Emulating Passive Microwave Satellite Products: Information Content and Uncertainties

Abstract

The ability to continuously monitor Earth systems is critical to understanding and predicting our weather and climate. To achieve this goal over land, we confidently rely on ground-based observations. Over oceans, however, measurements are rarely possible without using satellite sensors. More than half a century into their development, satellite instrumentation and retrievals offer indispensable products to modern meteorology and research. Yet, user and science communities still lack desired information on the atmosphere and its underlying processes. This talk will explore the possibility of using the existing broadly available spatiotemporally “continuous” satellite products, such as geostationary InfraRed (IR) radiances, to replicate the information typically offered by sparse low-earth-orbit passive microwave sensor observations. Using nothing but the GOES-R series Advanced Baseline Imager (ABI) IR bands on input, we predict (emulate) otherwise unavailable PMW brightness temperatures at frequencies and accuracy suitable for operational use in precipitationrelated applications. A machine learning approach for delivering quantitative traceable uncertainty estimates will be discussed.