Startup Uses Butterfly Effect To Predict Weather Disasters

AI saves the lives of soldiers, prevents crop damage, and protects chemical plants

Artificial intelligence is being used to predict extreme weather events like hurricanes and heat waves, three months before they happen.

It is saving the lives of soldiers in the field, preventing damage to farmers’ crops and protecting chemical industrial plants around the world. 

Technology developed by Emnotion, an Israeli startup, is inspired by the “butterfly effect”.

American mathematician and meteorologist Edward Lorenz coined the term in 1972 to explain how something very small – a distant butterfly flapping its wings – could have a disproportionate impact on something very big – like a devastating tornado several weeks later.

It sounds almost too far-fetched to be true, but recent research suggests that 72 per cent of hurricanes and other tropical storms in the Atlantic are related to low air pressure in Africa, nearly 4,000 miles away.

Whether or not we understand it, everything on our planet is interconnected – including weather and climate.  

Emnotion takes an unimaginable volume of global meteorological data, processes the most important information, and applies artificial intelligence to assess how specific conditions in one microclimate could ultimately cause a weather disaster in another. It specializes in providing “unique and rare forecasts” and hyper-local alerts.

“You need to know how to actually calculate the butterfly effect to understand how climate conditions impact each other,” says Ilya Shapira, CEO and Co-founder of Emnotion.

His company analyzes tiny, highly-localized variations in temperature, wind, humidity and more to forecast extreme weather events up to three months before they happen. 

“There is a higher frequency of out-of-season events due to climate change, including anomalies like hurricanes, typhoons, and so on,” says Shapira.

He says existing climate models are based on historical data and couldn’t predicting extreme weather events. So he set about creating something that could.

“What we have is an absolutely new approach to climate monitoring and forecasting,” he says. It doesn’t use any measuring equipment. Instead it gathers climate data from third parties and open sources and uses Big Data analysis, machine learning and other technologies.

Many cities around the world – including Tel Aviv, close to where Emnotion is based – don’t have meteorological stations, but the company is filling in the gaps, and has already mapped close to 90 per cent of the world.

Extreme weather forecasting is a must for a wide range of clients in logistics, manufacturing and agriculture, among others, who are prepared to pay for Emnotion’s updates. 

The company worked with small and medium-scale farmers in Asia, as part of four-year long collaboration with the UN Development Programme.

“In many cases, small farmers cannot afford high-quality forecasts, but the risks are higher for them than they are for the biggest farmers, because it’s all they have – one dunam, or a half hectare of land,” says Shapira.

“Each event can cause them to go bankrupt. So when they have the ability to understand and better be prepared, it saves them a lot of money.

“Farmers have three to five different apps on their phone with weather forecasts for like 10 days ahead,” he says. “But the problem is that it doesn’t give them the solution for being prepared, because statistically, after nine days, a general weather forecast has a 50 per cent chance of being incorrect.”

Emnotion also works with ICL (Israel Chemicals Ltd), which manufactures chemical fertilizers, metals, and other chemical products for the agriculture, food and engineered materials markets. 

“We help them monitor and forecast the event and its impact on safety issues on their plants around the globe. It’s important to minimize the costs of the events, and be better prepared for the next one.”

Using extreme weather forecasting can help avoid industrial accidents, such as the Arkema plant explosion of 2017, in which flooding from Hurricane Harvey disabled the refrigeration system at a chemical plant in Texas, USA, causing it to explode.

Two wastewater tanks also overflowed during the storm, releasing more than 23,000 pounds of contaminants like heavy metals into nearby homes and soil in neighborhoods near the plant. 

Aside from farmers and industries, Emnotion provides the IDF with a six-month forecast that updates hourly. It not only saves money, but potentially saves the lives of soldiers in the field from heat waves and flash flooding during operations and field training.

“We help them to better plan their training, help them better plan their training and save millions of shekels on not last minute canceling of logistics,” says Shapira.

But climate forecasting model can’t predict every extreme weather event. To be able to determine when tsunamis, earthquakes, and volcanic eruptions will take place, more research is required. 

Emnotion’s main competition is from governments, producing their own climate models using supercomputers and thousands of scientists.

“But from what we see, governments understand that they need the help of private companies, like more flexible thinking,” he says. 

Since its foundation in 2016, Emnotion has relied entirely on funding from Shapira and other co-founders plus income from projects, but it has just launched its first-ever funding round.

Subscribe to NoCamels weekly newsletter and get our top stories