Most meteorologists predict the weather by looking at the sky, but one startup thinks it can do better by looking at the ground.

A startup called Understory thinks it can provide better forecasts by measuring conditions on the ground, instead of the traditional detection that emphasizes atmospheric conditions. That is, after all, the location where it matters for most life and property risks.

“The primary sources [of weather information for enterprises] use a lot of radar and satellite measurement,” Understory founder/CEO Alex Kubicek told VentureBeat. “They do use a little ground-level observation,” he noted, “but within 40 kilometers of the middle of Boston there might be three [on the ground] stations.”

“We would put 50 stations in the same area,” Kubicek said.

AI Weekly

The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.

Included with VentureBeat Insider and VentureBeat VIP memberships.

Understory is announcing today it has received $1.9 million in a seed round to implement the idea in the real world.

The Boston-based company, founded in 2012 and until recently known as Subsidence, adopted the name Understory from the term that describes the area in a rainforest underneath the forest canopy.

Its approach employs “ground-truth” detection, the company’s description of the increased resolution and fidelity that it believes is available from detecting rain, hail, wind and other weather at ground level.

Each Understory station, about a foot wide by two feet tall, is part of a data grid that the company networks together via cellular connections. Sensors inside each station sample wind, rain and hail in three dimensions. That way, for instance, it can record the size and the angle of hailstones, not just the fact that it’s hailing.

There are also sensors for heat, pressure, humidity, lightning strikes, and light. The data, integrated with historical and forecasted weather events, is used to generate datasets, views and actionable information in real-time.

At some point, Kubicek told us, the service may be available to individuals, but the first focus in on enterprises, particularly insurance, agriculture, and broadcasting companies, which all have a strong economic interest in knowing what the weather will be. The enterprise client will have access to a dashboard and an API. The company’s first client is American Family Insurance in Kansas City, Missouri, where serious weather events are not uncommon.

A ground-based network like Understory’s, Kubicek said, is most useful “for any event that will cause the weather to change dramatically,” such as hail, a thunderstorm, or a tornado.

“When you have a blue sky, that means the weather is stable,” and there is less need to know exactly and immediately where the weather is hitting and where it is traveling.

The company had previously landed only $68,000 from the Bolt hardware accelerator where it was hatched, and from another accelerator. Over the next 18 months, Kubicek said, the new funding will be used to “build up the software side, and to create a short-term forecasting model.”

The seed round was led by True Ventures, with participation from RRE Ventures, Vegas Tech Fund, SK Ventures and Andrew C. Payne.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More