August 2015 in the Pacific Northwest

[Front Ranger]

I can see why there would be some apprehension, but again, the models (paricularly the MM5's) were pretty consistent.

 

It's easy to criticize coaches after a call goes terribly wrong, and it's easy to criticize pro mets when their forecast is way off. You can explain why they were wrong until you're blue in the face, but if you do it after the fact, it's kinda weak.

 

There, I spelled it out.

 

[Deweydog]

Why do you keep citing your prediction from last Tuesday? Seems silly.

You seem stuck on this hindsight issue. That was a forecast based partly on what ended up playing out today.

 

I'm not a genius or a professional, but I do know our climate almost kinda okayishly.

[IbrChris]

I can see why there would be some apprehension, but again, the models (paricularly the MM5's) were pretty consistent.

Live and (hopefully) learn.

[IbrChris]

You seem stuck on this hindsight issue. That was a forecast based partly on what ended up playing out today.

 

I'm not a genius or a professional, but I do know our climate almost kinda okayishly

There was obviously enough disagreements among the models that the MM5 was probably an outlier. Outliers generally get dismissed. I think a look at real-time obs (sat, radar etc) would have borne out the prudence of dropping highs closer to 88-90. I wouldn't have gone lower...I can't fault NWS for not going 80 or even 85 today absent a major change in boundary layer airmass. Thick altocu killed our heating, end of story.

[Front Ranger]

You seem stuck on this hindsight issue. That was a forecast based partly on what ended up playing out today.

 

I'm not a genius or a professional, but I do know our climate almost kinda okayishly.

 

Except you've been citing days of mm5 runs leading up to today. Obviously, that wasn't the case last Tuesday. 

 

I think you're a little scattershot here.

[IbrChris]

In other news, Monday looks warmer than today even with more marine influence. It's an upside down world.

[Deweydog]

Live and (hopefully) learn.

 

It's a pretty common forecast anomaly, and really the consequences are so trivial if it goes wrong I can see why they just ride it out.  Lives aren't exactly going to be lost if it's 80 as opposed to 90.  We've already seen it play out once this summer during the last weekend of June.  July 2006 was another biggie when the cloud shield moved up from the south and brought the one ridiculously warm night.  Good times.  

[BLI snowman]

Or, as I surmised with a coworker this morning, NWS generally has a cool bias to start a heat event and a warm bias on the back side. Call it forecaster inertia...or that they get "burned" a couple times and just say "F#ck it let's add 2-3 degrees to what we would have forecast."

The same is true in the winter with arctic events. Too cold at first and too warm once the air is entrenched.

[Deweydog]

There was obviously enough disagreements among the models that the MM5 was probably an outlier. Outliers generally get dismissed. I think a look at real-time obs (sat, radar etc) would have borne out the prudence of dropping highs closer to 88-90. I wouldn't have gone lower...I can't fault NWS for not going 80 or even 85 today absent a major change in boundary layer airmass. Thick altocu killed our heating, end of story.

 

I could definitely see staying middle of the road.  I never paid a whole lot of attention with what highs they were going for today, but I imagine given the players on the field they were mainly in the mid 90's which would have basically required a full sun scenario.  

[IbrChris]

It's a pretty common forecast anomaly, and really the consequences are so trivial if it goes wrong I can see why they just ride out.  Lives aren't exactly going to be lost if it's 80 as opposed to 90.  We've already seen it play out once this summer during the last weekend of June.  July 2006 was another biggie when the cloud shield moved up from the south and brought the one ridiculously warm night.  Good times.  

More trivial than the bust on Thursday when GFS MOS was showing consistently a high of 102-103 at PDX and NWS was adamant about 97-98 until inside of 24 hrs. Course MOS won that round and the result was a heat advisory upgraded to an excessive heat warning. I'm not one to care about highlights but the general public probably senses an unpleasant difference between 97 and 103.

[Front Ranger]

There was obviously enough disagreements among the models that the MM5 was probably an outlier. Outliers generally get dismissed. I think a look at real-time obs (sat, radar etc) would have borne out the prudence of dropping highs closer to 88-90. I wouldn't have gone lower...I can't fault NWS for not going 80 or even 85 today absent a major change in boundary layer airmass. Thick altocu killed our heating, end of story.

 

Yup. And the reality is, sometimes those things are just about impossible to predict. There's always a chance it could happen, but again...no one expected highs around 80 in the Willamette Valley today. Surprises happen.

 

Part of the fun of following weather.

[IbrChris]

I could definitely see staying middle of the road.  I never paid a whole lot of attention with what highs they were going for today, but I imagine given the players on the field they were mainly in the mid 90's which would have basically required a full sun scenario.  

Ah...going back and looking at T-24 output for today's high I see the SREF had 82 for PDX and NAM MOS showed 87. Guess those shouldn't have been written off.

[Deweydog]

Except you've been citing days of mm5 runs leading up to today. Obviously, that wasn't the case last Tuesday. 

 

I think you're a little scattershot here.

 

Pattern recognition is a factor.  One may be inclined to devalue a model solution, such as column-integrated cloud water, if the pattern doesn't seem ripe for it.  This one did and it was apparent in the models even early last week.  Doesn't mean you'd go hog wild about a full-on overcast situation but you might take a stab at a less extreme number.  

[Deweydog]

More trivial than the bust on Fri when GFS MOS was showing consistently a high of 102-103 at PDX and NWS was adamant about 97-98 until inside of 24 hrs. Course MOS won that round and the result was a heat advisory upgraded to an excessive heat warning. I'm not one to care about highlights but the general public probably senses an unpleasant difference between 97 and 103.

 

I don't think most people around here can tell much of a difference.  Hot is hot.  103 just sounds so much more sensational.  

