Hurricane forecasts (NHC) may be running headlong into the butterfly effect

[tolakram]

"When you look at the improvements in hurricane track forecasting, they're astounding," said study co-author Chris Landsea, who is a scientist at the National Hurricane Center. "They've dropped two-thirds in a generation. But we know we’re not going to get to zero errors."

https://arstechnica.com/science/2019/08/hurricane-forecasters-may-be-reaching-the-limits-of-predictability/

[wxman57]

I think that one of the main limiting factors, as briefly mentioned near the end of the article, is the relatively poor amount of atmospheric and oceanic data going into the model initialization. We've seen that track forecasting improves significantly once low and high-level reconnaissance data are availble. However, we really have very little high-resolution data across the oceans. Perhaps future generations of satellites and increasing use of drones to sample the tropical atmosphere will provide additional data to help improve forecasts.