Space weather forecasting is significant because of two reasons; the spread of Low Earth Orbit (LEO) technology such as satellites, and it’s capacity to affect GPS systems, electricity grids, aircraft and sea navigation systems. Space weather studies the interaction of solar plasma and solar wind in the magnetosphere (the upper atmosphere which houses the Hubble Telescope and International Space Station). With accurate forecasts of geomagnetic storms, operators could reposition the satellite or send weather warnings preventing grid failures and save millions of dollars. Researchers also track the 6000 tons of space junk to avoid collision with moving satellites. The biggest challenge is atmospheric drag (force extended on object) which increases with solar activity and affects several variables. Because the variables are in thousands, AI models are useful in predicting local and global space weather. Statler researchers are trying to reduce drag uncertainty in AI models to improve forecasts. Research focuses on improving model accuracy and accommodating uncertainty estimates in forecasts. One study used 20 years of input in a Space Environment Technologies High Accuracy Satellite Drag Model (HASDM) to improve models. Researchers worked two techniques: dropping a variable and adjusting probability distribution (calculation of different possible outcomes) in these models.