[Met_help] [rt.rap.ucar.edu #73528] History for Anomaly correlation
John Halley Gotway via RT
met_help at ucar.edu
Thu Sep 24 15:32:33 MDT 2015
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi,
I have two sets of model data with 6 hourly forecasts fields for about 3
months (~360 time steps). What MET5 tool should be used to find anomaly
correlation of couple of variable fields between these models ? Any help.
Thanks.
Debasish
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Anomaly correlation
From: John Halley Gotway
Time: Thu Sep 24 10:10:00 2015
Debasish,
I see that you have output from two different forecast models. And I
see
that you want to compute the correlation between them.
"Anomaly Correlation" is typically computed relative to some
climatology
dataset, whereas the Pearson Correlation is computed between any two
difference datasets.
My guess is that you actually just want to compute the Pearson
Correlation
between data from the two models. If that's correct, you would use
the
Grid-Stat tool in MET. It compares gridded fields to eachother and,
among
other things, computes continuous statistics for that comparison. The
continuous statistics are written out in the CNT line type and the
Pearson
Correlation is written out in the PR_CORR column of that output line
type.
You can find examples of running Grid-Stat in our online MET tutorial:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
Hope that helps.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT
<met_help at ucar.edu>
wrote:
>
> Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> Transaction: Ticket created by debasish.hazra5 at gmail.com
> Queue: met_help
> Subject: Anomaly correlation
> Owner: Nobody
> Requestors: debasish.hazra5 at gmail.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
>
>
> Hi,
>
> I have two sets of model data with 6 hourly forecasts fields for
about 3
> months (~360 time steps). What MET5 tool should be used to find
anomaly
> correlation of couple of variable fields between these models ? Any
help.
>
> Thanks.
> Debasish
>
>
------------------------------------------------
Subject: Anomaly correlation
From: Debasish Hazra
Time: Thu Sep 24 11:00:14 2015
Thanks John. Two models dimensions are (lat,lon,time), so after
computing
continuous statistics, Pearson Correlation (PR_CORR column in the
output)
will be in (lat,lon) for each time step or just a single number for
each
time step ?
Debasish
---------- Forwarded message ----------
From: John Halley Gotway via RT <met_help at ucar.edu>
Date: Thu, Sep 24, 2015 at 12:10 PM
Subject: Re: [rt.rap.ucar.edu #73528] Anomaly correlation
To: debasish.hazra5 at gmail.com
Debasish,
I see that you have output from two different forecast models. And I
see
that you want to compute the correlation between them.
"Anomaly Correlation" is typically computed relative to some
climatology
dataset, whereas the Pearson Correlation is computed between any two
difference datasets.
My guess is that you actually just want to compute the Pearson
Correlation
between data from the two models. If that's correct, you would use
the
Grid-Stat tool in MET. It compares gridded fields to eachother and,
among
other things, computes continuous statistics for that comparison. The
continuous statistics are written out in the CNT line type and the
Pearson
Correlation is written out in the PR_CORR column of that output line
type.
You can find examples of running Grid-Stat in our online MET tutorial:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
Hope that helps.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT
<met_help at ucar.edu>
wrote:
>
> Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> Transaction: Ticket created by debasish.hazra5 at gmail.com
> Queue: met_help
> Subject: Anomaly correlation
> Owner: Nobody
> Requestors: debasish.hazra5 at gmail.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
>
>
> Hi,
>
> I have two sets of model data with 6 hourly forecasts fields for
about 3
> months (~360 time steps). What MET5 tool should be used to find
anomaly
> correlation of couple of variable fields between these models ? Any
help.
>
> Thanks.
> Debasish
>
>
------------------------------------------------
Subject: Anomaly correlation
From: John Halley Gotway
Time: Thu Sep 24 11:17:59 2015
Debasish,
You have two options.
(1) The Grid-Stat tool is run once for each valid time. It compares
gridded forecast and observed data across many grid points for that
single
point in time. You have the ability to define multiple "masking
regions"
(i.e. spatial areas) over which to compute statistics, but you'd only
get 1
correlation value for each spatial masking region. For example,
suppose
you're comparing data over the continental United States, and you'd
defined
48 different masking regions - one for each state. You'd end up with
48
different correlation values, each one computed using the grid points
that
fell within that state.
