What level of prediction granularity can we achieve?

[taylorfturner]

What level of prediction granularity can we generate from this model? If I have a lat/long point or "lasso" of lat/long for a defined area, can I get prediction of weather data points for that lat/long area? Ideally down to a 1,000 yds ^ 2?

[taylorfturner]

May have answered my own question: https://huggingface.co/ibm-granite/granite-geospatial-wxc-downscaling. Appears that IBM's downscaled model predicts a 12.5 km^2.

I imagine this is a "single" point estimate for this entire square and not further granularity within that 12.5km square area?

[WillTrojak]

At the moment, that's what we have; however, we (IBM Research) are actively working on this.

[taylorfturner]

At the moment, that's what we have; however, we (IBM Research) are actively working on this.

Thanks, @WillTrojak - any idea on timing for delivery of this? Way to publicly track progress?