[Met_help] [rt.rap.ucar.edu #89845] History for Metviewer CSI plot without date range selection
John Halley Gotway via RT
met_help at ucar.edu
Mon May 6 14:07:56 MDT 2019
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
To Whom It May Concern:
Case XML : 20190422_151527
I want to compare CSI score for August 2018 between different AQM model
output. But I recall previously I can select the time range for
comparison, e.g., AUG 2018 versus SEP 2018. But this time I have no option
for selecting the time range.
My Question is : Can CSI be computed for a specific time range? or it just
use the wntire data in the database? IS it a problem in pareparing the
database ? i.e., mv_emc_cmaq.
Thank you.
Ho-Chun
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Mon Apr 22 11:25:32 2019
Ho-Chun,
I see you have a question about using METviewer to limit the data
going
into a plot to 1 month. Thanks for pointing me to your example XML.
The answer is yes, of course. You can control what is held constant
on the
plot using the "Fixed Values" section. Your plot already has
"FCST_LEAD =
24" included in that spot. So I just added a new Fixed Values entry
for
"FCST_INIT_BEG" to specify which model initialization times should be
included. On the right side, there's a little both with check mark in
it
which is the "date picker". I clicked on that and selected "Start =
2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
hours"...
and then clicked "Select Values". And that picked all the
initializations
for Aug 2018.
Granted, you have a long list of dates... so an easier time picker
would be
nice.
I updated the title and then remade the plot. If you look at the
history,
you'll find it dated as XML: 20190422_172306.
Hopefully that helps.
Thanks,
John
On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> Transaction: Ticket created by ho-chun.huang at noaa.gov
> Queue: met_help
> Subject: Metviewer CSI plot without date range selection
> Owner: Nobody
> Requestors: ho-chun.huang at noaa.gov
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
>
> To Whom It May Concern:
>
> Case XML : 20190422_151527
>
> I want to compare CSI score for August 2018 between different AQM
model
> output. But I recall previously I can select the time range for
> comparison, e.g., AUG 2018 versus SEP 2018. But this time I have no
option
> for selecting the time range.
>
> My Question is : Can CSI be computed for a specific time range? or
it just
> use the wntire data in the database? IS it a problem in pareparing
the
> database ? i.e., mv_emc_cmaq.
>
> Thank you.
>
> Ho-Chun
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Tatiana Burek
Time: Mon Apr 22 11:29:00 2019
Hi Ho-Chun,
Yes, you can specify the time range for the each series.
Please look at plot # 20190422_171127
I added 2 series variable (both fcst_valid_beg). One is for August
2018 and another for September 2018. I apply the grouping to both. I
end up with 6 series. Each series represents single model/single
period. Is it what you try to accomplish?
Tatiana
On Mon Apr 22 09:25:34 2019, ho-chun.huang at noaa.gov wrote:
> To Whom It May Concern:
>
> Case XML : 20190422_151527
>
> I want to compare CSI score for August 2018 between different AQM
model
> output. But I recall previously I can select the time range for
> comparison, e.g., AUG 2018 versus SEP 2018. But this time I have no
option
> for selecting the time range.
>
> My Question is : Can CSI be computed for a specific time range? or
it just
> use the wntire data in the database? IS it a problem in pareparing
the
> database ? i.e., mv_emc_cmaq.
>
> Thank you.
>
> Ho-Chun
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon Apr 22 12:29:27 2019
John,
Thanks. Since you are on duty today, I am going to take advantage of
it.
Say, I want to know the performance of the day1 (FCST 1-24) and Day 2
(FCST 25-48), i.e., verify each hours for day1 and day2 period and
produce
a daily mean CSI. Then average over Aug 2018 for monthly CSI score.
How
should I do it?
Does 20190422_182543 plot what I need?
Ho-Chun
On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> Ho-Chun,
>
> I see you have a question about using METviewer to limit the data
going
> into a plot to 1 month. Thanks for pointing me to your example XML.
>
> The answer is yes, of course. You can control what is held constant
on the
> plot using the "Fixed Values" section. Your plot already has
"FCST_LEAD =
> 24" included in that spot. So I just added a new Fixed Values entry
for
> "FCST_INIT_BEG" to specify which model initialization times should
be
> included. On the right side, there's a little both with check mark
in it
> which is the "date picker". I clicked on that and selected "Start =
> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
hours"...
> and then clicked "Select Values". And that picked all the
initializations
> for Aug 2018.
>
> Granted, you have a long list of dates... so an easier time picker
would be
> nice.
>
> I updated the title and then remade the plot. If you look at the
history,
> you'll find it dated as XML: 20190422_172306.
>
> Hopefully that helps.
>
> Thanks,
> John
>
> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> > Queue: met_help
> > Subject: Metviewer CSI plot without date range selection
> > Owner: Nobody
> > Requestors: ho-chun.huang at noaa.gov
> > Status: new
> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> >
> > To Whom It May Concern:
> >
> > Case XML : 20190422_151527
> >
> > I want to compare CSI score for August 2018 between different AQM
model
> > output. But I recall previously I can select the time range for
> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I have
no
> option
> > for selecting the time range.
> >
> > My Question is : Can CSI be computed for a specific time range?
or it
> just
> > use the wntire data in the database? IS it a problem in
pareparing the
> > database ? i.e., mv_emc_cmaq.
> >
> > Thank you.
> >
> > Ho-Chun
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon Apr 22 13:50:55 2019
Hi, Please verify whether plot # 20190422_194456
NOTE there is score for 06Z and 12Z AQM runs, and 48 hours FCST.
(1) Model 5x, 82, 8H
(2) 12 Z run from 20180801 to 20180831
(3) average CSI for DAY 2 forecast (hourly)
If I change FCST LEAD to 1, then it is DAY one mean CSI score,
correct?
Thank you for your help.
Tatiana, thanks for showing me the "gourp" method. I will assume your
approach is the mean score of both 06z and 12Z runs.
Ho-Chun
On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate <
ho-chun.huang at noaa.gov> wrote:
> John,
>
> Thanks. Since you are on duty today, I am going to take advantage
of it.