(2) The second option is to use the Series-Analysis tool. While the
Grid-Stat tool compares a single forecast and observation field and
computes spatial averages, the Series-Analysis tool compares a series
of
forecast and observation fields. Usually that series is defined as a
time
series. Rather than computing spatially averaged statistics like
Grid-Stat, Series-Analysis computes one or more statistics at each
grid
point over the series of forecast/observation pairs. It writes a
NetCDF
output file containing a gridded statistics. The advantage to
Series-Analysis is seeing how your model performance varies over the
domain.
Which tool best fits your needs?
Thanks,
John
On Thu, Sep 24, 2015 at 11:00 AM, Debasish Hazra via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
>
> Thanks John. Two models dimensions are (lat,lon,time), so after
computing
> continuous statistics, Pearson Correlation (PR_CORR column in the
output)
> will be in (lat,lon) for each time step or just a single number
for each
> time step ?
>
> Debasish
>
> ---------- Forwarded message ----------
> From: John Halley Gotway via RT <met_help at ucar.edu>
> Date: Thu, Sep 24, 2015 at 12:10 PM
> Subject: Re: [rt.rap.ucar.edu #73528] Anomaly correlation
> To: debasish.hazra5 at gmail.com
>
>
> Debasish,
>
> I see that you have output from two different forecast models. And
I see
> that you want to compute the correlation between them.
>
> "Anomaly Correlation" is typically computed relative to some
climatology
> dataset, whereas the Pearson Correlation is computed between any two
> difference datasets.
>
> My guess is that you actually just want to compute the Pearson
Correlation
> between data from the two models. If that's correct, you would use
the
> Grid-Stat tool in MET. It compares gridded fields to eachother and,
among
> other things, computes continuous statistics for that comparison.
The
> continuous statistics are written out in the CNT line type and the
Pearson
> Correlation is written out in the PR_CORR column of that output line
type.
>
> You can find examples of running Grid-Stat in our online MET
tutorial:
>
>
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
>
> Hope that helps.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
> On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> > Transaction: Ticket created by debasish.hazra5 at gmail.com
> > Queue: met_help
> > Subject: Anomaly correlation
> > Owner: Nobody
> > Requestors: debasish.hazra5 at gmail.com
> > Status: new
> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
> >
> >
> > Hi,
> >
> > I have two sets of model data with 6 hourly forecasts fields for
about 3
> > months (~360 time steps). What MET5 tool should be used to find
anomaly
> > correlation of couple of variable fields between these models ?
Any help.
> >
> > Thanks.
> > Debasish
> >
> >
>
>
------------------------------------------------
Subject: Anomaly correlation
From: Debasish Hazra
Time: Thu Sep 24 11:58:49 2015
Thanks John. Is there a way, so that we can provide a spatial region
(say
over Africa) grid ,compute the 2 option using Series analysis and that
will
result correlation stat over that "masked" grid box, rather than at
each
grid for all 300 time steps ?
Debasish.
On Thu, Sep 24, 2015 at 1:17 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Debasish,
>
> You have two options.
>
> (1) The Grid-Stat tool is run once for each valid time. It compares
> gridded forecast and observed data across many grid points for that
single
> point in time. You have the ability to define multiple "masking
regions"
> (i.e. spatial areas) over which to compute statistics, but you'd
only get 1
> correlation value for each spatial masking region. For example,
suppose
> you're comparing data over the continental United States, and you'd
defined
> 48 different masking regions - one for each state. You'd end up
with 48
> different correlation values, each one computed using the grid
points that
> fell within that state.
>
> (2) The second option is to use the Series-Analysis tool. While the
> Grid-Stat tool compares a single forecast and observation field and
> computes spatial averages, the Series-Analysis tool compares a
series of
> forecast and observation fields. Usually that series is defined as
a time
> series. Rather than computing spatially averaged statistics like
> Grid-Stat, Series-Analysis computes one or more statistics at each
grid
> point over the series of forecast/observation pairs. It writes a
NetCDF
> output file containing a gridded statistics. The advantage to
> Series-Analysis is seeing how your model performance varies over the
> domain.
>
> Which tool best fits your needs?
>
> Thanks,
> John
>
> On Thu, Sep 24, 2015 at 11:00 AM, Debasish Hazra via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
> >
> > Thanks John. Two models dimensions are (lat,lon,time), so after
computing
> > continuous statistics, Pearson Correlation (PR_CORR column in the
output)
> > will be in (lat,lon) for each time step or just a single number
for
> each
> > time step ?
