Jump to content
  • Member Statistics

    17,611
    Total Members
    7,904
    Most Online
    NH8550
    Newest Member
    NH8550
    Joined

Central PA & The Fringes - December 2013 Part III


PennMan

Recommended Posts

this storm has shown me that the models need some major work still. I work in IT. I teach IT at the undergrad level. I know that the models get revised every so often, it looks like that every model has it's own bias and they use a weighted average to make a prediction. Using Moore's law (technology doubles every 18 to 24 months as cost goes down) I do not see the accuracy getting better.

 

Just think, with the drop in cost and increase in performance, we could have models that run continuously and that are far more accurate that what we have today.

Link to comment
Share on other sites

  • Replies 672
  • Created
  • Last Reply

this storm has shown me that the models need some major work still. I work in IT. I teach IT at the undergrad level. I know that the models get revised every so often, it looks like that every model has it's own bias and they use a weighted average to make a prediction. Using Moore's law (technology doubles every 18 to 24 months as cost goes down) I do not see the accuracy getting better.

 

Just think, with the drop in cost and increase in performance, we could have models that run continuously and that are far more accurate that what we have today.

What about getting better initial data input too? The best mathemtics / resolution still will have flaws if accurate information is not input into the models.  I do agree though that you think we might see upgrades come in faster, then again the latest upgrade to the Euro seems to have decreased quality of forecasts.

Link to comment
Share on other sites

Garbage in garbage out. Our credo... I do not know thing one about how models are programmed.  Obviously getting better data input is a key. I would have to think about how I would approach building a model. they should be able to run past results into a model and see where it breaks down. basically if you know what is supposed to happen, put in the data and tweak the code until you get the correct results.

 

I know there are billions of data points that would need to be captured but it should be possible.

Link to comment
Share on other sites

There are some absolute beauties in the individuals.

just seeing the 1.0 eeking west is surely a sight for sleepy eyes....but worth getting up for...so far.  Nice to see many panels (but not all) not showing the donut hole.  couple scary ones remain but the mean is a nice step.

 

and agreed...a couple of them would bring us to our knees....in SNOW.

 

Nut

Link to comment
Share on other sites

Anyone want a run if real winter? Here we go! (From CTP btw)

Thursday

Cloudy. Snow in the afternoon. Snow accumulation of 1 to 3 inches. Highs in the lower 30s. Northeast winds 5 to 10 mph. Chance of snow 80 percent.

Thursday Night

Snow...mainly in the evening. Additional moderate snow accumulation. Blustery and much colder with lows around 10 above. Northwest winds 15 to 25 mph. Gusts up to 40 mph after midnight. Chance of snow 90 percent.

Friday

Mostly sunny and blustery. Much colder with highs around 15. Northwest winds 15 to 25 mph with gusts up to 40 mph.

Friday Night

Mostly clear. Lows 5 to 10 above.

...

Monday Night

Partly cloudy. Much colder with lows around zero.

Tuesday

Sunny. Much colder with highs 10 to 15. Wind chill values as low as 15 below.

Link to comment
Share on other sites

Good trends overnight...liking the chances for a 3-6/4-8 event at this point.

 

EDIT: BGM issued a WSW for NEPA zones, says 6-9 possible.

i started to post before I saw your edit.  I think your first line is possible in state college based on srefs.  if your in hazleton, i think the second line is all but a lock.

 

Nut

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...