Long-range weather forecasts have and continue to improve
The accuracy of weather forecasting has come a long way and new data sheds light on just how much forecasts have improved.
Author Hannah Ritchie presents data in ‘Our World in Data’ from the European Centre for Medium-range Weather Forecasts; analyzing what was forecast versus the subsequent weather on a large scale at 500 mb hpa geopotential height — which is a common metric to measure air pressure and analyze weather patterns.
The data shows threeday forecasts have been incredibly accurate since the 1980s. Five-day forecasts are also highly accurate and seven-day forecasts fall just short of being 80 per cent accurate.
Ten-day forecasts are not as on point but are getting better.
Ritchie’s data also shows that hurricane forecasting has improved with a decline in forecast errors in the Atlantic.
Overall, forecast skill is dependent on the type of weather, but this data gives a fresh perspective that weather forecasting remains incredibly accurate.
We can thank satellites, better supercomputers, improved data, and data assimilation for more accurate forecasts.
Of course, meteorology is a science, and large errors can occur in the short-term — it won’t always be 100 per cent accurate. Let’s also remember that while storm identification can happen five to seven days out, it’s often one to three days before details become clear.
Something I’ve heard more often is that the weather is less accurate now than it was 30 years ago. I think we’ve become so accustomed to accurate weather that when it is wrong, we notice it more.
“We’re so far removed from the days of very poor short term weather forecasts that it’s easy to forget just how worse it used to be,” scientist Tomer Burg said on X, formerly Twitter.
I must agree.