[lacker]

Yeah, the weather forecasts for air quality have been quite inaccurate. Since Friday, the projections for Oakland CA have constantly been, “it’ll clear up in the next 24 hours.” Only now, Tuesday morning, has the AQI dropped below 100.

It seems like it’s just a harder problem than regular weather forecasting. Nobody has much experience with giant smoke clouds like this one. So I can forgive the weather service some inaccuracy. Just don’t rely on the AQI forecasts for much.

[wnissen]

Yep. Predictions are obviously hard, but there's no wind to blow the smoke away, so I don't know why it is routinely predicted to do so 48 hours in the future. There is enough smoke piled up over the ocean (at least to judge from the satellite pictures) that it would take several days of steady winds to clear everything out. Any wind strong enough to theoretically clear the smoke quickly would probably spark new fires instead. And yet even with multiple fires nowhere near containment, the Spare The Air forecast still says "moderate" air in two days. I'll believe it when I breathe it.

The "incident meteorologist" for the North Complex fire was quoted yesterday as saying there was no system predicted that would cause a notable improvement in air quality for the next two weeks. Again, assuming no new fires, which is a stretch given that we're right in the middle of the traditional fire season.

The only forecast I rely on is the National Weather Service. Their site is painfully dated but the accuracy is unmatched.

[GaryNumanVevo]

I've found the HRRR experimental smoke models to be quite good. It's a bit cumbersome to use, but good enough for 18H forecasts

https://rapidrefresh.noaa.gov/hrrr/HRRRsmoke/

[cmason]

In my experience these have been completely wrong recently for my neck of the woods (PDX) even at 6-12 hours. Cliff explains in one of his posts that this is because of not modeling the inversion. I’m having better luck with the Copernicus CAMS forecast in Windy