>
> Say, I want to know the performance of the day1 (FCST 1-24) and Day
2
> (FCST 25-48), i.e., verify each hours for day1 and day2 period and
produce
> a daily mean CSI. Then average over Aug 2018 for monthly CSI
score. How
> should I do it?
>
> Does 20190422_182543 plot what I need?
>
> Ho-Chun
>
> On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
>> Ho-Chun,
>>
>> I see you have a question about using METviewer to limit the data
going
>> into a plot to 1 month. Thanks for pointing me to your example
XML.
>>
>> The answer is yes, of course. You can control what is held
constant on
>> the
>> plot using the "Fixed Values" section. Your plot already has
"FCST_LEAD =
>> 24" included in that spot. So I just added a new Fixed Values
entry for
>> "FCST_INIT_BEG" to specify which model initialization times should
be
>> included. On the right side, there's a little both with check mark
in it
>> which is the "date picker". I clicked on that and selected "Start
=
>> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
hours"...
>> and then clicked "Select Values". And that picked all the
initializations
>> for Aug 2018.
>>
>> Granted, you have a long list of dates... so an easier time picker
would
>> be
>> nice.
>>
>> I updated the title and then remade the plot. If you look at the
history,
>> you'll find it dated as XML: 20190422_172306.
>>
>> Hopefully that helps.
>>
>> Thanks,
>> John
>>
>> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate via
RT <
>> met_help at ucar.edu> wrote:
>>
>> >
>> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
>> > Transaction: Ticket created by ho-chun.huang at noaa.gov
>> > Queue: met_help
>> > Subject: Metviewer CSI plot without date range selection
>> > Owner: Nobody
>> > Requestors: ho-chun.huang at noaa.gov
>> > Status: new
>> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>> >
>> >
>> > To Whom It May Concern:
>> >
>> > Case XML : 20190422_151527
>> >
>> > I want to compare CSI score for August 2018 between different AQM
model
>> > output. But I recall previously I can select the time range for
>> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I have
no
>> option
>> > for selecting the time range.
>> >
>> > My Question is : Can CSI be computed for a specific time range?
or it
>> just
>> > use the wntire data in the database? IS it a problem in
pareparing the
>> > database ? i.e., mv_emc_cmaq.
>> >
>> > Thank you.
>> >
>> > Ho-Chun
>> >
>> >
>>
>>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Wed Apr 24 06:33:26 2019
Hi,
I got the response from John and Tatiana, thanks.
I still have question on John's XML: 20190422_17230.
Mostly, I am not sure how to use the combination of FCST_LEAD (24) and
FCST_INIT_BEG (06Z and 12Z runs)
Does it mean the statistic includes all evaluation that are validated
at
06Z and 12Z of each day? Also, what is the meaning of FCST_LEAD = 1,
evaluation of all FCST hours? or only at 07Z and 13Z of each day
(combine
usage of FCST_INIT_BEGabove)?
Ho-Chun
On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate <
ho-chun.huang at noaa.gov> wrote:
> John,
>
> Thanks. Since you are on duty today, I am going to take advantage
of it.
>
> Say, I want to know the performance of the day1 (FCST 1-24) and Day
2
> (FCST 25-48), i.e., verify each hours for day1 and day2 period and
produce
> a daily mean CSI. Then average over Aug 2018 for monthly CSI
score. How
> should I do it?
>
> Does 20190422_182543 plot what I need?
>
> Ho-Chun
>
> On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
>> Ho-Chun,
>>
>> I see you have a question about using METviewer to limit the data
going
>> into a plot to 1 month. Thanks for pointing me to your example
XML.
>>
>> The answer is yes, of course. You can control what is held
constant on
>> the
>> plot using the "Fixed Values" section. Your plot already has
"FCST_LEAD =
>> 24" included in that spot. So I just added a new Fixed Values
entry for
>> "FCST_INIT_BEG" to specify which model initialization times should
be
>> included. On the right side, there's a little both with check mark
in it
>> which is the "date picker". I clicked on that and selected "Start
=
>> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
hours"...
>> and then clicked "Select Values". And that picked all the
initializations
>> for Aug 2018.
>>
>> Granted, you have a long list of dates... so an easier time picker
would
>> be
>> nice.
>>
>> I updated the title and then remade the plot. If you look at the
history,
>> you'll find it dated as XML: 20190422_172306.
>>
>> Hopefully that helps.
>>
>> Thanks,
>> John
>>
>> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate via
RT <
>> met_help at ucar.edu> wrote:
>>
>> >
>> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
>> > Transaction: Ticket created by ho-chun.huang at noaa.gov
>> > Queue: met_help
>> > Subject: Metviewer CSI plot without date range selection
>> > Owner: Nobody
>> > Requestors: ho-chun.huang at noaa.gov
>> > Status: new
>> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>> >
>> >
>> > To Whom It May Concern:
>> >
>> > Case XML : 20190422_151527
>> >
>> > I want to compare CSI score for August 2018 between different AQM
model
>> > output. But I recall previously I can select the time range for
>> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I have
no
>> option
>> > for selecting the time range.
>> >
>> > My Question is : Can CSI be computed for a specific time range?
or it
>> just
>> > use the wntire data in the database? IS it a problem in
pareparing the
>> > database ? i.e., mv_emc_cmaq.
>> >
>> > Thank you.
>> >
>> > Ho-Chun
>> >
>> >
>>
>>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Wed Apr 24 12:04:14 2019
Ho Chun,
Sorry for the delay.
So you had written about plot # 20190422_194456, and asked...
If I change FCST LEAD to 1, then it is DAY one mean CSI score,
correct?
I believe the answer is no. When I look at the available FCST_LEAD
values
for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48, and then
a
jump to 54.
I am assuming these FCST_LEAD values are in hours. So a day ONE would
either be just FCST_LEAD = 24... or maybe you'd prefer to define day
one as
all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
I also see that you're setting VALID_HOUR = 00, 01, 02... 23 in the
"Fixed
Values" section. This really isn't necessary since that's the full
set of
valid hours that are possible. So specifying this will just make the
query
a little bit slower but have no impact on the result.
