Could 5G make weather satellites less effective?

Question

I've seen a few articles pop up in sources stating that 5G signals will run too close to the 23.8 MHz signal that weather satellites need in order to accurately produce weather forecasts.

• National Geographic

  One of the crucial bands, says William Blackwell, an atmospheric scientist and engineer at MIT, is around 23.8 GHz. Water vapor absorbs in this microwave band, leaving behind a faint signal that can be read by satellite-mounted instruments that look at the microwave part of the spectrum. The problem now is that telecommunications companies are interested in using parts of the spectrum right next to that water vapor signal.

• Wired

  Jordan Gerth, a research meteorologist at the University of Wisconsin, Madison, has been studying this issue as part of a group at the American Meteorological Society. He says that while the FCC can switch which regions of the spectrum it allocates to phone companies, forecasters are stuck. That’s because water vapor emits a faint signal in the atmosphere at a frequency (23.8 GHz) that is extremely close to the one sold for next-generation 5G wireless communications (24 GHz). Satellites like NOAA's GOES-R and the European MetOp monitor this frequency to collect data that is fed into prediction models for upcoming storms and weather systems.

  I've also seen a few dissenting articles claiming that the predictions of a negative impact on weather satellites rely on bad, outdated data, or that they don't take into account how 5G will differ from 4G.

• The Hill

  It certainly sounds bad, but NOAA’s doomsday predictions are devoid of scientific justification. When federal agencies raised interference concerns in 2017, the evidence they presented was ultimately withdrawn after its methodological shortcomings were exposed. NOAA’s current claims are based on a similarly flawed study, which has not been released publicly, that doesn’t take into account how 5G networks work in the real world. For instance, it fails to consider that 5G — unlike 4G — will use beamforming to precisely transmit signals to a particular user and prevent out-of-band emissions. In addition, the 5G spectrum will be more than 250 MHz away from the passive weather sensors NOAA is concerned about, providing further protection from interference.

  It goes on to argue the relevant weather sensor isn't even in use.

• PC Mag

  The headline says "No, 5G Won't Ruin Your Weather Forecasts", but the body basically concedes that the 24 GHz bandwidth will need be used with limited power to protect the weatehr forecasts.

I'm inclined to agree with the dissenters, but I also recognize that I am not an expert in signals. To my mind, 200 - 250 MHz seems like a pretty wide gap in between two signals, easy enough for any satellite worth its salt to focus on and filter down to, especially if it's a "focused" signal like 5G is supposed to be.

Will 5G broadcasts in the 24 GHz impact weather forecasts?

Answer 1

This issue was discussed during the last World Radiocommunication Conference, held in November 2019 by the International Telecommunication Union (ITU), under Agenda Item 1.13. The ITU has worked on this topic since 2015, and many countries, industries and international organizations have submitted studies on the compatitbility between 5G (or IMT in ITU's world) and meteo satellite services (or Earth Exploration Satellite Service, EESS). You can find the characteristics used for EESS sensors in ITU Recommandation RS.1861 and interference criteria in Rec. RS.2017. The antenna pattern for 5G is given in REC. M.2101.

You can see that for sensors at 24 GHz, it is recommended that a power of -166 dBW/200 MHz should not be exceeded more than 0.01% of a year. Otherwise, the measurement would be degraded.

With those characteristics, ITU members run simulations of compatibility to find the maximum authorized power in the EESS band. One of the biggest issues is the assumptions that you make for your study. And that is where Agenda Item 1.13 was difficult. If you have a look at this document (the source is the World Meteorological Organization, which is a biaised source in this context but this is the only document that I found available), and more precisely at the section 3 on assumptions, you can have an idea of the points of disagreement: the antenna pattern for 5G, the number of base stations and the aggregate impact of all services (IMT is not the only one using the band) on EESS. And depending on the assumptions and the studies (there were studies from 5G industrials, from scientific agencies and from countries), you have different results, and the range of possibilities is wide according to this document (look for Agenda Item 1.13, page 158 of the English version). This document also gives an idea of all disagreements between the involved speakers.

Therefore at the question, "could 5G make weather satellite less effective", I think the answer is yes, considering that under assumptions agreed by many scientific agencies and countries, studies show that the value agreed during WRC-19 is not enough to respect the protection criteria of weather sensors.

Answer 2

An article on Nature web page suggests the reason behind the concern.

The 5G transmissions will involve many frequencies, but the key one under discussion is 23.8 gigahertz. Water vapour in the atmosphere naturally produces a weak signal at this frequency, which satellites use to measure humidity. Those data feed into weather forecasts. But if a 5G station is transmitting a signal near the 23.8-gigahertz frequency, a weather satellite might pick it up and interpret it as water vapour. And that bad data could degrade forecasts.

It's an interesting bit of science to discover that water vapor in the atmosphere generates radio frequencies. I'd known that clouds can generate static electricity (lightning!) but figured that was pretty broad spectrum.

It's understandable that a 5G signal created by man could certainly overpower receivers on a satellite designed to pick up the weak stuff generated in the atmosphere.

No reference is made in the article about directional aspects of the signal. One could expect that it's as omni-directional as can be managed, as you would not want to exclude the 5G user by intentionally creating dead zones. The exception to that, referenced in the article, is to shut off the 5G signal when the respective satellites are taking measurements.