[Met_help] [rt.rap.ucar.edu #86903] History for Combining model/forecast files

Wed Sep 12 10:48:16 MDT 2018

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

Good afternoon,

I am currently running MET software and have learned how to implement it at discrete times (i.e., forecast hour 3). I was wondering if it was possible to use the software to compute statistics over a time window (i.e., 2nd - 8th forecast hours) and get POD, FSS, GSS, etc. over that time span. I know the suite includes "pcp_combine" which is helpful for accumulating variables. I would prefer to compare model and observations at each time and compute statistics over a given time span.

Thank you for your help!
Marcus

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

Subject: Combining model/forecast files
From: John Halley Gotway
Time: Mon Sep 10 14:36:01 2018

Hi Marcus,

I see you have a question about how to a aggregate results through
time
using MET.  Generally, the MET tools are intended to be run once per
output
time.  For example, if verifying precip using the Grid-Stat tool,
you'd run
it separately for forecast hours 2, 3, 4,..., and 8.

Then you'd run the STAT-Analysis tool aggregate results through time.
You
can read about the STAT-Analysis tool here:

https://dtcenter.org/met/users/support/online_tutorial/METv6.1/tutorial.php?name=stat_analysis&category=index

Since you're interested in categorical stats, be sure to configure
Grid-Stat to write the CTC output line.  And let's say your output
files
live in a directory named "out_grid_stat".  You'd run a STAT-Analysis
job
like this:

   stat_analysis -lookin out_grid_stat -job aggregate_stat -line_type
CTC
-out_line_type CTS -by VX_MASK,FCST_THRESH -out_stat
agg_CTC_to_CTS.stat

This job...
(1) Reads .stat files it find when searching recursively through the
out_grid_stat directory.
(2) Uses only CTC lines from the input.
(3) Aggregates the CTC lines together and derive contingency table
stats
(CTS line type), including PODY and GSS.
(4) Applies this logic for each unique entry it finds in the VX_MASK
and
FCST_THRESH columns.
(5) Write the resulting output CTS lines to a file named
agg_CTC_to_CTS.stat

Hope that helps point you in the right direction.

One other thing to mention.  In our internal use at the DTC, we load
the
ascii output of MET in the METviewer database and display system.
Then we
aggregate results through time using METviewer while also plotting the
data.

Thanks,
John

On Mon, Sep 10, 2018 at 1:53 PM Johnson, Marcus R. via RT
<met_help at ucar.edu>
wrote:

>
> Mon Sep 10 13:53:42 2018: Request 86903 was acted upon.
> Transaction: Ticket created by marcus.johnson at ou.edu
>        Queue: met_help
>      Subject: Combining model/forecast files
>        Owner: Nobody
>   Requestors: marcus.johnson at ou.edu
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=86903 >
>
>
> Good afternoon,
>
> I am currently running MET software and have learned how to
implement it
> at discrete times (i.e., forecast hour 3). I was wondering if it was
> possible to use the software to compute statistics over a time
window
> (i.e., 2nd - 8th forecast hours) and get POD, FSS, GSS, etc. over
that time
> span. I know the suite includes "pcp_combine" which is helpful for
> accumulating variables. I would prefer to compare model and
observations at
> each time and compute statistics over a given time span.
>
> Thank you for your help!
> Marcus
>
>

------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #86903] Combining model/forecast files
From: Johnson, Marcus R.
Time: Tue Sep 11 12:45:53 2018

Thank you so much for your help John. That clarifies things and lets
me know exactly what I need to do.

Marcus

-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Monday, September 10, 2018 3:36 PM
To: Johnson, Marcus R.
Subject: Re: [rt.rap.ucar.edu #86903] Combining model/forecast files