I see that you have two daily initializations for this data, at 06 and
12
each day. Unless you tell METviewer to plot them separately, they
will be
combined.
There's just so many ways you can slice/dice the data, it isn't clear
to me
exactly what you're looking for.
I see that you're interested in the 3 models you listed. And you want
to
plot the CSI statistic versus the PM2.5 threshold. But how many lines
do
you want on the plot? And how do you want to define the timing info
that
goes into each? What lead times should be used (just 24 hours or 01-
24
hours)? What initialization hours do you want to include? And it
looks
like you want to limit this to August 2018. Is that correct?
Once we know exactly what you want we can help you with the METviewer
configuration.
Or if it's easier, we could do a telecon and work on this together.
Thanks,
John
On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
> Hi,
>
> I got the response from John and Tatiana, thanks.
>
> I still have question on John's XML: 20190422_17230.
>
> Mostly, I am not sure how to use the combination of FCST_LEAD (24)
and
> FCST_INIT_BEG (06Z and 12Z runs)
>
> Does it mean the statistic includes all evaluation that are
validated at
> 06Z and 12Z of each day? Also, what is the meaning of FCST_LEAD =
1,
> evaluation of all FCST hours? or only at 07Z and 13Z of each day
(combine
> usage of FCST_INIT_BEGabove)?
>
> Ho-Chun
>
> On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate <
> ho-chun.huang at noaa.gov> wrote:
>
> > John,
> >
> > Thanks. Since you are on duty today, I am going to take advantage
of it.
> >
> > Say, I want to know the performance of the day1 (FCST 1-24) and
Day 2
> > (FCST 25-48), i.e., verify each hours for day1 and day2 period and
> produce
> > a daily mean CSI. Then average over Aug 2018 for monthly CSI
score.
> How
> > should I do it?
> >
> > Does 20190422_182543 plot what I need?
> >
> > Ho-Chun
> >
> > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> >> Ho-Chun,
> >>
> >> I see you have a question about using METviewer to limit the data
going
> >> into a plot to 1 month. Thanks for pointing me to your example
XML.
> >>
> >> The answer is yes, of course. You can control what is held
constant on
> >> the
> >> plot using the "Fixed Values" section. Your plot already has
> "FCST_LEAD =
> >> 24" included in that spot. So I just added a new Fixed Values
entry for
> >> "FCST_INIT_BEG" to specify which model initialization times
should be
> >> included. On the right side, there's a little both with check
mark in
> it
> >> which is the "date picker". I clicked on that and selected
"Start =
> >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
> hours"...
> >> and then clicked "Select Values". And that picked all the
> initializations
> >> for Aug 2018.
> >>
> >> Granted, you have a long list of dates... so an easier time
picker would
> >> be
> >> nice.
> >>
> >> I updated the title and then remade the plot. If you look at the
> history,
> >> you'll find it dated as XML: 20190422_172306.
> >>
> >> Hopefully that helps.
> >>
> >> Thanks,
> >> John
> >>
> >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate
via RT <
> >> met_help at ucar.edu> wrote:
> >>
> >> >
> >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> >> > Queue: met_help
> >> > Subject: Metviewer CSI plot without date range selection
> >> > Owner: Nobody
> >> > Requestors: ho-chun.huang at noaa.gov
> >> > Status: new
> >> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> >
> >> >
> >> >
> >> > To Whom It May Concern:
> >> >
> >> > Case XML : 20190422_151527
> >> >
> >> > I want to compare CSI score for August 2018 between different
AQM
> model
> >> > output. But I recall previously I can select the time range
for
> >> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I
have no
> >> option
> >> > for selecting the time range.
> >> >
> >> > My Question is : Can CSI be computed for a specific time range?
or it
> >> just
> >> > use the wntire data in the database? IS it a problem in
pareparing
> the
> >> > database ? i.e., mv_emc_cmaq.
> >> >
> >> > Thank you.
> >> >
> >> > Ho-Chun
> >> >
> >> >
> >>
> >>
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Wed Apr 24 12:28:49 2019
Hi, John:
>From your email I think I should remove the valid hour list entirely.
I
simply has trouble to imagine the relationship between FCST_INI_BEG
and
FCST_LEAD.I do not know if it only evaluated
at one hour = FCST hr 24
or
Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours 25-48)
performance.
I really do not know why there is 54 (only Perry will know).
What I want is to compare hourly from FCST hr 1 to FCST hr 24 between
model
and observed and has a combined one CSI score for Day 1 performance
(statistic include all 24 hours). Then again for Day 2 CSI score
(from
FCST 25-FCST 48) in separate plot. [+ selected VX_MASK + ether 06z or
12Z
run].
If you look at plot_20190424_1340030, what is your interpretation of
the
intention of the plot?
I actually want to evaluate DAY2 PM25 CSI score for 12Z runs, in Aug
2018,
for 3 model, for NWC, Does it match with your understanding ?
Ho-Chun
On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> Ho Chun,
>
> Sorry for the delay.
>
> So you had written about plot # 20190422_194456, and asked...
> If I change FCST LEAD to 1, then it is DAY one mean CSI score,
correct?
>
> I believe the answer is no. When I look at the available FCST_LEAD
values
> for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48, and
then a
> jump to 54.
>
> I am assuming these FCST_LEAD values are in hours. So a day ONE
would
> either be just FCST_LEAD = 24... or maybe you'd prefer to define day
one as
> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
>
> I also see that you're setting VALID_HOUR = 00, 01, 02... 23 in the
"Fixed
> Values" section. This really isn't necessary since that's the full
set of
> valid hours that are possible. So specifying this will just make
the query
> a little bit slower but have no impact on the result.
>
> I see that you have two daily initializations for this data, at 06
and 12
> each day. Unless you tell METviewer to plot them separately, they
will be
> combined.
>
> There's just so many ways you can slice/dice the data, it isn't
clear to me
> exactly what you're looking for.
