Jeremy Lien
Research Mentor(s): Cheng Li
Mentor Department: Climate and Space Sciences and Engineering
Authors: Jeremy Lien, Alexander Zvoch
Session: Session 2 (10:00am – 10:50am)
Presentation Type: Poster 23
Abstract
Space is the ultimate frontier for human beings, and Mars is the nearest candidate planet for future human exploration and dwelling. Thus, a precise method to predict Martian weather forecasts will become a necessity in the near future. The distinct characteristics of the Martian atmosphere, including its distinct chemical composition and drastic seasonal and diurnal temperature changes, present significant challenges for accurate weather prediction. Therefore, developing reliable weather forecast models is a critical step in ensuring the success of future manned missions, explorations, and potential colonialization efforts. An accurate weather prediction model can be developed utilizing the wealth of topographic, temperature, pressure, dust storm activity, and wind pattern data collected by instruments involved in past missions such as the Mars Orbital Laser Altimeter (MOLA) aboard the MGS satellite. After the raw data is collected, several algorithms are used to sort, parse, bin, and resize the data to create a tensor-based, three-dimensional representation of the dataset that is more computationally efficient. Next, the processed data, along with data from other orbiters, landers, and rovers, is used to train a model in PyTorch to create a realistic simulation model of Mars’s weather. Ultimately, such a model will not only enhance the scientific understanding of Martian meteorology but also contribute to safer and more effective mission planning and operations, advancing our goal of sustainable exploration and possible human settlement on the red planet.
