Models are a total borefest.

It has been raining here in Orange for a good part of the day, heavy at times, and there are still off and on showers occurring this evening as well, with one quick downpour about 10-15 minutes ago.

Yup. Herpes and Hasselholf FTW

Model performance, breakdown 12/24/25 8:58 pm SYNOPSIS: I want to discuss the model performance or lack thereof and breakdown what they were forecasting, what happened, and why. As of the 18z model data yesterday afternoon it was appearing that the odds had increased to ~50% for gusts 50mph or greater for PDX metro and the Willamette Valley. This was weighed heavily by the 18z ECMWF deterministic and EPS which trended towards the windier solutions along with hi-res guidance from the HRRR/RAP and GEFS/SREF ensembles. There was still the possibility for a significant, damaging wind storm although on the lowest end of probabilities. There was as well uncertainties which lent to a ~35% chance that we wouldn't see anything more than 20-30mph breezes. The NAM, NAM-NEST/HRRR/RAP were very bullish painting a possible damaging wind storm with gusts 55-80mph, but this was the bottom percentile of what we thought was possible. IF we had seen a compact, deep low around 977-983mb which is what the hi-res had been showing then yeah, a more impactful wind storm would have occurred. Heading into Tuesday evening we were preparing for strong winds and power outages. BUT.... That all changed when the 00z runs came in. The 00z ECMWF backed away from the potential for a wind storm and EPS dramatically reduced the number of its 50 member ensemble showing a wind storm. The subsequent models followed and hi-res guidance also slowly but surely backed off now forecasting a weaker low, but failed to handle the double barrel structure. How did we go from being fairly confident we'd see strong winds to the next model cycle backing off almost entirely on this? How did the models do so poorly even within 24 hours? The Pacific Ocean. There simply aren't enough ship reports or other instrumentation across the North Pacific to sample the surface and atmosphere which makes forecasting in the Pacific Northwest the most difficult in the country. By 00z though there was much better initialization with that data being ingested into the models. You could see the double barrel structure on Water Vapor as well. Some models did better than others, and some were abysmal even within 12 hours! Model grading (Inside 18-24 hours *before 00z runs Tuesday night) [B+] HRDPS(Canadian - High Resolution Deterministic Prediction Service), RDPS(Canadian - Regional Deterministic Prediction System, ICON(German - Icosahedral Nonhydrostatic) Often jokingly referred to as the "Hasselhoff" model. - SYNOPSIS: These models excelled with good run-to-run consistency forecasting a weaker low to track inland and not north along the Oregon coast. They did not however accurately forecast a double barreled structure. Perhaps we should have not discounted the Canadian? Hmmm.... [B-] ECMWF deterministic, ECMWF EPS 50 member ensemble, GEM deterministic. - SYNOPSIS: The ECMWF and EPS consistently had forecast a weaker low with only minor wavering primarily the 18z which trended windier. The ECMWF did not however forecast a double barrel structure. EPS was never bullish on a wind storm. The GEM was also not on board. [C] GFS deterministic - SYNOPSIS: Decent good run-to-run consistency with a 987-990mb tracking north along the Oregon coast. Did not however forecast a double barrel low. [C-] UW-WRF-GFS - SYNOPSIS: Too aggressive on the depth of the low and compact structure. The Wind Gust model however was about on par with what we thought we were going to see with gusts 45-55mph. [D+] SREF, GEFS - SYNOPSIS: Both of these ensembles failed with a more compact low moving north along the Oregon coast. [D-] HRRR/RAP, NAM/NAM-NEST/3k(North American Mesoscale-Nested) - SYNOPSIS: Models struggled mightily forecasting a stronger, compact low moving north along the Oregon coast. By far overperformed the depth of the low and forecast peak wind gusts. Just as with the others did not depict a double barrel structure. A colossal failure. What can we learn from this? When we are within Day 3-5 days(72-120 hours) of a forecast event to never go 'all in' on any one model. Don't be fooled by a few deterministic runs. Utilize ensembles and look for trends, then build run-to-run consistency and model agreement. It is best to rely on and heavily weigh the EPS, and GEFS or a blend with the CMCE/SREF in the mix. Hi-res guidance and mesoscale models sometimes aren't what they should be. This time they performed terribly mainly depicting a much stronger low than other models. BUT honestly there were just so many hi-res models to ignore showing their agreement for a major wind storm. We should look for specific features to clue us in on certain things. If models show a baggy, elongated low we can often but not always rule out strong winds as the pressure gradient is too relaxed. There are times though where the southern lobe swings north with its tighter gradient couplet leading to wind issues. These systems often have a double barreled structure with a northern low and secondary southern lobe/low. That was the case today. Cut the NWS and local TV mets some slack. They handled things just fine and given the available model data going with 40-55mph gust was the right call. They never screamed "WIND STORM!" and explained why there was a degree of confidence to expect strong winds, but also why there were still uncertainties remaining. You want to inform the public, but you also don't want to freak them out at the same time. This was an extremely challenging forecast due to a total lack of model agreement even 12-24 hours out. I do not envy their jobs. Wind Storms are difficult to forecast and often models change wildly. There are still some things human forecasters can use that computer models don't have such as gut instincts based on knowledge from past storms understanding what to look for and why. The superior ECMWF and EPS can rarely be discounted even if there are a plethora of other models in disagreement. I guess there's a reason why the ECMWF is widely known as "King EURO" right? Until the next event!

Yeah I spend a few days a month almost every month right on Moses Lake for about 15 years…Wind was a way of life over there and it’s not the same as hearing it roar through the trees here.