>
> I see that you're interested in the 3 models you listed. And you
want to
> plot the CSI statistic versus the PM2.5 threshold. But how many
lines do
> you want on the plot? And how do you want to define the timing info
that
> goes into each? What lead times should be used (just 24 hours or
01-24
> hours)? What initialization hours do you want to include? And it
looks
> like you want to limit this to August 2018. Is that correct?
>
> Once we know exactly what you want we can help you with the
METviewer
> configuration.
>
> Or if it's easier, we could do a telecon and work on this together.
>
> Thanks,
> John
>
> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> > Hi,
> >
> > I got the response from John and Tatiana, thanks.
> >
> > I still have question on John's XML: 20190422_17230.
> >
> > Mostly, I am not sure how to use the combination of FCST_LEAD (24)
and
> > FCST_INIT_BEG (06Z and 12Z runs)
> >
> > Does it mean the statistic includes all evaluation that are
validated at
> > 06Z and 12Z of each day? Also, what is the meaning of FCST_LEAD =
1,
> > evaluation of all FCST hours? or only at 07Z and 13Z of each day
(combine
> > usage of FCST_INIT_BEGabove)?
> >
> > Ho-Chun
> >
> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate <
> > ho-chun.huang at noaa.gov> wrote:
> >
> > > John,
> > >
> > > Thanks. Since you are on duty today, I am going to take
advantage of
> it.
> > >
> > > Say, I want to know the performance of the day1 (FCST 1-24) and
Day 2
> > > (FCST 25-48), i.e., verify each hours for day1 and day2 period
and
> > produce
> > > a daily mean CSI. Then average over Aug 2018 for monthly CSI
score.
> > How
> > > should I do it?
> > >
> > > Does 20190422_182543 plot what I need?
> > >
> > > Ho-Chun
> > >
> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >> Ho-Chun,
> > >>
> > >> I see you have a question about using METviewer to limit the
data
> going
> > >> into a plot to 1 month. Thanks for pointing me to your example
XML.
> > >>
> > >> The answer is yes, of course. You can control what is held
constant
> on
> > >> the
> > >> plot using the "Fixed Values" section. Your plot already has
> > "FCST_LEAD =
> > >> 24" included in that spot. So I just added a new Fixed Values
entry
> for
> > >> "FCST_INIT_BEG" to specify which model initialization times
should be
> > >> included. On the right side, there's a little both with check
mark in
> > it
> > >> which is the "date picker". I clicked on that and selected
"Start =
> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By: 6
> > hours"...
> > >> and then clicked "Select Values". And that picked all the
> > initializations
> > >> for Aug 2018.
> > >>
> > >> Granted, you have a long list of dates... so an easier time
picker
> would
> > >> be
> > >> nice.
> > >>
> > >> I updated the title and then remade the plot. If you look at
the
> > history,
> > >> you'll find it dated as XML: 20190422_172306.
> > >>
> > >> Hopefully that helps.
> > >>
> > >> Thanks,
> > >> John
> > >>
> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate
via RT
> <
> > >> met_help at ucar.edu> wrote:
> > >>
> > >> >
> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> > >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> > >> > Queue: met_help
> > >> > Subject: Metviewer CSI plot without date range selection
> > >> > Owner: Nobody
> > >> > Requestors: ho-chun.huang at noaa.gov
> > >> > Status: new
> > >> > Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > >
> > >> >
> > >> >
> > >> > To Whom It May Concern:
> > >> >
> > >> > Case XML : 20190422_151527
> > >> >
> > >> > I want to compare CSI score for August 2018 between different
AQM
> > model
> > >> > output. But I recall previously I can select the time range
for
> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I
have no
> > >> option
> > >> > for selecting the time range.
> > >> >
> > >> > My Question is : Can CSI be computed for a specific time
range? or
> it
> > >> just
> > >> > use the wntire data in the database? IS it a problem in
pareparing
> > the
> > >> > database ? i.e., mv_emc_cmaq.
> > >> >
> > >> > Thank you.
> > >> >
> > >> > Ho-Chun
> > >> >
> > >> >
> > >>
> > >>
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Wed Apr 24 12:30:20 2019
Hi, John:
Yes, I can set up a teleconference if I still have confusion on the
data
included.
Ho-Chun
On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA Affiliate <
ho-chun.huang at noaa.gov> wrote:
> Hi, John:
>
> From your email I think I should remove the valid hour list
entirely. I
> simply has trouble to imagine the relationship between FCST_INI_BEG
and
> FCST_LEAD.I do not know if it only evaluated
>
> at one hour = FCST hr 24
> or
> Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours 25-48)
> performance.
>
> I really do not know why there is 54 (only Perry will know).
>
> What I want is to compare hourly from FCST hr 1 to FCST hr 24
between
> model and observed and has a combined one CSI score for Day 1
performance
> (statistic include all 24 hours). Then again for Day 2 CSI score
(from
> FCST 25-FCST 48) in separate plot. [+ selected VX_MASK + ether 06z
or 12Z
> run].
>
> If you look at plot_20190424_1340030, what is your interpretation of
the
> intention of the plot?
>
> I actually want to evaluate DAY2 PM25 CSI score for 12Z runs, in Aug
2018,
> for 3 model, for NWC, Does it match with your understanding ?
>
> Ho-Chun
>
> On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
>> Ho Chun,
>>
>> Sorry for the delay.
>>
>> So you had written about plot # 20190422_194456, and asked...
>> If I change FCST LEAD to 1, then it is DAY one mean CSI score,
correct?
>>
>> I believe the answer is no. When I look at the available FCST_LEAD
values
>> for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48, and
then a
>> jump to 54.
>>
>> I am assuming these FCST_LEAD values are in hours. So a day ONE
would
>> either be just FCST_LEAD = 24... or maybe you'd prefer to define
day one
>> as
>> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
>>
>> I also see that you're setting VALID_HOUR = 00, 01, 02... 23 in the
"Fixed
>> Values" section. This really isn't necessary since that's the full
set of
>> valid hours that are possible. So specifying this will just make
the
>> query
>> a little bit slower but have no impact on the result.
>>
>> I see that you have two daily initializations for this data, at 06
and 12
>> each day. Unless you tell METviewer to plot them separately, they
will be
>> combined.
