.webp)
AI, it has been a topic that has engulfed the business world. In fact, it's hard to find a company that is not at least considering how they can implement it into their operations. The business of weather is no different - both in the private and public sector.
Computer models are a huge focal point in weather forecasting, by simulating the atmosphere in the future and projecting what the weather will be days and weeks down the line. While they have improved tremendously in time, they still have their flaws and deficiencies . This is where AI comes in.
AI behaves like a super human brain in the sense that it can learn from its mistakes and ingest information & quickly process it. It can also store past weather data and use that to make decisions or comparisons about future weather.
“of the top issues in the world today, AI and climate change and extreme weather have got to be maybe 2 of the top 5, so it’s really natural to start thinking about how we can use the most advanced technology in the world to help address one of the biggest problems in the world.” Alex Levy, CEO Atmo
While the current computer models, which simulate future weather based on current observations, have improved over time, the process to get there has been timely and costly. "The old way of doing things didn’t have a learning process associated with it, so let’s say errors or biases crept in there....that meant you missed a major event “ says Alex. He continues by saying, "one of the advantages of this new AI weather approach is that we can let 1000 flowers bloom, we can make 1000 forecasting models and customize them to the particular needs of particular people."
So essentially, it’s a forecasting mechanism with high-level intellectual capabilities...and the detail or resolution is impressive. Atmo specifically is able to get down to a street level view, city block by block at 5 to 10 minute increments, which is roughly 100 times more detailed than current forecasts available (watch the full interview above to see more).
Atmo has a product like this for the San Francisco region. And like the Bay Area, back here in the Capital Region there are multiple microclimates over a relatively small geographical area. As we have seen, the forecast could be much different in the Berkshires versus say the Hudson Valley. Specifically for Atmo, this level of detail is super useful for clients such as the military or those involved in space exploration.
Well, Alex believes there still needs to be meteorologists to analyze this vast quantity of data. In essence, it's more raw material for meteorologists to turn into the analysis that people want to hear. So it’s not all dooms day when it comes to AI. In fact, the confluence of AI and weather has the potential to pay huge dividends for both you and meteorologists like me.
