.10512265.1.  In Press.