>>
>> There's just so many ways you can slice/dice the data, it isn't
clear to
>> me
>> exactly what you're looking for.
>>
>> I see that you're interested in the 3 models you listed. And you
want to
>> plot the CSI statistic versus the PM2.5 threshold. But how many
lines do
>> you want on the plot? And how do you want to define the timing
info that
>> goes into each? What lead times should be used (just 24 hours or
01-24
>> hours)? What initialization hours do you want to include? And it
looks
>> like you want to limit this to August 2018. Is that correct?
>>
>> Once we know exactly what you want we can help you with the
METviewer
>> configuration.
>>
>> Or if it's easier, we could do a telecon and work on this together.
>>
>> Thanks,
>> John
>>
>> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA Affiliate via
RT <
>> met_help at ucar.edu> wrote:
>>
>> >
>> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>> >
>> > Hi,
>> >
>> > I got the response from John and Tatiana, thanks.
>> >
>> > I still have question on John's XML: 20190422_17230.
>> >
>> > Mostly, I am not sure how to use the combination of FCST_LEAD
(24) and
>> > FCST_INIT_BEG (06Z and 12Z runs)
>> >
>> > Does it mean the statistic includes all evaluation that are
validated at
>> > 06Z and 12Z of each day? Also, what is the meaning of FCST_LEAD
= 1,
>> > evaluation of all FCST hours? or only at 07Z and 13Z of each day
>> (combine
>> > usage of FCST_INIT_BEGabove)?
>> >
>> > Ho-Chun
>> >
>> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate <
>> > ho-chun.huang at noaa.gov> wrote:
>> >
>> > > John,
>> > >
>> > > Thanks. Since you are on duty today, I am going to take
advantage of
>> it.
>> > >
>> > > Say, I want to know the performance of the day1 (FCST 1-24)
and Day 2
>> > > (FCST 25-48), i.e., verify each hours for day1 and day2 period
and
>> > produce
>> > > a daily mean CSI. Then average over Aug 2018 for monthly CSI
score.
>> > How
>> > > should I do it?
>> > >
>> > > Does 20190422_182543 plot what I need?
>> > >
>> > > Ho-Chun
>> > >
>> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
>> > > met_help at ucar.edu> wrote:
>> > >
>> > >> Ho-Chun,
>> > >>
>> > >> I see you have a question about using METviewer to limit the
data
>> going
>> > >> into a plot to 1 month. Thanks for pointing me to your
example XML.
>> > >>
>> > >> The answer is yes, of course. You can control what is held
constant
>> on
>> > >> the
>> > >> plot using the "Fixed Values" section. Your plot already has
>> > "FCST_LEAD =
>> > >> 24" included in that spot. So I just added a new Fixed Values
entry
>> for
>> > >> "FCST_INIT_BEG" to specify which model initialization times
should be
>> > >> included. On the right side, there's a little both with check
mark
>> in
>> > it
>> > >> which is the "date picker". I clicked on that and selected
"Start =
>> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and "By:
6
>> > hours"...
>> > >> and then clicked "Select Values". And that picked all the
>> > initializations
>> > >> for Aug 2018.
>> > >>
>> > >> Granted, you have a long list of dates... so an easier time
picker
>> would
>> > >> be
>> > >> nice.
>> > >>
>> > >> I updated the title and then remade the plot. If you look at
the
>> > history,
>> > >> you'll find it dated as XML: 20190422_172306.
>> > >>
>> > >> Hopefully that helps.
>> > >>
>> > >> Thanks,
>> > >> John
>> > >>
>> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA Affiliate
via
>> RT <
>> > >> met_help at ucar.edu> wrote:
>> > >>
>> > >> >
>> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
>> > >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
>> > >> > Queue: met_help
>> > >> > Subject: Metviewer CSI plot without date range
selection
>> > >> > Owner: Nobody
>> > >> > Requestors: ho-chun.huang at noaa.gov
>> > >> > Status: new
>> > >> > Ticket <URL:
>> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
>> > >
>> > >> >
>> > >> >
>> > >> > To Whom It May Concern:
>> > >> >
>> > >> > Case XML : 20190422_151527
>> > >> >
>> > >> > I want to compare CSI score for August 2018 between
different AQM
>> > model
>> > >> > output. But I recall previously I can select the time range
for
>> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this time I
have no
>> > >> option
>> > >> > for selecting the time range.
>> > >> >
>> > >> > My Question is : Can CSI be computed for a specific time
range?
>> or it
>> > >> just
>> > >> > use the wntire data in the database? IS it a problem in
pareparing
>> > the
>> > >> > database ? i.e., mv_emc_cmaq.
>> > >> >
>> > >> > Thank you.
>> > >> >
>> > >> > Ho-Chun
>> > >> >
>> > >> >
>> > >>
>> > >>
>> >
>> >
>>
>>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Thu Apr 25 14:38:00 2019
Ho Chun,
OK, I think I understand what you want.
You want 3 lines on the plot... for 3 different models. You want to
limit
the analysis to August, 2018. And you want separate plots for each
combination of:
- VX_MASK region (NWC, plus other ones too probably)
- Initialization hour (06 or 12).
- Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
I just made an example plot on METviewer for:
- month = aug 2018
- vx_mask = NWC
- init_hour = 06
- forecast = day 1
And I named it *cmaq_aug2018_NWC_init06_day1*. So if you look in the
history, you'll find it there. In METviewer 2.9, you have the option
of
giving a plot a name using the "Job title" box.
But this is just one permutation of what you want. If you only want a
couple of ones, you could easily go make them by hand using the GUI.
But
you could also use the batch engine to automate this. Perhaps you
want
separate plots for:
- 12 months of the yera
- 2 initialization hours (06 and 12)
- 2 forecast lead time groups (day 1 and day 2)
- 14 subregions
That would be 672 plots... way too many to make by hand. But we could
do
them all using a single XML file through the METviewer batch engine.
Make sense?
Thanks,
John
On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
> Hi, John:
>
> Yes, I can set up a teleconference if I still have confusion on the
data
> included.
