AI & Early Warning Severe WX Alerts

 New Machine Learning Algorithms Spot Lightning Days Before Other WX Systems Do

                               Source:  Stock Lightning Images

New AI System to Forecast Lightning and More

There are few things more difficult to track and predict than the dynamic forces of nature that result in real-time weather.  Now, researchers at the University of Washington have invented an artificial intelligence, machine learning system that results in earlier and more precise weather forecasting.  The team started their artificial intelligence system with weather reports and data on lightning which is one of the most destructive and unpredictable forces of nature.  An example of the destructive force of unpredictable lightning is the massive 2020 California Lightning Complex fire that resulted in nearly 470,000 acres of total fire devastation and 23 deaths.  Early forecasting of that severe lightning event could have facilitated earlier fire preparation and possibly earlier containment.  

New AI Weather Tech

According to the UWASH team, this is the world's first AI system that deploys algorithms to predict lightning.  They say it demonstrates that machine learning works best on what humans don't fully understand. For scientists, one of the most confounding natural phenomena is lightning. The team combined weather forecasts and location points of lightning strikes from 2010 and 2016 and then extended the data to 2019.  The data they used is from the southeastern US.

AI Provides a 2 Day Leadtime on Lightning Forecasts

Their analysis has shown that machine learning using computer algorithms can improve lightning forecasts significantly over commonly used weather forecasting systems. In fact, their new AI system provides 2-day lead time over current weather forecasting systems. The algorithms provided better forecasts on their own without any direct programming from humans. 

What's the Importance of This?

It could help predict and prevent wildfires by earlier intervention. The AI system also is capable of creating more accurate weather forecasts and providing more accurate, long range climate change models. This is the world's first, AI machine learning system that predicts lightning strikes.   It shows the potential that machine learning AI methods can be used for other types of weather forecasting like tornadoes, hailstorms, the complex dynamics of blizzards and much more.