[Met_help] [rt.rap.ucar.edu #62830] History for Series-Analysis Tool

[John Halley Gotway via RT]

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

Hi,John

Now, I Interest Series-Analysis Tool. I try to read the MET Users Guide
4.1, but it is not clear, you are an example tutorial about
Series-Analysis Tool please ?

                                                                Best regards
                                                                  K.SUJIRA

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

Subject: Series-Analysis Tool
From: John Halley Gotway
Time: Wed Aug 28 09:53:40 2013

Hello,

I created a new help desk ticket for your recent question about the
Series-Analysis tool.

We are currently working on the online tutorial for METv4.1 which will
include an example of running the Series-Analysis tool.  Unfortunately
though, it's not available yet.

The tool is described in Chapter 6 of the MET User's Guide:
   http://www.dtcenter.org/met/users/docs/users_guide/MET_Users_Guide_v4.1.pdf

The Series-Analysis tool provides a grid-to-grid comparison.  So your
forecast and observation files need to be gridded.  Then you define a
"series" of matching forecast and observation files to be paired up.
That series could a multiple vertical levels of model output - but,
most commonly, it'd be a time series of data.  Then you just pass the
list of forecast files and observation files to the tool, along with a
config file:

series_analysis \
-fcst fcst_file1 fcst_file2 fcst_file3 \
-obs   obs_file1  obs_file2 obs_file3  \
-out series_stats.nc \
SeriesAnalysisConfig

In the config file, you tell the tool what field should be selected
from each file, and you tell it which statistics should be computed
from each line type.

The output is a gridded NetCDF file containing 1 variable for each of
the output statistics you've selected.  The value at each grid point
is the statistic computed over all pairs in the series at that grid
point.

The nice thing here is that you're able to see the spatial structure
of model errors.

Hopefully that helps you get started.

Thanks,
John Halley Gotway
met_help at ucar.edu

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