>
> Ho-Chun
>
> On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA Affiliate <
> ho-chun.huang at noaa.gov> wrote:
>
> > Hi, John:
> >
> > From your email I think I should remove the valid hour list
entirely. I
> > simply has trouble to imagine the relationship between
FCST_INI_BEG and
> > FCST_LEAD.I do not know if it only evaluated
> >
> > at one hour = FCST hr 24
> > or
> > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours 25-
48)
> > performance.
> >
> > I really do not know why there is 54 (only Perry will know).
> >
> > What I want is to compare hourly from FCST hr 1 to FCST hr 24
between
> > model and observed and has a combined one CSI score for Day 1
performance
> > (statistic include all 24 hours). Then again for Day 2 CSI score
(from
> > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK + ether
06z or
> 12Z
> > run].
> >
> > If you look at plot_20190424_1340030, what is your interpretation
of the
> > intention of the plot?
> >
> > I actually want to evaluate DAY2 PM25 CSI score for 12Z runs, in
Aug
> 2018,
> > for 3 model, for NWC, Does it match with your understanding ?
> >
> > Ho-Chun
> >
> > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> >> Ho Chun,
> >>
> >> Sorry for the delay.
> >>
> >> So you had written about plot # 20190422_194456, and asked...
> >> If I change FCST LEAD to 1, then it is DAY one mean CSI score,
> correct?
> >>
> >> I believe the answer is no. When I look at the available
FCST_LEAD
> values
> >> for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48, and
then a
> >> jump to 54.
> >>
> >> I am assuming these FCST_LEAD values are in hours. So a day ONE
would
> >> either be just FCST_LEAD = 24... or maybe you'd prefer to define
day one
> >> as
> >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> >>
> >> I also see that you're setting VALID_HOUR = 00, 01, 02... 23 in
the
> "Fixed
> >> Values" section. This really isn't necessary since that's the
full set
> of
> >> valid hours that are possible. So specifying this will just make
the
> >> query
> >> a little bit slower but have no impact on the result.
> >>
> >> I see that you have two daily initializations for this data, at
06 and
> 12
> >> each day. Unless you tell METviewer to plot them separately,
they will
> be
> >> combined.
> >>
> >> There's just so many ways you can slice/dice the data, it isn't
clear to
> >> me
> >> exactly what you're looking for.
> >>
> >> I see that you're interested in the 3 models you listed. And you
want
> to
> >> plot the CSI statistic versus the PM2.5 threshold. But how many
lines
> do
> >> you want on the plot? And how do you want to define the timing
info
> that
> >> goes into each? What lead times should be used (just 24 hours or
01-24
> >> hours)? What initialization hours do you want to include? And
it looks
> >> like you want to limit this to August 2018. Is that correct?
> >>
> >> Once we know exactly what you want we can help you with the
METviewer
> >> configuration.
> >>
> >> Or if it's easier, we could do a telecon and work on this
together.
> >>
> >> Thanks,
> >> John
> >>
> >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA Affiliate
via RT <
> >> met_help at ucar.edu> wrote:
> >>
> >> >
> >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >> >
> >> > Hi,
> >> >
> >> > I got the response from John and Tatiana, thanks.
> >> >
> >> > I still have question on John's XML: 20190422_17230.
> >> >
> >> > Mostly, I am not sure how to use the combination of FCST_LEAD
(24) and
> >> > FCST_INIT_BEG (06Z and 12Z runs)
> >> >
> >> > Does it mean the statistic includes all evaluation that are
validated
> at
> >> > 06Z and 12Z of each day? Also, what is the meaning of
FCST_LEAD = 1,
> >> > evaluation of all FCST hours? or only at 07Z and 13Z of each
day
> >> (combine
> >> > usage of FCST_INIT_BEGabove)?
> >> >
> >> > Ho-Chun
> >> >
> >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA Affiliate
<
> >> > ho-chun.huang at noaa.gov> wrote:
> >> >
> >> > > John,
> >> > >
> >> > > Thanks. Since you are on duty today, I am going to take
advantage
> of
> >> it.
> >> > >
> >> > > Say, I want to know the performance of the day1 (FCST 1-24)
and
> Day 2
> >> > > (FCST 25-48), i.e., verify each hours for day1 and day2
period and
> >> > produce
> >> > > a daily mean CSI. Then average over Aug 2018 for monthly
CSI
> score.
> >> > How
> >> > > should I do it?
> >> > >
> >> > > Does 20190422_182543 plot what I need?
> >> > >
> >> > > Ho-Chun
> >> > >
> >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> >> > > met_help at ucar.edu> wrote:
> >> > >
> >> > >> Ho-Chun,
> >> > >>
> >> > >> I see you have a question about using METviewer to limit the
data
> >> going
> >> > >> into a plot to 1 month. Thanks for pointing me to your
example
> XML.
> >> > >>
> >> > >> The answer is yes, of course. You can control what is held
> constant
> >> on
> >> > >> the
> >> > >> plot using the "Fixed Values" section. Your plot already
has
> >> > "FCST_LEAD =
> >> > >> 24" included in that spot. So I just added a new Fixed
Values
> entry
> >> for
> >> > >> "FCST_INIT_BEG" to specify which model initialization times
should
> be
> >> > >> included. On the right side, there's a little both with
check mark
> >> in
> >> > it
> >> > >> which is the "date picker". I clicked on that and selected
"Start
> =
> >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and
"By: 6
> >> > hours"...
> >> > >> and then clicked "Select Values". And that picked all the
> >> > initializations
> >> > >> for Aug 2018.
> >> > >>
> >> > >> Granted, you have a long list of dates... so an easier time
picker
> >> would
> >> > >> be
> >> > >> nice.
> >> > >>
> >> > >> I updated the title and then remade the plot. If you look
at the
> >> > history,
> >> > >> you'll find it dated as XML: 20190422_172306.
> >> > >>
> >> > >> Hopefully that helps.
