Identification of Planetary Boundary Layer (PBL) Regimes Using a Flexible, Multi-Sensor Deep Learning Algorithm

The Planetary Boundary Layer (PBL) remains an immensely complex part of the atmosphere, where many vital earth system processes occur. Current satellite microwave technology and retrieval methods are not sensitive enough to detect the PBL temperature and water vapor structure. Recent developments in hyperspectral microwave (HMW) technology show the capability to perform PBL-sensitive thermodynamic sounding of the atmosphere. In addition, recent developments in the use of AI and denoising techniques on space-borne backscatter lidars (BSL) show the capability to identify the PBL height more accurately. Using a flexible artificial intelligence data fusion retrieval, optimal HMW instrument configurations can be tested and PBL height can be retrieved using a combination of passive (HMW) and active (BSL) remote sensing instruments. This work will demonstrate how our deep learning model can capture a wide range of PBL regimes and their associated temperature and water vapor profile variability, to better optimize future model configurations. The outcome of these experiments will facilitate the design of our upcoming suborbital field campaign and inform on current technology gaps and instrument development needs that are key to plan the next generation of NASA and NOAA space and suborbital missions.