[Met_help] [rt.rap.ucar.edu #57787] History for question about mode output

[Paul Oldenburg via RT]

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

Hi,
 
I am using the mode tool and have the following questions.
 
the verification and forecast maps are daily precipitation on a lat/lon grid
(0.5 deg) over the CONUS. The dimensions are: (150,60) = 9000 points. Since
I am verifying the forecasts only over land, the number of data points is
3389.
 
1) The output ascii file is:  mode_1920000L_20050109_000000V_240000A_cts
The variable TOTAL" is equal to: 8584
Shouldn't the total number of matched pairs be 9000?
 
2) In traditional forecast verification (e.g.,grid_stat), skill is computed
over collocated gridpoints in the verification/forecast maps.
 
In my case study, the verification/forecast maps have two objects each, but
one of the objects in the forecast does not match the objects in the
verification.  
 
I am having problems to understand how the forecast skill is computed in the
mode tool. For example, how the variables FY_OY, FY_ON etc are computed?
 
I attached the output files here.
 
Thanks again for the help.
 
Cheers,
 
Charles.
 
*====================================================*
Dr. Charles Jones
Researcher
Earth Research Institute (ERI)
University of California, Santa Barbara, CA 93106-3060
 
phone: (805) 893-5824
fax:   (805) 893-2578
email:  <mailto:cjones at icess.ucsb.edu> cjones at eri.ucsb.edu
URL:   <http://www.icess.ucsb.edu/asr/> http://www.eri.ucsb.edu/asr/
|=====================================================*
 


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

Subject: Re: [rt.rap.ucar.edu #57787] question about mode output
From: Paul Oldenburg
Time: Fri Aug 10 09:57:02 2012

Charles,

1.  The MODE algorithm does not use the outermost ring of points in
the input grid, for computational reasons.  This
means that the total number of points considered in your case is 9000
- 2*150 - 2*58 = 8584.

2.  The statistics FY_OY, FY_ON, etc. are calculated grid point by
grid point from the trimmed model and observation
fields using a threshold of >0 as follows for each of the three
fields:

RAW:    calculated after the raw thresholding is applied
FITLER: calculated after the filtering step - which is often unused as
in your case
OBJECT: calculated after the object identification step

If you are interested in seeing the gridded data used to calculate the
three contingency tables, set the nc_pairs_flag
to TRUE in your MODE config file (assuming you are using METv4.0) to
generate a NetCDF with the gridded raw and object
information.

If you have any further questions, please let me know.

Paul


On 08/09/2012 05:29 PM, cjones via RT wrote:
>
> Thu Aug 09 17:29:33 2012: Request 57787 was acted upon.
> Transaction: Ticket created by cjones at icess.ucsb.edu
>         Queue: met_help
>       Subject: question about mode output
>         Owner: Nobody
>    Requestors: cjones at icess.ucsb.edu
>        Status: new
>   Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=57787 >
>
>
> Hi,
>
> I am using the mode tool and have the following questions.
>
> the verification and forecast maps are daily precipitation on a
lat/lon grid
> (0.5 deg) over the CONUS. The dimensions are: (150,60) = 9000
points. Since
> I am verifying the forecasts only over land, the number of data
points is
> 3389.
>
> 1) The output ascii file is:
mode_1920000L_20050109_000000V_240000A_cts
> The variable TOTAL" is equal to: 8584
> Shouldn't the total number of matched pairs be 9000?
>
> 2) In traditional forecast verification (e.g.,grid_stat), skill is
computed
> over collocated gridpoints in the verification/forecast maps.
>
> In my case study, the verification/forecast maps have two objects
each, but
> one of the objects in the forecast does not match the objects in the
> verification.
>
> I am having problems to understand how the forecast skill is
computed in the
> mode tool. For example, how the variables FY_OY, FY_ON etc are
computed?
>
> I attached the output files here.
>
> Thanks again for the help.
>
> Cheers,
>
> Charles.
>
> *====================================================*
> Dr. Charles Jones
> Researcher
> Earth Research Institute (ERI)
> University of California, Santa Barbara, CA 93106-3060
>
> phone: (805) 893-5824
> fax:   (805) 893-2578
> email:  <mailto:cjones at icess.ucsb.edu> cjones at eri.ucsb.edu
> URL:   <http://www.icess.ucsb.edu/asr/> http://www.eri.ucsb.edu/asr/
> |=====================================================*
>
>


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