> >> > >>
> >> > >> Thanks,
> >> > >> John
> >> > >>
> >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA
Affiliate via
> >> RT <
> >> > >> met_help at ucar.edu> wrote:
> >> > >>
> >> > >> >
> >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> >> > >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> >> > >> > Queue: met_help
> >> > >> > Subject: Metviewer CSI plot without date range
selection
> >> > >> > Owner: Nobody
> >> > >> > Requestors: ho-chun.huang at noaa.gov
> >> > >> > Status: new
> >> > >> > Ticket <URL:
> >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> >> > >
> >> > >> >
> >> > >> >
> >> > >> > To Whom It May Concern:
> >> > >> >
> >> > >> > Case XML : 20190422_151527
> >> > >> >
> >> > >> > I want to compare CSI score for August 2018 between
different AQM
> >> > model
> >> > >> > output. But I recall previously I can select the time
range for
> >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this time
I have
> no
> >> > >> option
> >> > >> > for selecting the time range.
> >> > >> >
> >> > >> > My Question is : Can CSI be computed for a specific time
range?
> >> or it
> >> > >> just
> >> > >> > use the wntire data in the database? IS it a problem in
> pareparing
> >> > the
> >> > >> > database ? i.e., mv_emc_cmaq.
> >> > >> >
> >> > >> > Thank you.
> >> > >> >
> >> > >> > Ho-Chun
> >> > >> >
> >> > >> >
> >> > >>
> >> > >>
> >> >
> >> >
> >>
> >>
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon Apr 29 06:49:34 2019
Hi, John:
Thanks for the help. My desktop dead this morning and I need to wait
for
the new machine to get back into the Metviewer to try. All my save
xml are
on the dead machine.
Will get back to you and try again on your setting. Yes, the next
thing is
figuring out how to use the batch machine. Perry expressed maybe to
wait
till Amazon Cloud is available to us.
Ho-Chun
On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> Ho Chun,
>
> OK, I think I understand what you want.
>
> You want 3 lines on the plot... for 3 different models. You want to
limit
> the analysis to August, 2018. And you want separate plots for each
> combination of:
> - VX_MASK region (NWC, plus other ones too probably)
> - Initialization hour (06 or 12).
> - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
>
> I just made an example plot on METviewer for:
> - month = aug 2018
> - vx_mask = NWC
> - init_hour = 06
> - forecast = day 1
>
> And I named it *cmaq_aug2018_NWC_init06_day1*. So if you look in
the
> history, you'll find it there. In METviewer 2.9, you have the
option of
> giving a plot a name using the "Job title" box.
>
> But this is just one permutation of what you want. If you only want
a
> couple of ones, you could easily go make them by hand using the GUI.
But
> you could also use the batch engine to automate this. Perhaps you
want
> separate plots for:
> - 12 months of the yera
> - 2 initialization hours (06 and 12)
> - 2 forecast lead time groups (day 1 and day 2)
> - 14 subregions
>
> That would be 672 plots... way too many to make by hand. But we
could do
> them all using a single XML file through the METviewer batch engine.
>
> Make sense?
>
> Thanks,
> John
>
> On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> > Hi, John:
> >
> > Yes, I can set up a teleconference if I still have confusion on
the data
> > included.
> >
> > Ho-Chun
> >
> > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA Affiliate <
> > ho-chun.huang at noaa.gov> wrote:
> >
> > > Hi, John:
> > >
> > > From your email I think I should remove the valid hour list
entirely.
> I
> > > simply has trouble to imagine the relationship between
FCST_INI_BEG and
> > > FCST_LEAD.I do not know if it only evaluated
> > >
> > > at one hour = FCST hr 24
> > > or
> > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours 25-
48)
> > > performance.
> > >
> > > I really do not know why there is 54 (only Perry will know).
> > >
> > > What I want is to compare hourly from FCST hr 1 to FCST hr 24
between
> > > model and observed and has a combined one CSI score for Day 1
> performance
> > > (statistic include all 24 hours). Then again for Day 2 CSI
score (from
> > > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK + ether
06z or
> > 12Z
> > > run].
> > >
> > > If you look at plot_20190424_1340030, what is your
interpretation of
> the
> > > intention of the plot?
> > >
> > > I actually want to evaluate DAY2 PM25 CSI score for 12Z runs, in
Aug
> > 2018,
> > > for 3 model, for NWC, Does it match with your understanding ?
> > >
> > > Ho-Chun
> > >
> > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >> Ho Chun,
> > >>
> > >> Sorry for the delay.
> > >>
> > >> So you had written about plot # 20190422_194456, and asked...
> > >> If I change FCST LEAD to 1, then it is DAY one mean CSI
score,
> > correct?
> > >>
> > >> I believe the answer is no. When I look at the available
FCST_LEAD
> > values
> > >> for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48,
and
> then a
> > >> jump to 54.
> > >>
> > >> I am assuming these FCST_LEAD values are in hours. So a day
ONE would
> > >> either be just FCST_LEAD = 24... or maybe you'd prefer to
define day
> one
> > >> as
> > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> > >>
> > >> I also see that you're setting VALID_HOUR = 00, 01, 02... 23 in
the
> > "Fixed
> > >> Values" section. This really isn't necessary since that's the
full
> set
> > of
> > >> valid hours that are possible. So specifying this will just
make the
> > >> query
> > >> a little bit slower but have no impact on the result.
> > >>
> > >> I see that you have two daily initializations for this data, at
06 and
> > 12
> > >> each day. Unless you tell METviewer to plot them separately,
they
> will
> > be
> > >> combined.
> > >>
> > >> There's just so many ways you can slice/dice the data, it isn't
clear
> to
> > >> me
> > >> exactly what you're looking for.
> > >>
> > >> I see that you're interested in the 3 models you listed. And
you want
> > to
> > >> plot the CSI statistic versus the PM2.5 threshold. But how
many lines
> > do
> > >> you want on the plot? And how do you want to define the timing
info
> > that
> > >> goes into each? What lead times should be used (just 24 hours
or
> 01-24
> > >> hours)? What initialization hours do you want to include? And
it
> looks
> > >> like you want to limit this to August 2018. Is that correct?
> > >>
> > >> Once we know exactly what you want we can help you with the
METviewer
> > >> configuration.
> > >>
> > >> Or if it's easier, we could do a telecon and work on this
together.
