Norman WFO issues crap forecast

[Jonathan Whitehead]

Id like to begin this thread by saying that Im not taking shots at the Norman WFO, I know jack when it comes to forecasting.
However I noticed something that caught my attention.
Yesterday (Thr) I checked the Norman WFO website to get a glimpse of the wknd forecast. They didnt have any chance of rain whatsoever for the next two days. I go on there today and they have heavy rain with a slight risk of severe storms from storms that originated in Kansas overnight (MCS). Im more curious than taking shots at the forecasters over how they could or maybe the models coulda missed this just 24hrs earlier. It seems to me that the forecasters have been doing a really good job of picking up on these MCS's that have been coming at us in abundance this summer.

 

[Tony Laubach]

MCS's have been pretty good, but at the same time, the tracks of these can sometimes be off.. in some of the reports I heard from storms that slid out of Nebraska, some of them had some odd tracks to them.. an extact position of an MCS over hundreds of miles is hard to pinpoint (try forecasting a Hurricane's exact track). Those variences can lead to weather differences in many places.. although I haven't read the discussions or what lead up to this, I do know that an MCS is easier to forecast, but track variences and life-span can contribute to weather leading from that, and sometimes that's a bit harder to forecast.

As for the instance of last night, again, I don't really know, but a forecaster 24 plus hours in advance may not know the exact magnitude of what's coming, even if they said clear skies. Weather changes rapidly.

 

[Richard Halter]

A mesoscale convective complex is just that, mesoscale. As such, it is too small, and many times the dynamics surrounding its steering forces are too small for the resolution of models to pick up with a great degree of accuracy. As such, you can have the forecast for their trajectories and their developements to be off by a few hundred miles over the course of 24 hours.

Heck, that happens with hurricanes and mid-latitude synoptic systems as well. No reason to get so peeved.


rotest: :argue:

 

[David Draun]

Talk about the erratic tracks of MCS's That MCS from the plains last night split up into northern and southern componnents. The southern part backbuilt to the west as it moved south into KS/OK, the northern one has set it's sights on Chicago within an hour or two.

 

[Glen Romine]

Id like to begin this thread by saying that Im not taking shots at the Norman WFO, I know jack when it comes to forecasting.

Your title sure sounds like a stab to me - and I'm certain others will interpret it that way as well. You suggest that you are a met student at OU - so maybe you should start learning about forecasting instead of attacking those that do when a forecast goes awry. Predicting the weather is not like predicting the tides - the atmosphere is a much more complicated system. Models of the atmosphere are surprisingly good at offering a glimpse into the future - but cannot provide a perfect forecast - otherwise we would just need to run a single model (just one - don't need the dozens that are out there now 'cause they'd all give the same answer) once out to infinity and we'd never need forecasters again. But, 1) You can never know the true state of the atmosphere to begin with for initialization of the model, because observations are just a sampling of the true atmosphere, 2) even with a perfect model initial state, it is not possible to solve the equations of motion for the atmosphere without using simplifications on some level - and this adds in error that grows with time (a.k.a. nonlinear instability) - so forecasts from models will NEVER be perfect - though they are always getting better. Instead, models offer an array of "advice", sometimes good and sometimes bad, and the forecaster is left to consume this advice along with personal experience and guidance from current observational trends to try and guess what the future may hold. Sometimes this is easy to do - and other times it can be extremely difficult.

If you think the forecasts from the NWS aren't good enough, a) make your own, B) write to your congressman to increase funding for NOAA to buy better computers (to run more precise models), more observations platforms (like additional and more frequent RAOBs or more wind profiler sites), and to pay for more forecasters, or c) don't complain about it here and just accept that forecasts will inherently be bad sometimes and be happy when the forecasters get it right.

My 2 cents,
Glen

 

[Bill Tabor]

Yep, what Glen above me said is true.