[Deweydog]

Ah...going back and looking at T-24 output for today's high I see the SREF had 82 for PDX and NAM MOS showed 87. Guess those shouldn't have been written off.

 

I'm sure they were still shocked and awed by the 80.  

[Jesse]

You mean algorithm? It's ok, that's kinda insider lingo.

 

Yeah some vendors are gonna be **** some of the time, not all vendors are gonna be **** at the same time and no one vendor is likely to be **** all the time.

 

Example:

NWS ingests raw data as a base value in arriving at their forecasts. It's then adjusted (bias-corrected for example) and a MOS is produced for GFS, NAM and ECMWF though we are not able to access the Euro MOS. MOS is simply an algorithm that takes into account known biases of the models and adjusts for climatology (especially beyond day 5). It's a set of empirical equations. Together this forms an algorithm and serves as an adjusted final forecast.

 

It's not a stretch to consider a vendor as a similar entity that takes raw data and performs a series of modifications to that (their algorithm) to arrive at a result they claim is better than the raw numbers. If the number of vendors/members is sufficiently large the flaws in the algorithms for each (assuming they are different) will weigh less in the consensus. For example I include only one TWC forecast, even though they own both the Weather Channel and Wunderground and the forecasts on those two sites while very similar aren't identical.

 

As far as the robustness of the Consensus forecast, the more members the better generally. If I polled the forum members every morning for a 7 day forecast I bet the consensus of the members would generally be superior to the forecast of any one member. Again, not rocket science here.

Great post. You are quite the charmer!

[Front Ranger]

Pattern recognition is a factor.  One may be inclined to devalue a model solution, such as column-integrated cloud water, if the pattern doesn't seem ripe for it.  This one did and it was apparent in the models even early last week.  Doesn't mean you'd go hog wild about a full-on overcast situation but you might take a stab at a less extreme number.  

 

OLM hit 88 today, because they managed more afternoon sunshine. In the exact same pattern as Portland. But had those clouds extended north just a bit faster, maybe OLM fails to hit 80. No way to really know that in advance, despite pattern recognition, models, or arthritic knees.

 

Which is why Andrew's original statement that today was going to end up a lot cooler than expected for the Willamette Valley was correct. 

[IbrChris]

I'm sure they were still shocked and awed by the 80. 

I've heard most TV mets are alcoholics, perhaps most mets in general? This might be a contributing factor.

[IbrChris]

Great post. You are quite the charmer!

I aim to please. A happy Jesse is definitely more fun to be around. What was your high in Stevenson today?

[Deweydog]

I've heard most TV mets are alcoholics, perhaps most mets in general? This might be a contributing factor.

 

Jim Bosley always struck me as a drunk.  

[Deweydog]

OLM hit 88 today, because they managed more afternoon sunshine. In the exact same pattern as Portland. But had those clouds extended north just a bit faster, maybe OLM fails to hit 80. No way to really know that in advance, despite pattern recognition, models, or arthritic knees.

 

Which is why Andrew's original statement that today was going to end up a lot cooler than expected for the Willamette Valley was correct. 

 

The models didn't show nearly the consistent cloud shield over Washington.  Olympia is in Washington.  The models done good.  

[Front Ranger]

The models didn't show nearly the consistent cloud shield over Washington.  Olympia is in Washington.  The models done good.  

 

Portland is almost in WA. That was close! 

[IbrChris]

I don't think most people around here can tell much of a difference.  Hot is hot.  103 just sounds so much more sensational.  

We should avoid tryggers, right Jesse?

[bainbridgekid]

Looks like SEA did indeed fall short of 90 by a few degrees.

BFI still hit 90 though!

 

http://w1.weather.gov/obhistory/KBFI.html

[Deweydog]

Portland is almost in WA. That was close! 

 

The models are just that good!

[IbrChris]

Jim Bosley always struck me as a drunk.  

Dat hate mail....I don't know how Mark Nelsen can handle it. I guess just compartmentalize like a boss.

[Deweydog]

***FYI TO THE FORUM***

 

It will probably be cloudy in the morning.  

 

Temperatures will respond accordingly.  

[IbrChris]

4 pages on Aug 2nd...at this rate we will hit exactly 62 pages for the month*

*Disclaimer: This is a mathematical derivation not a personal forecast.

[Deweydog]

Anyway, gonna be a warm night.  

[Jesse]

Anyway, gonna be a warm night.

PDX is running -9 compared to yesterday!!

[Deweydog]

PDX is running -9 compared to yesterday!!

That'll make the warm lows that much more surprising!

[Jesse]

That'll make the warm lows that much more surprising!

Not to anyone who looks at an MM5 model but fails to mention that fact until tomorrow morning!

[Deweydog]

Not to anyone who looks at an MM5 model but fails to mention that fact until tomorrow morning!

Too bad they aren't available to the public multiple times daily.

[Jesse]

Too bad they aren't available to the public multiple times daily.

I know.

 

Let's get IbrChris on it. He seems to be brilliant and well-connected.

[MossMan]

Warmest late evening since this toasty spell started, 69 degrees currently.

[Front Ranger]

Too bad they aren't available to the public multiple times daily.

 

[Front Ranger]

BFI still hit 90 though!

 

http://w1.weather.gov/obhistory/KBFI.html

 

BFD.

[ShawniganLake]

0z gfs indicates that the Portland 90+ record will be long gone by mid August.

[Jesse]

Who thinks that this Friday will overperform and hit 90 at PDX to make up for today's underperformance?

[Front Ranger]

Who thinks that this Friday will overperform and hit 90 at PDX to make up for today's underperformance?

 

The other shoe has to drop, right?