Mobile Meteorology

June 17, 2013 — The idea to measure rainfall with cell phone signals arose eight years ago over a cup of coffee. Hagit Messer-Yaron, professor of electrical engineering and former chief scientist in Israel’s Ministry of Science, was meeting with a meteorologist friend in a university cafeteria. The friend was struggling to find high-resolution weather data for his climate models, used in the short-term forecasting of events like flash floods, as well as in the description of long-term global changes. Fine-grained weather information is unavailable across much of the globe, and where it is available it’s often measured inconsistently — an essential problem for accurate modeling and prediction. Messer-Yaron had an idea: why not look to the effects of weather on cellular signals, now ubiquitous across the world?

For decades, Messer-Yaron had studied signal processing and cellular communication. She knew rainfall interfered with wave patterns in measurable ways, and that information might be gleaned from this interference. But could the interference be used to actually measure the rainfall creating it? Through an existing connection with cellular companies in Israel, Messer-Yaron and a doctoral student gathered months of signal data. When compared to daily weather records, individual precipitation events affected the signals in notably different ways. The team designed an algorithm to match signal noise against particular rain events and ran a pilot test in 2005. “It worked like a miracle,” Messer-Yaron says.

“The potential is vast in the countries of Africa, where water is such an issue.” — Harald KunstmannWith a few independent science teams from Europe, Messer-Yaron has pushed far beyond the pilot phase. She hopes meteorologists worldwide will soon be able to gather precipitation data using the microwave signals passing between cell phone towers.

“This is one of the most sensitive ways to measure precipitation,” notes Harald Kunstmann, a professor of hydrology at the Karlsruhe Institute of Technology in Germany.

Clear Picture

Measuring rainfall through signal interference hinges on properties of wave attenuation first described in a 1930 theoretical paper, “The Effect of Rain and Fog on the Propagation of Very Short Radio Waves.” Particulates in the air — precipitation most dramatically — alter otherwise smooth wave patterns. Wireless communication companies have long known about and countered these effects to maintain a clear signal, but with understandably little interest in the possible applications of the information contained in the signal disruptions. By designing and calibrating an algorithm to separate this disturbance from a baseline clear-weather signal, Messer-Yaron and others have been able to extract a detailed fingerprint of rain.

The readings provide a complement to, not replacement for, the two methods now commonly used for measuring precipitation: satellite radar and ground-based rain gauges.

Radar images provide a valuable big picture, but have trouble giving precise readings of what’s happening on the ground. Rain gauges are precise, but provide only a single point of information. Microwaves, however, are relatively close to the ground, traveling at heights of 100–200 feet, and traverse a linear path. “It happens to be that the weakness of the radar and the weakness of the gauge is the strength of the microwave link,” says Kunstmann. Together, these three measures capture very clear spatial and temporal pictures of rainfall — particularly important in the case of severe weather events where conditions change quickly.

A Signal Possibility

Cell phone rain measurement technology is emerging at a time when water-monitoring networks are underfunded and in some cases being reduced. “Less is known with each passing decade,” according to a 2009 United Nations World Water Report. Meanwhile, the World Economic Forum this year ranked the water supply crisis as one of the top five global threats. These concerns hold particular weight in developing countries, which often struggle to maintain meteorological infrastructure but are deeply dependent on the vagaries of weather.

“The potential is vast in the countries of Africa, where water is such an issue,” says Kunstmann. The bulk of cellular subscriptions, nearing 7 billion this year, are in the developing world, with 500 million across Africa alone.

Key remaining challenges are integrating data from the three different sources to paint a single picture of precipitation and, perhaps equally challenging, negotiating deals with cellular service companies.

“Israel is a small place, with three major cell providers,” says Messer-Yaron. “All three of them gave the data for free — no contracts, no expectations. This is a nice model for collaboration between industry and universities, but I think my colleagues in other countries have not been so lucky.”

Kunstmann’s lab, for instance, spent three years working to build trust with a point person at a major cellular provider. Only after that did they gain access to the data.

Despite this potential hurdle, the method has great promise. Teams continue to refine the algorithm’s accuracy and universal applicability. “The best solution is not yet found,” says Christian Chwala, one of Kunstmann’s doctoral students. Looking ahead, these groups are hoping not only to accurately measure rain, but also to distinguish different types of precipitation, such as sleet and snow. Even more ambitiously, they hope to one day measure any kind of airborne impurity through microwave attenuation.

“Once you start to think about reverse engineering the signal, the opportunities are endless,” says Messer-Yaron. Though precipitation has the most dramatic effects on cell signals, she says, future research and development “absolutely” could open up a full spectrum of global-scale phenomena, such as pollution, to this type of monitoring. Perhaps with the help of a little more coffee.