How Does the Growth of 5G mmWave Deployment Affect the Accuracy of Numerical Weather Forecasting?

The allocation of the 5G mmWave spectrum in the 26 GHz range, known as 3GPP band $\mathbf{n 2 5 8}$, has raised wide concern among the remote sensing and weather forecast communities due to the adjacency of this band with a frequency band used by passive sensors in Earth Exploration-Satellite Service (EESS). The concern stems from the potential radio frequency interference (RFI) caused by transmissions in the $\mathbf{n} 258$ band into the 23.8 GHz frequency, one of the key frequencies employed by weather satellite passive sensing instruments, such as AMSUA and ATMS, to measure atmospheric water vapor using its emission spectrum. Such RFI can bias satellite observations and compromise weather forecasting. In this paper, we develop a modeling and numerical framework to evaluate the potential effect of the 5G mmWave $\mathbf{n} 258$ band’s commercial deployment on numerical weather forecast accuracy. We first estimate and map the spatio-temporal distribution of 5G mmWave base stations at the county-level throughout the contiguous United States (US) using a model for technology adoption prediction. Then, the interference power received by the AMSU-A radiometer is estimated for a single base station based on models for signal transmission, out-of-band radiation, and radio propagation. Then, the aggregate interference power for each satellite observation footprint is calculated. Using the contaminated microwave observations, a series of simulations using a numerical weather prediction (NWP) model are conducted to study the impact of $\mathbf{5 G}$-induced contamination on weather forecasting accuracy. For example, our results show that when the interference power at the radiometer from a single base station is at a level of $-\mathbf{1 7 5} \mathrm{dBW}$ for a network of base stations with spectral efficiency of $15 \mathrm{bit} / \mathrm{s} / \mathrm{Hz} / \mathrm{BS}$, the aggregate interference power has limited impact in the year 2025 but can result in an induced noise in brightness temperature (contamination) of up to 17 K in the year 2040. Furthermore, that level of RFI can significantly impact the $\mathbf{1 2}$-hour forecast of a severe weather event such as the Super Tuesday Tornado Outbreak with forecasting errors of up to 10 mm in precipitation or a mean absolute error of $\mathbf{1 2. 5 \%}$. It is also estimated that when the level of interference power received by the radiometer from a single base station is $-\mathbf{2 0 0} \mathrm{dBW}$, then there is no impact on forecasting errors even in 2040.