Modeling Microwave Radiometer Sensing Performance as a Function of Spectrum Use

Passive microwave radiometers operating in the Earth Exploration-Satellite Service (EESS-passive) bands are increasingly exposed to unwanted out-of-band emissions due to growing competition for spectrum from emerging 5G, 6G, and beyond wireless technologies. At the same time, new and advanced radiometer missions are driving demand for wider bandwidths to improve geophysical retrieval accuracy, further increasing exposure to adjacent-band transmissions. Motivated by the growing interest in spectrum sharing, this work proposes a modeling framework for quantifying radiometer retrieval performance under a range of alternative spectrum access scenarios. The proposed modeling framework uses a physics-based forward model to simulate ideal radiometer measurements, to which noise is added as a function of channel bandwidth, and applies linear and quadratic retrieval models with and without logarithmic linearization of the input data. For integrated water vapor retrievals using the 18.7, 23.8, and 36.5 GHz channels, the quadratic retrieval model achieves the best overall performance, yielding numerous bandwidth combinations for which the RMSE is below 1 mm. Both retrieval models benefit substantially from the logarithmic transformation, which further expands the range of bandwidth combinations achieving sub-1 mm RMSE. These results indicate that spectrum sharing opportunities can be maximized through retrieval designs that reduce model mismatch error.