[Met_help] [rt.rap.ucar.edu #80934] History for topography mask

[John Halley Gotway via RT]

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

I noticed (what appears to my team to be) an issue in VSDB, and I wanted to know if this is a potential issue in MET or MET+ as well.  For GFS pressure-level output, there are values written out even if the surface pressure is less than the pressure level; for example, if a surface pressure in the Himalayas is 500 hPa, a lapse rate is applied to write out temperature values at this location for pressures 550-1000 hPa.  VSDB, however, does not mask for topography, so any potential air temperature differences at the Himalayan surface are extrapolated down to 1000 hPa, and included in bulk statistics calculations at 1000 hPa.  Does MET include a topography mask, or, if given the same GFS output, would it still include these below-surface pressure-level points in bulk statistical calculations?  Thanks.

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

Sean,

Good question.  MET does have the ability to mask based on topography but it requires a 2-step process, where you generate a topography mask that you then apply during computation.  We have a development item to be automatically more sensitive to topography when computing scores. That is slated for inclusion sometime during this next year.

Tara

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

Subject: topography mask
From: John Halley Gotway
Time: Tue Jun 20 11:22:22 2017

Hello Sean,

I see you have a question about how topography data can be used in the
MET tools.

I agree with Tara's analysis.  MET is likely doing the same thing as
the VSDB tools... at least that's how we designed it!

Here's how I interpret your question... please correct me if I'm
wrong.  When verifying GFS as some pressure level (e.g. Temperature at
850mb), can MET be configured to exclude grid points where the surface
pressure is < 850?

The answer to that question is yes, but it may not be as convenient as
you'd like.  MET can do "data masking" where it reads gridded data and
applies a threshold to define a mask.  That is supported in both
gen_vx_mask and the MET statistic tool config files as well.  But if
you need to *intersect* multiple masks (i.e. apply this topo logic but
only in the northern hemisphere), that would need to be done in
gen_vx_mask.

Next, when running grid_stat or point_stat, all the defined masks are
applied to all of the verification tasks.  You can't define them
separately for each task.  You'd only want to apply this 850mb topo
mask when verifying data at 850mb.  The 700mb topo mask would look a
lot different!  Practically speaking then, in each grid_stat or
point_stat run, you'd only verify a single pressure level.  That's
possible to do, but a lot messier scripting it up.

So I also agree with Tara that it'd be preferable to provide a more
direct way of supporting this logic in the point_stat and grid_stat
config files.

Thanks,
John

------------------------------------------------
Subject: topography mask
From: sean.casey at noaa.gov
Time: Tue Jun 20 11:35:19 2017

John,

Thank you for the explanation!  It sounds like there are plans to make
this
process run more smoothly in the future.  I'll discuss what you've
told me
with my colleagues and get back to you if I have any questions once we
decide how to proceed here.

-Sean

On Tue, Jun 20, 2017 at 1:22 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> Hello Sean,
>
> I see you have a question about how topography data can be used in
the MET
> tools.
>
> I agree with Tara's analysis.  MET is likely doing the same thing as
the
> VSDB tools... at least that's how we designed it!
>
> Here's how I interpret your question... please correct me if I'm
wrong.
> When verifying GFS as some pressure level (e.g. Temperature at
850mb), can
> MET be configured to exclude grid points where the surface pressure
is <
> 850?
>
> The answer to that question is yes, but it may not be as convenient
as
> you'd like.  MET can do "data masking" where it reads gridded data
and
> applies a threshold to define a mask.  That is supported in both
> gen_vx_mask and the MET statistic tool config files as well.  But if
you
> need to *intersect* multiple masks (i.e. apply this topo logic but
only in
> the northern hemisphere), that would need to be done in gen_vx_mask.
>
> Next, when running grid_stat or point_stat, all the defined masks
are
> applied to all of the verification tasks.  You can't define them
separately
> for each task.  You'd only want to apply this 850mb topo mask when
> verifying data at 850mb.  The 700mb topo mask would look a lot
different!
> Practically speaking then, in each grid_stat or point_stat run,
you'd only
> verify a single pressure level.  That's possible to do, but a lot
messier
> scripting it up.
>
> So I also agree with Tara that it'd be preferable to provide a more
direct
> way of supporting this logic in the point_stat and grid_stat config
files.
>
> Thanks,
> John
>



--
Sean P. F. Casey
Cooperative Institute for Marine and Atmospheric Sciences
Atlantic Oceanographic and Meteorological Laboratory
Hurricane Research Division
5830 University Research Court
Suite 2884
College Park, MD 20740
206-851-4791

------------------------------------------------
Subject: topography mask
From: John Halley Gotway
Time: Tue Jun 20 17:32:00 2017

Sean,

I added a development task for this feature in JIRA, the software we
currently use to track MET development.  For others on this ticket,
here's
the JIRA issue:
  https://sdg/jira/browse/MET-830

This certainly won't make it into the 6.1 release, but I tentatively
assigned it to the release after that.

Thanks,
John

On Tue, Jun 20, 2017 at 11:35 AM, sean.casey at noaa.gov via RT <
met_help at ucar.edu> wrote:

>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=80934 >
>
> John,
>
> Thank you for the explanation!  It sounds like there are plans to
make this
> process run more smoothly in the future.  I'll discuss what you've
told me
> with my colleagues and get back to you if I have any questions once
we
> decide how to proceed here.
>
> -Sean
>
> On Tue, Jun 20, 2017 at 1:22 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Hello Sean,
> >
> > I see you have a question about how topography data can be used in
the
> MET
> > tools.
> >
> > I agree with Tara's analysis.  MET is likely doing the same thing
as the
> > VSDB tools... at least that's how we designed it!
> >
> > Here's how I interpret your question... please correct me if I'm
wrong.
> > When verifying GFS as some pressure level (e.g. Temperature at
850mb),
> can
> > MET be configured to exclude grid points where the surface
pressure is <
> > 850?
> >
> > The answer to that question is yes, but it may not be as
convenient as
> > you'd like.  MET can do "data masking" where it reads gridded data
and
> > applies a threshold to define a mask.  That is supported in both
> > gen_vx_mask and the MET statistic tool config files as well.  But
if you
> > need to *intersect* multiple masks (i.e. apply this topo logic but
only
> in
> > the northern hemisphere), that would need to be done in
gen_vx_mask.
> >
> > Next, when running grid_stat or point_stat, all the defined masks
are
> > applied to all of the verification tasks.  You can't define them
> separately
> > for each task.  You'd only want to apply this 850mb topo mask when
> > verifying data at 850mb.  The 700mb topo mask would look a lot
different!
> > Practically speaking then, in each grid_stat or point_stat run,
you'd
> only
> > verify a single pressure level.  That's possible to do, but a lot
messier
> > scripting it up.
> >
> > So I also agree with Tara that it'd be preferable to provide a
more
> direct
> > way of supporting this logic in the point_stat and grid_stat
config
> files.
> >
> > Thanks,
> > John
> >
>
>
>
> --
> Sean P. F. Casey
> Cooperative Institute for Marine and Atmospheric Sciences
> Atlantic Oceanographic and Meteorological Laboratory
> Hurricane Research Division
> 5830 University Research Court
> Suite 2884
> College Park, MD 20740
> 206-851-4791
>
>

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