Forecast model comparison - which one is the most reliable?
-
We recently published a two-part deep-dive into how numerical weather prediction models actually work - from how the atmosphere is divided into a 3D grid and simulated step by step, to how global models differ from each other and why ensemble forecasting exists.
👉 Part 1: Understanding NWP Systems
👉 Part 2: Inside Today's Forecasting Systems
What you might not know is that meteoblue runs its own model families - NMM (operational since 2007, optimised for complex terrain) and NEMS (the newer generation since 2013, with improved cloud and precipitation forecasting) - alongside data from ECMWF, GFS, ICON, UKMO, and a dozen others. You can compare them all side by side in our MultiModel view.
A few questions for the community:
-
Which model do you follow most closely for your region, and has one ever clearly outperformed the others during a major event?
-
For those in complex terrain (Alps, Balkans, Scandinavia) - do you find high-resolution regional models noticeably better than global ones?
-
What's your take on AI models like ECMWF's AIFS entering the picture - genuine step forward or still unproven?
-