> >
> > Debasish
> >
> > ---------- Forwarded message ----------
> > From: John Halley Gotway via RT <met_help at ucar.edu>
> > Date: Thu, Sep 24, 2015 at 12:10 PM
> > Subject: Re: [rt.rap.ucar.edu #73528] Anomaly correlation
> > To: debasish.hazra5 at gmail.com
> >
> >
> > Debasish,
> >
> > I see that you have output from two different forecast models.
And I see
> > that you want to compute the correlation between them.
> >
> > "Anomaly Correlation" is typically computed relative to some
climatology
> > dataset, whereas the Pearson Correlation is computed between any
two
> > difference datasets.
> >
> > My guess is that you actually just want to compute the Pearson
> Correlation
> > between data from the two models. If that's correct, you would
use the
> > Grid-Stat tool in MET. It compares gridded fields to eachother
and,
> among
> > other things, computes continuous statistics for that comparison.
The
> > continuous statistics are written out in the CNT line type and the
> Pearson
> > Correlation is written out in the PR_CORR column of that output
line
> type.
> >
> > You can find examples of running Grid-Stat in our online MET
tutorial:
> >
> >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
> >
> > Hope that helps.
> >
> > Thanks,
> > John Halley Gotway
> > met_help at ucar.edu
> >
> > On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT <
> met_help at ucar.edu>
> > wrote:
> >
> > >
> > > Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> > > Transaction: Ticket created by debasish.hazra5 at gmail.com
> > > Queue: met_help
> > > Subject: Anomaly correlation
> > > Owner: Nobody
> > > Requestors: debasish.hazra5 at gmail.com
> > > Status: new
> > > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528
> >
> > >
> > >
> > > Hi,
> > >
> > > I have two sets of model data with 6 hourly forecasts fields
for
> about 3
> > > months (~360 time steps). What MET5 tool should be used to find
anomaly
> > > correlation of couple of variable fields between these models ?
Any
> help.
> > >
> > > Thanks.
> > > Debasish
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Anomaly correlation
From: John Halley Gotway
Time: Thu Sep 24 12:16:30 2015
Debasish,
Yes, the Series-Analysis configuration file includes an option of
masking:
mask = {
grid = "";
poly = "/path/to/masking/polyline.txt";
};
The easiest thing would be defining a lat/lon polyline masking region.
You'd put a list of lat/lon points into an ASCII file and put the path
to
that file into the Series-Analysis config file. Then Series-Analysis
would
only compute time series statistics at the grid points falling within
the
region defined by that mask.
All that does is save computing time. I'd suggest starting by running
over
the full model domain. If you find that it runs too slowly, then you
could
use the polyline mask to limit the analysis to the grid points falling
in
Africa.
Feel free to give it a try and let us know what questions or issues
you
encounter.
Thanks,
John
On Thu, Sep 24, 2015 at 11:58 AM, Debasish Hazra via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
>
> Thanks John. Is there a way, so that we can provide a spatial region
(say
> over Africa) grid ,compute the 2 option using Series analysis and
that will
> result correlation stat over that "masked" grid box, rather than at
each
> grid for all 300 time steps ?
>
> Debasish.
>
> On Thu, Sep 24, 2015 at 1:17 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Debasish,
> >
> > You have two options.
> >
> > (1) The Grid-Stat tool is run once for each valid time. It
compares
> > gridded forecast and observed data across many grid points for
that
> single
> > point in time. You have the ability to define multiple "masking
regions"
> > (i.e. spatial areas) over which to compute statistics, but you'd
only
> get 1
> > correlation value for each spatial masking region. For example,
suppose
> > you're comparing data over the continental United States, and
you'd
> defined
> > 48 different masking regions - one for each state. You'd end up
with 48
> > different correlation values, each one computed using the grid
points
> that
> > fell within that state.
> >
> > (2) The second option is to use the Series-Analysis tool. While
the
> > Grid-Stat tool compares a single forecast and observation field
and
> > computes spatial averages, the Series-Analysis tool compares a
series of
> > forecast and observation fields. Usually that series is defined
as a
> time
> > series. Rather than computing spatially averaged statistics like
> > Grid-Stat, Series-Analysis computes one or more statistics at each
grid
> > point over the series of forecast/observation pairs. It writes a
NetCDF
> > output file containing a gridded statistics. The advantage to
> > Series-Analysis is seeing how your model performance varies over
the
> > domain.
