I was setting up a scheduled Auto-Run forecast in WXSIM, and as I clicked on the " Mix with Bias-corrected external model data via WXSIM-Lite " button, I of course, noticed some notes explaining the internal procedure that takes place
when using bias corrected model data, and choosing the percentage influence
from that bias-corrected model data.
It was then when I noticed a statement that said " Studies done with GFS, being the only Global external model data being imported into WXSIM yielded optimal results with a weighting around 60%-65%.
the addition of mixing European (ECMWF) model data, specifically with an ECMWF:GFS ratio of 70:30 causes the optimal mix of WXSIM-Lite data to be slightly lower, around 55% - 60%"
While very straight forward, it caused me to be curious about just how these statistics were obtained?
Secondly, not Knowing how old these notes are, the question arises if any new data on what the ultimate mix for the most accurate forecast would be?
If anyone has figured out what seems to be working for them, I welcome your suggestions.
Thanks!