This helps crowdsource real-time weather forecasts and could one day help collect climate data in areas without weather stations, its developers say.
Once installed, the app automatically collects data and periodically uploads them to a server.
The app’s ability to record air temperature is based upon the discovery that the temperature of a smartphone battery correlates closely to the surrounding air temperature, published in Geophysical Research Letters this month (13 August).
“Lithium ion batteries have temperature sensors to prevent damage caused by attempts to charge them when the battery is too hot,” the paper says.
But these sensors do not provide a direct air temperature measurement — due to heat being emitted by both the smartphone and its user. So the researchers used a model that estimates the outside temperature based on smartphone readings.
The fact that battery temperature correlates with ambient air temperature was discovered by accident, James Robinson, one of the authors of the paper and co-founder of OpenSignal, tells SciDev.Net.
“When data from many phones are joined together, they become even more powerful and will allow us to make weather predictions of unprecedented detail.”
The team was researching energy consumption in relation to poor mobile network signal, a condition that is known to reduce battery performance.
“We started playing with the data and decided to look at average battery temperature versus historic weather temperature, and we found a really strong correlation,” says Robinson.
“Many smartphones have a variety of sensors,” says Robinson. “When data from many phones are joined together, they become even more powerful and will allow us to make weather predictions of unprecedented detail.
Developing countries often invest fewer resources in collecting weather data.
“As smartphones become more popular in developing countries, WeatherSignal could provide a valuable source of weather data — either supplementing existing sources or as the only source for some places,” he says.
“We’re open to working with as many people as possible,” Robinson says. “For instance, we plan on making historic data available to academics and organisations such as the World Meteorological Organization.”
Enzo Campetella, an Argentina-based meteorologist and WeatherSignal user, tells SciDev.Net that although the app has potential, “there are still several stages to accomplish” before it is completely reliable for use in meteorology.
“In meteorology, it is essential that data are comparable, so it is essential that they are collected following the same rules or standards,” he says.
And, in countries where weather stations are scarce, “the possibility of comparing data is much lower”, he explains.
The WeatherSignal team admit that “many additional high-quality urban observations [are] needed to reﬁne the air temperature estimates from smartphones and to expand their possibilities”.
Having more data is also crucial, so they are also working to get as many people as possible using the app.