Good morning all,
It’s exciting to see four week’s worth of evaluation scores on the AI Weather Quest website: Leaderboards • AI Weather Quest .
Something I’ve noticed is that several variable-averaged scores are dramatically reduced by poor precipitation predictions. I was expecting precipitation to be the hardest variable to forecast, but not to the extent of < -1 RPSSs. Is there any information I can provide to support participating teams?
Before the beginning of the next competitive phase (13th Nov, 5 weeks), I am very happy to take technical questions regarding evaluation.
Thanks,
Josh