Atmospheric corrosion of aluminum alloy AA2024-T3 samples was monitored at three different sites in
Florida: NRL-Key West, Kennedy Space Center and Daytona Beach/Battelle site. The samples were
exposed over different periods of exposure and for different time periods. A systematic approach was
used to correlate exposure conditions with the amount of corrosion as measured by mass loss. An
analytical approach was developed that involved collecting meteorological data for the three sites from
multiple sources, performing data quality checks, and analyzing the trends in both the corrosion data and
the meteorological data. Machine learning through feature selection and regression approaches was
used to identify leading meteorological factors that quantitatively control extent of corrosion. Key features
determined to have a quantitative effect on corrosion rate and mass loss per unit area were collected and
include mean precipitation, the range of temperatures, the minimum wind speed, the standard deviation
of ozone exposure, and the maximum solar irradiance. This approach could be applied to other materials
of interest, different locations, and adapted to other corrosion metrics such as localized corrosion depth
and pit volumes.