[Met_help] [rt.rap.ucar.edu #85061] History for Ensemble-Stat (UNCLASSIFIED)

Tue Jul 9 12:03:43 MDT 2019

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

CLASSIFICATION: UNCLASSIFIED

Would it be possible to have Ensemble-Stat generate some additional probability information which would enhance/compliment the current NetCDF ensemble forecast values output?

I have attached an image file containing an example of the output from the NOMADS Ensemble Probability Tool which is online at:
https://nomads.ncdc.noaa.gov/EnsProb/

This site provides the probability info generated from the real-time NCEP Global Ensemble model and requires the user to provide inputs to specify the requisite conditions for a given location and the output is computed for each model lead time over the time domain of the model run. For MET users I would think that the input conditions and location as well as the desired outputs could be handled by the configuration file. 

Thanks.

R/
John

Mr. John W. Raby
U.S. Army Research Laboratory
White Sands Missile Range, NM 88002
(575) 678-2004 DSN 258-2004
FAX (575) 678-1230 DSN 258-1230
Email: john.w.raby2.civ at mail.mil

CLASSIFICATION: UNCLASSIFIED

----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Ensemble-Stat (UNCLASSIFIED)
From: John Halley Gotway
Time: Thu May 10 13:59:44 2018

Hi John,

Thanks for sending an example of the requested capability.  Sounds
interesting.

To put it succinctly, you’d like to generate a probability field
defined by
applying thresholds to multiple input fields.  So it becomes a rather
complex definition of an “event”.  Is that correct?

Presumably, you’d then want to verify that probability field by
applying
that same cross-field logic to the observation data.

Currently, when you define “N” fields in ensemble-stat, grid-stat, and
point-stat, you get separate output for all N fields.  One idea would
be
adding a flag to process all N fields *together* with the logic you
requested.  I wonder if that’d be sufficient?

This sounds interesting but will require some significant development.
We
can discuss it more internally here.

Thanks
John

On Wed, May 9, 2018 at 12:35 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:

>
> Wed May 09 12:34:50 2018: Request 85061 was acted upon.
> Transaction: Ticket created by john.w.raby2.civ at mail.mil
>        Queue: met_help
>      Subject: Ensemble-Stat (UNCLASSIFIED)
>        Owner: Nobody
>   Requestors: john.w.raby2.civ at mail.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=85061 >
>
>
> CLASSIFICATION: UNCLASSIFIED
>
> Would it be possible to have Ensemble-Stat generate some additional
> probability information which would enhance/compliment the current
NetCDF
> ensemble forecast values output?
>
> I have attached an image file containing an example of the output
from the
> NOMADS Ensemble Probability Tool which is online at:
> https://nomads.ncdc.noaa.gov/EnsProb/
>
> This site provides the probability info generated from the real-time
NCEP
> Global Ensemble model and requires the user to provide inputs to
specify
> the requisite conditions for a given location and the output is
computed
> for each model lead time over the time domain of the model run. For
MET
> users I would think that the input conditions and location as well
as the
> desired outputs could be handled by the configuration file.
>
> Thanks.
>
> R/
> John
>
> Mr. John W. Raby
> U.S. Army Research Laboratory
> White Sands Missile Range, NM 88002
> (575) 678-2004 DSN 258-2004
> FAX (575) 678-1230 DSN 258-1230
> Email: john.w.raby2.civ at mail.mil
>
>
>
> CLASSIFICATION: UNCLASSIFIED
>
>

------------------------------------------------
Subject: Ensemble-Stat (UNCLASSIFIED)
From: John Halley Gotway
Time: Thu May 10 14:43:12 2018

John,

I talked to Tara Jensen briefly about this.  She informed me that we
have
submitted a proposal recently that would cover this type of work.  If
that
proposal gets funded, then we're in business.  If not, we'll need to
keep
our eyes open for funding agencies that would be interested in this
type of
enhancement.

