Revisiting Atmospheric Features of Mars Orbiter Laser Altimeter Data Using Machine Learning Algorithms

The Mars Orbiter Laser Altimeter (MOLA) instrument has been drawing a map of Mars' topography between September 1997 and June 2001. It has also been able to observe clouds during the mission duration, providing data for the low Martian atmosphere for nearly 1.5 Mars years. The Mars Global Surveyor, which carried MOLA, also carried two other instruments that also observed clouds during the same time period (the Mars Orbiter Camera and the Thermal Emission Spectrometer). Combining observations from these three data sets could provide a complete recap of most atmospheric structures during MY24 and MY25. However, previous studies of MOLA data set often had to use stringent detection criteria. Using machine learning clustering methods, we end up finding way more atmospheric returns. Our results are presented in the form of an atmospheric features catalog that we then use to compare MOLA observations with Mars Orbiter Camera and Thermal Emission Spectrometer results, but also with more recent missions. We study the development of recurrent phenomenon in the Martian atmosphere, like the aphelion cloud belt or the south polar hood, but also spontaneous events such as regional dust storms. Methods could be tuned even more finely by using more complex clustering methods or deep learning algorithms to clearly distinguish atmospheric structures.