Jump to content

2021-22 Global Model QPF/Snowfall Evaluation For PHL. 11/28-29 Trace Event. Constanza Bizzaro November Continues As GFS Is 1st & EC Family EPS Is Last.


Recommended Posts

Tom & I want to thank everyone for all the looks this thread gets. Our plans are to keep the same models/ensembles as last season.  We will see if the UKMET remains available, might force a change.

There were 47 precipitation events evaluated which included 12 predicted or observed snow events from Philly. Yes if you were north and west of Philly, the modeling snowy bias need not apply. That is the way last season went and with another nina season coming may happen again. 

Because it had such a hot start, the EPS was the most accurate model for PHL last season and had the least wet bias. Yes the models inherently are too wet for PHL.  26 of the 47 events every model/ensemble averaged (yes there could have been a dry run or two) too wet; 15 events some were too wet, some too dry and there were only six events in which every model and ensemble averaged too dry. Most had to do with convective events.  We try to avoid convective events because the models skill goes downhill.  None of these events we included PHL was an isolated convective outlier.  We might have to end the evaluation sooner if next April goes the same way.  It is also the reason we are not starting with event #1 this week. 

If you have any comments or suggestions, please let us know. We'd like to see comments as the events unfurl. I never come close to catching everything.

 

 

snowy.JPG

wety.JPG

zdry.JPG

  • Like 6
Link to post
Share on other sites
  • Replies 52
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Tom & I want to thank everyone for all the looks this thread gets. Our plans are to keep the same models/ensembles as last season.  We will see if the UKMET remains available, might force a change

euro 10/27 0z: .82 10/27 12z: .91 10/28 0z: .6 10/28 12z: 1 10/29 0z: .94 eps 10/27 0z: 1.22 10/27 12z: .97 10/28 0z: 1.06 10/28 12z: .94  10/29

I think we can go thru 00z on the 26th (this evening) with this one. Plus I am "assuming" an ensemble (hello WEFs) might get us there, can we start the Sunday one with the 12z run today? (25th) F

Posted Images

On 10/27/2021 at 8:53 PM, tombo82685 said:

euro

10/27 0z: .82

10/27 12z: .91

10/28 0z: .6

10/28 12z: 1

eps

10/27 0z: 1.22

10/27 12z: .97

10/28 0z: 1.06

10/28 12z: .94 

Well it is coming in faster, so the 00z run this morning (29th) can be the last one.

Link to post
Share on other sites
On 10/24/2021 at 10:58 PM, Rainshadow23.0 said:

Tom & I want to thank everyone for all the looks this thread gets. Our plans are to keep the same models/ensembles as last season.  We will see if the UKMET remains available, might force a change.

There were 47 precipitation events evaluated which included 12 predicted or observed snow events from Philly. Yes if you were north and west of Philly, the modeling snowy bias need not apply. That is the way last season went and with another nina season coming may happen again. 

Because it had such a hot start, the EPS was the most accurate model for PHL last season and had the least wet bias. Yes the models inherently are too wet for PHL.  26 of the 47 events every model/ensemble averaged (yes there could have been a dry run or two) too wet; 15 events some were too wet, some too dry and there were only six events in which every model and ensemble averaged too dry. Most had to do with convective events.  We try to avoid convective events because the models skill goes downhill.  None of these events we included PHL was an isolated convective outlier.  We might have to end the evaluation sooner if next April goes the same way.  It is also the reason we are not starting with event #1 this week. 

If you have any comments or suggestions, please let us know. We'd like to see comments as the events unfurl. I never come close to catching everything.

 

 

snowy.JPG

wety.JPG

zdry.JPG

Error means how far off the average for each particular model was from the actual?  How many model runs were used to calculate the error?

 

Bias measures too wet or too dry for the same time period average as above?

Link to post
Share on other sites
48 minutes ago, susqushawn said:

Error means how far off the average for each particular model was from the actual?  How many model runs were used to calculate the error?