Basically the short answer is that the forecast was based on models - perhaps almost entirely and the model busted. This is just the state of the art nowadays. If you watch the models much you will notice that day to day they change sometimes dramatically. When that happens it may flood instead of not raining, or you may have tornadoes instead of snow, etc. You basically have to get online and watch these yourself. I never look at precip forecasts anymore just go here for instance http://www.rap.ucar.edu/weather/model/ - select the model, time period, and 'precipitation' for the graph and take a look. Usually the RUC does a better job in my opinion for current day convective than the ETA. Probably the WRF does better. My old fav that I thought did a pretty good job was the ETA KF (Kain Fritsch) convective paramertization, but hard to find nowadays. Next you should look at the soundings, forecast soundings, convective temperatures, and what is going on for lifting / forcing mechanisms etc to allow convection to occur and rain to break out. At times IMO forecasting rain (convection) is the most difficult part of a good severe forecast. Take yesterday for example, SPC had a moderate risk to cover the area around NE, SD, and KS in case the cap weakened and precip broke out. It never did. THey downgraded to a Slight and there wasn't much rain and storms until after dark. It's really a probability game - that's why things are presented as a % chance. But like I say the complexity of mapping fluid dynamics and interactions in a global (not to mention mesoscale / microscale) environment is extremely complex. Personally I am amazed that brilliant people have been able to create such complex models that do a considerably decent job at projecting weather trends a large majority of the time. Hat's off to all you 'long haired' severe weather researchers and programmers.

 

[Sheila_Ward]

Agree with Bill and Glen. Your title indicates you are taking a stab. If you want a discussion about the forecast and why events happened the way they did, that's fine. Saying the Norman WFO issues a "crap" forecast is the same as the other "they blew it" threads. Which invites the response, if you can do better then posts your forecasts and let's compare.

I believe your intent is to discuss why this particular event was hard to forecast 24hrs in advance. You might want to change the title of this thread to reflect that...or you're gonna get slammed :wink:

 

[Greg Stumpf]

If the original poster didn't intend to label the OUN WFO forecast as crap, then I'd suggest that the moderators change the subject line to something else.


greg

 

[mikegeukes]

Weather changes, new data comes in, forecasts will change.

Read the Area Forecast Discussions over the last couple of days,
will help give you clues in what the forecaster is thinking.
http://kamala.cod.edu/offs/KOKC/0407081522.fxus64.html
http://kamala.cod.edu/offs/KOKC/0407081943.fxus64.html
http://kamala.cod.edu/offs/KOKC/0407090751.fxus64.html
http://kamala.cod.edu/offs/KOKC/0407091538.fxus64.html
http://kamala.cod.edu/offs/KOKC/0407091954.fxus64.html

You should sit down, make your own forecast on your own,
before reading the discussions and forecasts from the NWS,
Compare your forecast to the local NWS office, and see if
you agree or not. Its a fun thing to do, you learn a lot and
you are less likely to critize weather forecasts or forecasters.

Mike

 

[Jeff Lawson]

Well, if a forecast that changes over the course of 24 hours is a "crap" forecast, I must be the crappiest forecaster in the world.

 

[Shane Adams]

Well, if a forecast that changes over the course of 24 hours is a "crap" forecast, I must be the crappiest forecaster in the world.

I guess that means I'm Crappy Dew (get it, dew? )

 

[Tim Vasquez]

Whoa, wait a second -- there's nothing wrong with criticizing a NWS forecast. Clearly the forecast busted. That's not in dispute. Chris wanted to know why the OUN office blew that forecast, and he did not point fingers at anyone.

Ok, we're all human; maybe it was a new forecaster, or they had a bad day and missed something. That's FINE. What about the meteorology behind this? What can we all learn from this difficult forecast? Ascribing it to poor model performance is nothing more than an unqualified assumption. Model performance issues often reveal themselves early into the run (verification), and the forecast process does include other techniques besides models. The search for that nugget of truth that would have made a better forecast is an honorable one. It is science.

Yes, Chris posted a controversial title to this thread, but why can we not keep the criticism to a sentence or two and dig on into the meteorology for that forecast date? Granted Chris may not know how difficult forecasting is, but this was a tricky situation that even the NWS had trouble with, and so this is really more about an interesting and challenging forecast situation than it is about Chris. What can be learned? Isn't that why we're here?

Tim