Thanks,
John

On Thu, May 10, 2018 at 1:59 PM, The RT System itself via RT <
met_help at ucar.edu> wrote:

>
> Thu May 10 13:59:45 2018: Request 85061 was acted upon.
> Transaction: Given to johnhg (John Halley Gotway) by RT_System
>        Queue: met_help
>      Subject: Ensemble-Stat (UNCLASSIFIED)
>        Owner: johnhg
>   Requestors: john.w.raby2.civ at mail.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=85061 >
>
>
> This transaction appears to have no content
>

------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #85061] Ensemble-Stat (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Thu May 10 15:22:01 2018

CLASSIFICATION: UNCLASSIFIED

John -

Thanks for your consideration of this product and your thoughts on
ways to implement it as 1) a method to provide additional ensemble
forecast value output which I think would be useful, and 2) to exploit
the possibility of providing a unique, additional aspect of ensemble
verification.

The initial concept based on the NOMADS tool is to select a location
and compute the results by applying the multiple thresholds. From the
current capability in Ensemble-Stat, the frequency of occurrence of
events defined using thresholds in the config file seems like a step
in the direction of computing the probabilities. I don't know all the
statistical details in how to generate the probabilities, but an
initial capability (Phase 1) might be to generate a field which shows
the probabilities for the thresholds set in the config file and output
that as a NetCDF along with the present forecast fields of mean,
median, range, etc.

Then, further development would include the ideas you've expressed in
your email below which sound reasonable to me. Perhaps the second
phase of this development would be to generate the forecast
probability field defined by applying thresholds to multiple input
fields.

I especially like your idea of using the observations for verification
of the probability field. Perhaps Phase 3 could be the application of
that same cross-field logic to the observation data per your
suggestion and then verify the fcst probability field against the
observed along the lines of your idea.

Regards your idea of adding a flag to process all N fields *together*
with the logic you requested also sounds interesting. I guess to fully
understand this I'd need to learn more of what you are thinking here.

I'll float your ideas about such a tool here within ARL to gain
forecast modeler inputs to further refine the concept and how it could
be used to support our model assessment interests.

R/
John

-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Thursday, May 10, 2018 2:00 PM
To: Raby, John W CIV USARMY RDECOM ARL (US)
<john.w.raby2.civ at mail.mil>
Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #85061] Ensemble-Stat
(UNCLASSIFIED)

All active links contained in this email were disabled.  Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.

----

Hi John,

Thanks for sending an example of the requested capability.  Sounds
interesting.

To put it succinctly, you’d like to generate a probability field
defined by applying thresholds to multiple input fields.  So it
becomes a rather complex definition of an “event”.  Is that correct?

Presumably, you’d then want to verify that probability field by
applying that same cross-field logic to the observation data.

Currently, when you define “N” fields in ensemble-stat, grid-stat, and
point-stat, you get separate output for all N fields.  One idea would
be adding a flag to process all N fields *together* with the logic you
requested.  I wonder if that’d be sufficient?

This sounds interesting but will require some significant development.
We can discuss it more internally here.

Thanks
John

On Wed, May 9, 2018 at 12:35 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:

>
> Wed May 09 12:34:50 2018: Request 85061 was acted upon.
> Transaction: Ticket created by john.w.raby2.civ at mail.mil
>        Queue: met_help
>      Subject: Ensemble-Stat (UNCLASSIFIED)
>        Owner: Nobody
>   Requestors: john.w.raby2.civ at mail.mil
>       Status: new
>  Ticket <Caution-url:
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=85061 >
>
>
> CLASSIFICATION: UNCLASSIFIED
>
> Would it be possible to have Ensemble-Stat generate some additional
> probability information which would enhance/compliment the current
> NetCDF ensemble forecast values output?
>
> I have attached an image file containing an example of the output
from
> the NOMADS Ensemble Probability Tool which is online at:
> Caution-https://nomads.ncdc.noaa.gov/EnsProb/
>
> This site provides the probability info generated from the real-time
> NCEP Global Ensemble model and requires the user to provide inputs
to
> specify the requisite conditions for a given location and the output
> is computed for each model lead time over the time domain of the
model
> run. For MET users I would think that the input conditions and
> location as well as the desired outputs could be handled by the
configuration file.
>
> Thanks.
>
> R/
> John
>
> Mr. John W. Raby
> U.S. Army Research Laboratory
> White Sands Missile Range, NM 88002
> (575) 678-2004 DSN 258-2004
> FAX (575) 678-1230 DSN 258-1230
> Email: john.w.raby2.civ at mail.mil
>
>
>
> CLASSIFICATION: UNCLASSIFIED
>
>

CLASSIFICATION: UNCLASSIFIED

------------------------------------------------