Tuesday, August 15, 2017

Eclipse Weather Forecast: Looks Good for BOTH Sides of the Cascades

I didn't want to make this forecast until there was some real skill to the forecast.  Today, 6 days from the big event, I believe I can provide actionable information, although I would not pretend there are no uncertainties.  

The bottom line:  it looks good over BOTH sides of the Cascades, except for the immediate coast.

Typically, there is little skill for weather forecasts greater than 10 days, marginal skill for 7-9 days, and rapidly increasing skill for weather predictions of less than a week.  

Furthermore, we have tools, and particularly ensemble forecasts (where we run our models many times), that can quantify the forecast uncertainties and tell us when we have periods of greater or lesser predictability.  

Back to the eclipse forecast.   As most of you know, the area of totality will run across northern and Oregon, roughly on an east-west track (see map).  Totality, which can last up to around 2 minutes, begins around 10:15 AM PDT Monday August 21st along the coast and 10:25 AM at Oregon's eastern border.  10:15 AM PDT is 17:15 UTC or universal time.  Keep that in mind for later.  Seattle will have a partial eclipse (about 92% coverage by the moon), with the darkest time around 10:20 AM PDT.

The big question for eclipse watchers in Washington and Oregon is whether they should be on the east or west side of the Cascades (or in the mountains!).  There are, in fact, three major threats for watching the eclipse:

  1. Low marine clouds along the coast and potentially the Willamette Valley
  2. Smoke from wildfires
  3. A major frontal system with deep clouds (like over the past weekend)

First, lets consider wildfire smoke.  Last week, there was dense smoke over the region, much of it from the big fires in British Columbia, with an assist from local fires on the eastern side of the Cascades.  With a shift to a very different large scale weather regime, the BC smoke is now heading more to the east, leaving Oregon and Washington in the clear.  Currently (Tuesday AM) air monitors show little smoke at ground level over most of the area.

And the 48 hour forecast of the Canadian smoke model (FireWork) show little smoke and the BC smoke is heading away from us.  I expect that to continue for the rest of the week based on current model forecasts.

There are some local fires burning in our region (see current fires below), but most are small or contained, and the amount of regional smoke is relatively small away from the fires (although there is a thin veil from them).  So away from their immediate vicinity, one should not expect a major impact.

 The weather should be typical this week, with no lightning over the central and northern Oregon Cascades...so one should not expect new wildfire initiation.    Furthermore,  the eclipse is close enough to solar noon that the sun will be relatively high in the sky, and thus the sun's light will not be going through a large amount of atmosphere, as occurs near sunrise and sunset.

Bottom line:  smoke is not going to be a major issue for this eclipse unless you are immediately downwind of a local fire.

But what about clouds? 

 Here is the latest UW WRF forecast for cloud water valid at 11 AM Monday morning.  We see a lot of low marine clouds over the Pacific, but they don't extend past the coastal mountains.  Salem and the Willamette Valley would be in the clear, as would those in Washington State and eastern Oregon.  Praise to the weather gods.

What about the European Center global model, the world's best?

Here is its forecast for 11 AM for total clouds and low-middle-high clouds.   No low and middle level clouds over the eclipse zone...which is very important.  There are some high clouds over Washington and NW Oregon, but these would be thin cirrus.  An irritant, but it probably would not ruin the show. And keep in mind that the position of these high clouds is very uncertain.

As I have discussed a hundred times in this blog, state-of-the-art forecasting does not look at one forecast, but rather uses ensembles of many forecasts to judge uncertainties and to provide probabilistic predictions.   So let us look at ensembles for this event.

First, take a look at the large, 51-member European Center ensemble prediction for total clouds at Salem, Oregon. The 51 rows are from the 51 members of the ensemble, and time is in UTC (0000 UTC 21 August is 5 PM 20 August PDT).  The situation around 1800 UTC 21 August is a mixed bag, but most members have little or no clouds--roughly a 25% chance of (high cirrus) clouds.

What about Redmond, Oregon on the other side of the Cascades?  Slightly better (20% chance of some high clouds).  But there is the threat of a thin veil of smoke there.

Along the coast at Newport, Oregon?   Around 40% chance of clouds.

Checking out an independent, large ensemble systems (the US-Canadian NAEFS) for Portland (closest available), suggests a high probability of very little clouds, with only a few members indicating 10-20% coverage (bottom row).

I could show you much more, but you get the message.   The best weather technology we have suggests a favorable situation for viewing the total eclipse in the Willamette Valley and in eastern Oregon. 

 Here in Seattle there may be some thin high clouds, but you should still enjoy the partial eclipse.
The place that is iffy is along the Oregon coast.  If you are there, a short trip into the coastal mountains of Oregon should do the trick (assuming the roads are not grid locked).

I believe that model solutions are relatively stable now--but by Friday we should be very confident in Monday's forecast.  Enjoy the eclipse and make sure you are careful not to look directly at the sun.


Tony said...

Like I ask my Dr.
"Which of the 2 pills would you chose if it would be you?",
I hope you will give us one more final opinion on say Saturday,
"if you would have reservations in both Salem and Madras,
where would you drive to early Sunday morning?"

Bryan Black said...

Looking good here in Monmouth in the Willamette Valley. Won't have to pay a dime to see it as I bought my glasses years ago. We are almost directly on the center of the totality path. One of the luckiest people on the planet.

Casey Connor said...

The forecast is much appreciated, thanks!

John K. said...

Will give it a go.. drive to Oregon Sunday night. But, figuring will probably have to turn back and scrub the mission at some point. An overpopulated world.

Matt said...

Thank you, Cliff. It's worth noting one more time that the difference between 99.9% coverage and totality is literally the difference between day and night. You've got to be within the path of totality to see the corona, and all the other goodies. GET TO THE PATH.

Vincent Kondo said...

Having been centerline on an annular eclipse in the past I'll have to second that Matt. Saw the May 1994 annular eclipse in Rochester NY..nice dimming but absolutely no possibility of looking at the solar disk directly much less observe any corona. Headed to Salem Saturday night. Joining the Mary's Peak enthusiasts while well off centerline looks like a stunning setting being on a mountain top.

Deek said...

Thanks Cliff. I look forward to your updates later this week. Re the high cirrus and thin smoke. If its not too thick this can actually be a bit of a bonus. I saw the 1979 eclipse from a location in Manitoba. On eclipse day there were high cirrus clouds but not so thick they obscured the event. This allowed us to see the onrushing shadow. Perhaps we would have seen the shadow anyway but I don't know because that was the only eclipse I have seen. The shadow was an awesome sight. It was an enormous thing that swept across the entire sky (roughly west to east) at an incredible speed (it seemed like only a second or two). It was otherworldly to see something so large, moving so fast in total silence.

JeffB said...

Deek, very cool description. I can't wait. It will be epic. Cliff thanks for the update. I saw '79 with cloud cover so all it did was get dark. This time going to ID so that there will be no clouds. Hoping to catch the shadow from a bluff.