> >
> > Which tool best fits your needs?
> >
> > Thanks,
> > John
> >
> > On Thu, Sep 24, 2015 at 11:00 AM, Debasish Hazra via RT <
> met_help at ucar.edu
> > >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
> > >
> > > Thanks John. Two models dimensions are (lat,lon,time), so after
> computing
> > > continuous statistics, Pearson Correlation (PR_CORR column in
the
> output)
> > > will be in (lat,lon) for each time step or just a single
number for
> > each
> > > time step ?
> > >
> > > Debasish
> > >
> > > ---------- Forwarded message ----------
> > > From: John Halley Gotway via RT <met_help at ucar.edu>
> > > Date: Thu, Sep 24, 2015 at 12:10 PM
> > > Subject: Re: [rt.rap.ucar.edu #73528] Anomaly correlation
> > > To: debasish.hazra5 at gmail.com
> > >
> > >
> > > Debasish,
> > >
> > > I see that you have output from two different forecast models.
And I
> see
> > > that you want to compute the correlation between them.
> > >
> > > "Anomaly Correlation" is typically computed relative to some
> climatology
> > > dataset, whereas the Pearson Correlation is computed between any
two
> > > difference datasets.
> > >
> > > My guess is that you actually just want to compute the Pearson
> > Correlation
> > > between data from the two models. If that's correct, you would
use the
> > > Grid-Stat tool in MET. It compares gridded fields to eachother
and,
> > among
> > > other things, computes continuous statistics for that
comparison. The
> > > continuous statistics are written out in the CNT line type and
the
> > Pearson
> > > Correlation is written out in the PR_CORR column of that output
line
> > type.
> > >
> > > You can find examples of running Grid-Stat in our online MET
tutorial:
> > >
> > >
> > >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
> > >
> > > Hope that helps.
> > >
> > > Thanks,
> > > John Halley Gotway
> > > met_help at ucar.edu
> > >
> > > On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT <
> > met_help at ucar.edu>
> > > wrote:
> > >
> > > >
> > > > Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> > > > Transaction: Ticket created by debasish.hazra5 at gmail.com
> > > > Queue: met_help
> > > > Subject: Anomaly correlation
> > > > Owner: Nobody
> > > > Requestors: debasish.hazra5 at gmail.com
> > > > Status: new
> > > > Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528
> > >
> > > >
> > > >
> > > > Hi,
> > > >
> > > > I have two sets of model data with 6 hourly forecasts fields
for
> > about 3
> > > > months (~360 time steps). What MET5 tool should be used to
find
> anomaly
> > > > correlation of couple of variable fields between these models
? Any
> > help.
> > > >
> > > > Thanks.
> > > > Debasish
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Anomaly correlation
From: Debasish Hazra
Time: Thu Sep 24 14:35:56 2015
Thanks John. I will try that and let you know, if I need any help.
Debasish.
On Thu, Sep 24, 2015 at 2:16 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Debasish,
>
> Yes, the Series-Analysis configuration file includes an option of
masking:
>
> mask = {
> grid = "";
> poly = "/path/to/masking/polyline.txt";
> };
>
> The easiest thing would be defining a lat/lon polyline masking
region.
> You'd put a list of lat/lon points into an ASCII file and put the
path to
> that file into the Series-Analysis config file. Then Series-
Analysis would
> only compute time series statistics at the grid points falling
within the
> region defined by that mask.
>
> All that does is save computing time. I'd suggest starting by
running over
> the full model domain. If you find that it runs too slowly, then
you could
> use the polyline mask to limit the analysis to the grid points
falling in
> Africa.
>
> Feel free to give it a try and let us know what questions or issues
you
> encounter.
>
> Thanks,
> John
>
>
>
>
> On Thu, Sep 24, 2015 at 11:58 AM, Debasish Hazra via RT
<met_help at ucar.edu
> >
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528 >
> >
> > Thanks John. Is there a way, so that we can provide a spatial
region (say
> > over Africa) grid ,compute the 2 option using Series analysis and
that
> will
> > result correlation stat over that "masked" grid box, rather than
at each
> > grid for all 300 time steps ?
> >
> > Debasish.
> >
> > On Thu, Sep 24, 2015 at 1:17 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Debasish,
> > >
> > > You have two options.
