Nieke Roos
14 October 2020

Wageningen University & Research, the Royal Netherlands Meteorological Institute (KNMI) and water research institute Deltares have used the signal strength of commercial mobile phone masts to accurately measure rainfall between two points. It also allowed them to make a weather forecast up to several hours ahead. The scientist derived the intensity of rain falling between two transmission masts from so-called radio beam connections in cell phone networks, aka commercial microwave links (CMLs).

“The communication companies are interested in having the clearest possible signal for mobile telephony. But we are looking at exactly the opposite,” explains Ruben Imhoff of the Wageningen research group Hydrology and Quantitative Water Management. The signal attenuation between two masts can be caused by rain between those two measuring points. “The harder it rains, the stronger the interference we get back in our data. We map out this disturbance and translate it into rain intensity, and then into very short-term precipitation forecasts, so-called nowcasting.”

The research team compared the data from KNMI rain radars, such as those used by Buienradar, with the data from the masts of the telecom companies on twelve days with rain. “We were pleasantly surprised when we saw the results,” says Imhoff. “The CML method turns out to work quite well, even at high precipitation intensities.” Sometimes the precipitation forecasts were even better than those of the rain radar, especially with heavy rainfall. “But then you have to take into account that the operational rain radar products weren’t corrected for anomalies. There’s still room for improvement in the case of radar as well, and we are working on that.”

WUR rainfall estimation
Comparison of estimated precipitation intensities for the Netherlands on 10-9-2011 22:00. The reference (radar image corrected with the rain measurements of the KNMI network), the operationally available radar image (U-Rad) and the precipitation intensities estimated with CML data. Credit: Wageningen University & Research

Can CML be an alternative to radar, which looks with a precision of one square kilometer, also above the sea and the Ijsselmeer? Imhoff: “Radar provides a stable image over a large area. You have radio beam connections at high densities in cities but less in rural areas. And the rainfall radar refreshes every five minutes. Currently, we receive the CML data every 15 minutes. A frequency increase, eg every five minutes, can bring about another improvement. That would make CMLs a valuable addition to the radar data in the Netherlands. These combined images can then be used for nowcasting, for example to better anticipate local showers.”

Bits&Chips event Save the date

This also applies to the new 5G masts. They’re positioned at even shorter distances from each other and therefore provide more accurate data. But the researchers can’t say how this will work out because it is not yet clear how well the CML system performs with the higher transmission frequencies of 5G. A research proposal on that topic is in preparation.

The CML system can prove its usefulness especially in countries where there are almost no rain radars, but where there are large cities with radio connections. WUR and KNMI are going to Nigeria and Sri Lanka to find out whether the system can be used there for short-term weather forecasts.