TROPICS observations of atmospheric vertical structure: performance assessments and comparisons with ATMS

New constellations to provide high-resolution atmospheric observations from microwave sounders operating in low-earth orbit are now coming online and are providing operationally useful data. The first of these missions, the NASA TROPICS (Time-Resolved Observations of Precipitation structure and storm Intensity with a Constellation of Smallsats) Earth Venture (EVI-3) mission, was launched 2023, and three of the four satellites remain operational. Publicly available TROPICS observations of temperature and humidity vertical profiles, as well as cloud ice and precipitation horizontal structure, are currently being used to conduct high-value science investigations of tropical cyclones. The constellation currently provides rapid-refresh microwave measurements with a median refresh rate of 70-90 minutes with three functional CubeSats. The TROPICS scanning microwave radiometer payload observes in 12 channels spanning approximately 90 to 205 GHz with spatial resolutions ranging from approximately 15 to 30 km. To assess the ability of TROPICS to provide estimates of atmospheric thermodynamic structure, a neural network atmospheric vertical profile (NNAVP) retrieval algorithm has been developed to infer non-linear and non-Gaussian relationships between the TROPICS brightness temperature measurements and the temperature and water vapor vertical profiles. The TROPICS NNAVP algorithm was trained against ERA-5 and validated with over 12,000 global radiosonde observations, yielding uncertainties that are compliant with the TROPICS baseline mission requirements of 2 K (in 3-km vertical layers spanning 0-20 km) for temperature profiles and 25 % (in 3-km vertical layers spanning 0-10 km) for water vapor profiles. The TROPICS NNAVP algorithm was also adapted for use with the Advanced Technology Microwave Sounder (ATMS) to permit direct comparisons of the instrument performance. In this paper, the TROPICS and ATMS NNAVP performance will be presented and compared. Additionally, the NOAA MIRS algorithm will be used to generate profile estimates from ATMS, and the NNAVP performance will be compared to NOAA MIRS performance.