> > >
> > > (1) The Grid-Stat tool is run once for each valid time. It
compares
> > > gridded forecast and observed data across many grid points for
that
> > single
> > > point in time. You have the ability to define multiple "masking
> regions"
> > > (i.e. spatial areas) over which to compute statistics, but you'd
only
> > get 1
> > > correlation value for each spatial masking region. For example,
> suppose
> > > you're comparing data over the continental United States, and
you'd
> > defined
> > > 48 different masking regions - one for each state. You'd end up
with
> 48
> > > different correlation values, each one computed using the grid
points
> > that
> > > fell within that state.
> > >
> > > (2) The second option is to use the Series-Analysis tool. While
the
> > > Grid-Stat tool compares a single forecast and observation field
and
> > > computes spatial averages, the Series-Analysis tool compares a
series
> of
> > > forecast and observation fields. Usually that series is defined
as a
> > time
> > > series. Rather than computing spatially averaged statistics
like
> > > Grid-Stat, Series-Analysis computes one or more statistics at
each grid
> > > point over the series of forecast/observation pairs. It writes
a
> NetCDF
> > > output file containing a gridded statistics. The advantage to
> > > Series-Analysis is seeing how your model performance varies over
the
> > > domain.
> > >
> > > Which tool best fits your needs?
> > >
> > > Thanks,
> > > John
> > >
> > > On Thu, Sep 24, 2015 at 11:00 AM, Debasish Hazra via RT <
> > met_help at ucar.edu
> > > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528
>
> > > >
> > > > Thanks John. Two models dimensions are (lat,lon,time), so
after
> > computing
> > > > continuous statistics, Pearson Correlation (PR_CORR column in
the
> > output)
> > > > will be in (lat,lon) for each time step or just a single
number for
> > > each
> > > > time step ?
> > > >
> > > > Debasish
> > > >
> > > > ---------- Forwarded message ----------
> > > > From: John Halley Gotway via RT <met_help at ucar.edu>
> > > > Date: Thu, Sep 24, 2015 at 12:10 PM
> > > > Subject: Re: [rt.rap.ucar.edu #73528] Anomaly correlation
> > > > To: debasish.hazra5 at gmail.com
> > > >
> > > >
> > > > Debasish,
> > > >
> > > > I see that you have output from two different forecast models.
And I
> > see
> > > > that you want to compute the correlation between them.
> > > >
> > > > "Anomaly Correlation" is typically computed relative to some
> > climatology
> > > > dataset, whereas the Pearson Correlation is computed between
any two
> > > > difference datasets.
> > > >
> > > > My guess is that you actually just want to compute the Pearson
> > > Correlation
> > > > between data from the two models. If that's correct, you
would use
> the
> > > > Grid-Stat tool in MET. It compares gridded fields to
eachother and,
> > > among
> > > > other things, computes continuous statistics for that
comparison.
> The
> > > > continuous statistics are written out in the CNT line type and
the
> > > Pearson
> > > > Correlation is written out in the PR_CORR column of that
output line
> > > type.
> > > >
> > > > You can find examples of running Grid-Stat in our online MET
> tutorial:
> > > >
> > > >
> > > >
> > >
> >
>
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.0/grid_stat/index.php
> > > >
> > > > Hope that helps.
> > > >
> > > > Thanks,
> > > > John Halley Gotway
> > > > met_help at ucar.edu
> > > >
> > > > On Thu, Sep 24, 2015 at 9:11 AM, Debasish Hazra via RT <
> > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > >
> > > > > Thu Sep 24 09:11:15 2015: Request 73528 was acted upon.
> > > > > Transaction: Ticket created by debasish.hazra5 at gmail.com
> > > > > Queue: met_help
> > > > > Subject: Anomaly correlation
> > > > > Owner: Nobody
> > > > > Requestors: debasish.hazra5 at gmail.com
> > > > > Status: new
> > > > > Ticket <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=73528
> > > >
> > > > >
> > > > >
> > > > > Hi,
> > > > >
> > > > > I have two sets of model data with 6 hourly forecasts
fields for
> > > about 3
> > > > > months (~360 time steps). What MET5 tool should be used to
find
> > anomaly
> > > > > correlation of couple of variable fields between these
models ? Any
> > > help.
> > > > >
> > > > > Thanks.
> > > > > Debasish
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
More information about the Met_help
mailing list