With all the recent crazy (and sadly), deadly weather events - from deadly tornadoes to a 50 degree drop in temp (from 105 F to mid-50s F) and 2 feet of hail in 1 hour in a Texas town in May - I thought it appropriate to take a quick look at how AI and ML are impacting meteorology. How does AI & ML compare to traditional methods and what challenges lay ahead?
Very simply put, traditional methods rely heavily on physics equations. Data, such as wind speed, pressure, precipitation, and temperature, is gathered from satellites and weather stations. Supercomputers can take that data and process it to make a forecast.
The Scientific American wrote, regarding AI forecasting, “AI tools are statistical models: they recognize patterns in training data sets composed of decades of observational weather records and information gleaned from physical forecasting. Thus these models may notice that the weather setup of a certain day resembles similar events in the past and make a forecast based on that pattern.”
Now with GraphCast, Google DeepMind’s AI model for global weather forecasting, predicting weather conditions “up to 10 days in advance more accurately and much faster than the industry gold-standard weather simulation system – the High Resolution Forecast (HRES),” according to Remi Lam. It can also offer earlier warnings of extreme weather events.
However AI is not without its own limitations, at the moment. For instance, the Scientific American reported that because AI models rely on past data they may be poorly equipped to forecast rare events, such as Hurricane Harvey or the rapid intensification we saw with Hurricane Otis. GraphCast also seems less able to forecast storm and rainfall intensity.
So, will AI and ML be replacing traditional forecasting methods? That does not seem to be the consensus at all. Rather collaboration seems to be the name of the game with many experts feeling AI can be another useful tool in their toolkit. This combo will undoubtedly be key to developing more accurate, life-saving weather warning systems. On top of that it will be nice to leave the house in the morning knowing for sure whether I really need to lug that umbrella around all day or not - not that I will remember it anyway but that’s another issue.
For more information check out these awesome resources/articles:
* Cover photo credits: Photo by NOAA on Unsplash