Authors: Watkins, Zachary (Georgia State University), Jones, William (Georgia State University), Sadykov, Viacheslav (Georgia State University), Kempton, Dustin (Georgia State University), He, Xiaochun (Georgia State University), Tobiska, W Kent (Space Environment Technologies), Mertens, Christopher (NASA Langley Research Center), Ranjan, Shubha (NASA Ames Research Center), Kitiashvili, Irina (NASA Ames Research Center), Spaulding, Ryan (NASA Ames Research Center), Deardorff, Donald Glenn (NASA Ames Research Center)

Continuous monitoring and analysis of the radiation environment in the Earth’s lower atmosphere are critical for the safety of aircraft and spacecraft crews and passengers. Addressing the problem requires a complex approach of integrating different data sources and enhancing the visualization and search capabilities. In addition, the development of the data-driven radiative environment prediction models demands a significant investment in data preparation. This work highlights the progress of expanding the Radiation Data Portal (RDP) database and construction of the Machine Learning-ready (ML-ready) data set for the prediction of the effective dose rates at airplane heights. The expanded version of RDP allows the users to explore the most recent measurements obtained from the Automated Radiation Measurements for Aerospace Safety (ARMAS) device obtained during the 2013-2023 time range and data sources describing the terrestrial and space environments (such as the measurements of the cosmic rays, solar wind, energetic particles, and geomagnetic activity). The ARMAS data of a quality suitable for ML purposes were down-selected semi-manually. The algorithm for separating the data into three partitions and preserving the same statistical properties for each parameter within each partition was developed and tested. We also discuss the results of the statistical analysis of radiation measurements and their comparison with the predictions of the Nowcast of the Atmospheric Ionizing Radiation for Aerospace Safety (NAIRAS V2) model.