 

Bias measures too wet or too dry for the same time period average as above?

Error, yes.  Model runs it depends on the event.  It is typically around 6 sounding runs before the onset of precipitation.  It can be less if the event moves up in timing like this morning's or for longer duration events.  We like to see the end to the event get to around 84 hours of go time before starting to look at the qpf/snowfall.

Bias, positive numbers too wet/snowy, negative too dry/not snowy enough.  This is the file for 2020-21.

2020-21qpf.xlsx

Link to post
Share on other sites
1 hour ago, Rainshadow23.0 said:

Well it is coming in faster, so the 00z run this morning (29th) can be the last one.

Can probably do 12z. Doubt philly even measured with this

Link to post
Share on other sites
18 hours ago, tombo82685 said:

Damn, didn't think they would. Didn't even measure anything at work 

Going to have to wait for tonight to make sure about this one.  Meanwhile the WEFS is saying I am going to make you do one for Tuesday.  We might have to start that one 00z tonight if it remains WEFy.

Link to post
Share on other sites

Yes it was wetter in other places, but as far as PHL goes, the least wet bias bias came closest, the UKMET.  Out of 30 possible forecasts, 29 were too wet. One ECMWF run nailed PHL's pcpn exactly.  No surprise to many, the wettest (last place) of them all was the WEFS, I mean GEFS. We started on exactly the same two days last year, so I would cancel any outdoor plans for Oct 29-30, 2022.

pcpn.JPG.8f34961443657f46cc3baadad4e17dce.JPG

 

 

past72.png

102931.JPG

Link to post
Share on other sites
  • DebShadow23.0 changed the title to 2021-22 Global Model QPF/Snowfall Evaluation For PHL. Inaugural 10/29-30 Event Went To The UKMET (Remember Me?) & The WEFS Came In Last.

euro

10/31 0z: .07

10/31 12z: .05

11/1 0z: .03

11/1 12z: .02

11/2 0z: .05

11/2 12z: .11

eps

10/31 0z: .07

10/31 12z: .05

11/1 0z: .05

11/1 12z: .04

11/2 0z: .06

11/2 12z: .11

Link to post
Share on other sites
On 10/31/2021 at 1:42 PM, tombo82685 said:

euro

10/31 0z: .07

10/31 12z: .05

11/1 0z: .03

11/1 12z: .02

eps

10/31 0z: .07

10/31 12z: .05

11/1 0z: .05

11/1 12z: .04

We will go thru the 12z run today (November 2nd) and thank-you very little WEFs for this.

Link to post
Share on other sites

Well another last place finish for the GFS Family, this time the OP GFS.  Rule Of thumb, if you are wetter than the GEFS, you are in trouble.  The first place model was the GGEM. Granted it was too dry and did not handle the north and west aspect that well.  The surprise so far early is no mention of the EC or EPS.  The EPS is not having the hot start it had last season.  Interesting on how we are following a similar path to last season (yeah early).  A soaking late October event, then a light one and now what appears to be a big time gap before the next event (assuming weekend system stays offshore).

 

awet.JPG

113.JPG

Link to post
Share on other sites
  • DebShadow23.0 changed the title to 2021-22 Global Model QPF/Snowfall Evaluation For PHL. 11/2 Does It Or Does It Not Event Went To The GGEM. Another Last Place Finish For The GFS Family.
5 hours ago, tombo82685 said:

eps

11/9 12z: .78

11/10 0z: .68

euro

11/9 12z: .97

11/10 0z: .55

I am going to say the WEFS will put us over the top today, so can we start with the 12z run today the separate Saturday wave on front precip.

Link to post
Share on other sites
  • DebShadow23.0 changed the title to 2021-22 Global Model QPF/Snowfall Evaluation For PHL. 11/28-29 Trace Event. Constanza Bizzaro November Continues As GFS Is 1st & EC Family EPS Is Last.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    No registered users viewing this page.


×
×
  • Create New...