Dutch startup Whiffle has developed a computing model that allows for hyperlocal weather forecasting. Even nearby trees, wind turbines, and buildings can be included in the model. Especially those data centers situated in areas prone to flooding or severe storms may find this beneficial.

Climate stress tests

As a result of climate change, weather forecasts are becoming increasingly unreliable. A country's digital infrastructure must be better prepared for disruptive weather conditions. In Western Europe and other parts of the world, there is a discussion about having data centers and other parts of a country's digital infrastructure perform climate stress tests. It is common practice for financial institutions to perform stress tests to determine the impact of various scenarios on their business. The same idea could be helpful when determining whether a country's digital infrastructure can withstand severe weather.

The technology was developed at the Delft University of Technology. Whiffle, a spin-off from the university, is bringing the technology to market. Earlier this month, the startup received an investment of 3 million euros, which it intends to use to grow its team and expand internationally.

'Weather finecasting'

Based on graphical processing units (GPUs), Whiffle's weather model - the company uses the term ‘weather finecasting’ - can capture local turbulence and underlying processes and conditions in the atmosphere in greater detail. It is the first operational weather model based on Large Eddy Simulation (LES), which is optimized by AI and machine learning.

Photo credit: Chutipon Pattanatitinon