> > >>
> > >> Thanks,
> > >> John
> > >>
> > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA Affiliate
via RT
> <
> > >> met_help at ucar.edu> wrote:
> > >>
> > >> >
> > >> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
>
> > >> >
> > >> > Hi,
> > >> >
> > >> > I got the response from John and Tatiana, thanks.
> > >> >
> > >> > I still have question on John's XML: 20190422_17230.
> > >> >
> > >> > Mostly, I am not sure how to use the combination of FCST_LEAD
(24)
> and
> > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > >> >
> > >> > Does it mean the statistic includes all evaluation that are
> validated
> > at
> > >> > 06Z and 12Z of each day? Also, what is the meaning of
FCST_LEAD =
> 1,
> > >> > evaluation of all FCST hours? or only at 07Z and 13Z of each
day
> > >> (combine
> > >> > usage of FCST_INIT_BEGabove)?
> > >> >
> > >> > Ho-Chun
> > >> >
> > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA
Affiliate <
> > >> > ho-chun.huang at noaa.gov> wrote:
> > >> >
> > >> > > John,
> > >> > >
> > >> > > Thanks. Since you are on duty today, I am going to take
advantage
> > of
> > >> it.
> > >> > >
> > >> > > Say, I want to know the performance of the day1 (FCST 1-
24) and
> > Day 2
> > >> > > (FCST 25-48), i.e., verify each hours for day1 and day2
period and
> > >> > produce
> > >> > > a daily mean CSI. Then average over Aug 2018 for monthly
CSI
> > score.
> > >> > How
> > >> > > should I do it?
> > >> > >
> > >> > > Does 20190422_182543 plot what I need?
> > >> > >
> > >> > > Ho-Chun
> > >> > >
> > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT <
> > >> > > met_help at ucar.edu> wrote:
> > >> > >
> > >> > >> Ho-Chun,
> > >> > >>
> > >> > >> I see you have a question about using METviewer to limit
the data
> > >> going
> > >> > >> into a plot to 1 month. Thanks for pointing me to your
example
> > XML.
> > >> > >>
> > >> > >> The answer is yes, of course. You can control what is
held
> > constant
> > >> on
> > >> > >> the
> > >> > >> plot using the "Fixed Values" section. Your plot already
has
> > >> > "FCST_LEAD =
> > >> > >> 24" included in that spot. So I just added a new Fixed
Values
> > entry
> > >> for
> > >> > >> "FCST_INIT_BEG" to specify which model initialization
times
> should
> > be
> > >> > >> included. On the right side, there's a little both with
check
> mark
> > >> in
> > >> > it
> > >> > >> which is the "date picker". I clicked on that and
selected
> "Start
> > =
> > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and
"By: 6
> > >> > hours"...
> > >> > >> and then clicked "Select Values". And that picked all the
> > >> > initializations
> > >> > >> for Aug 2018.
> > >> > >>
> > >> > >> Granted, you have a long list of dates... so an easier
time
> picker
> > >> would
> > >> > >> be
> > >> > >> nice.
> > >> > >>
> > >> > >> I updated the title and then remade the plot. If you look
at the
> > >> > history,
> > >> > >> you'll find it dated as XML: 20190422_172306.
> > >> > >>
> > >> > >> Hopefully that helps.
> > >> > >>
> > >> > >> Thanks,
> > >> > >> John
> > >> > >>
> > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA
Affiliate
> via
> > >> RT <
> > >> > >> met_help at ucar.edu> wrote:
> > >> > >>
> > >> > >> >
> > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted upon.
> > >> > >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> > >> > >> > Queue: met_help
> > >> > >> > Subject: Metviewer CSI plot without date range
selection
> > >> > >> > Owner: Nobody
> > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > >> > >> > Status: new
> > >> > >> > Ticket <URL:
> > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > >> > >
> > >> > >> >
> > >> > >> >
> > >> > >> > To Whom It May Concern:
> > >> > >> >
> > >> > >> > Case XML : 20190422_151527
> > >> > >> >
> > >> > >> > I want to compare CSI score for August 2018 between
different
> AQM
> > >> > model
> > >> > >> > output. But I recall previously I can select the time
range
> for
> > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this
time I
> have
> > no
> > >> > >> option
> > >> > >> > for selecting the time range.
> > >> > >> >
> > >> > >> > My Question is : Can CSI be computed for a specific time
range?
> > >> or it
> > >> > >> just
> > >> > >> > use the wntire data in the database? IS it a problem in
> > pareparing
> > >> > the
> > >> > >> > database ? i.e., mv_emc_cmaq.
> > >> > >> >
> > >> > >> > Thank you.
> > >> > >> >
> > >> > >> > Ho-Chun
> > >> > >> >
> > >> > >> >
> > >> > >>
> > >> > >>
> > >> >
> > >> >
> > >>
> > >>
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Mon Apr 29 10:33:26 2019
Ho-Chun,
Sorry to hear about the demise of your machine! I do think that this
use
case could serve as a nice example for learning how to use the batch
engine
in METviewer.
I asked Tatiana about the AWS instance, and she said that batch isn't
currently set up. The tricky part is figuring out who can log on to
what
machines and how on AWS.
But I could probably adapt your XML for running through batch on the
existing instance if that'd be helpful. Perhaps we could step through
the
details at the next METplus/NOAA telecon. Do you think that'd be
useful?
Thanks,
John
On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
> Hi, John:
>
> Thanks for the help. My desktop dead this morning and I need to
wait for
> the new machine to get back into the Metviewer to try. All my save
xml are
> on the dead machine.
>
> Will get back to you and try again on your setting. Yes, the next
thing is
> figuring out how to use the batch machine. Perry expressed maybe to
wait
> till Amazon Cloud is available to us.
>
> Ho-Chun
>
> On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Ho Chun,
> >
> > OK, I think I understand what you want.
> >
> > You want 3 lines on the plot... for 3 different models. You want
to
> limit
> > the analysis to August, 2018. And you want separate plots for
each
> > combination of:
> > - VX_MASK region (NWC, plus other ones too probably)
> > - Initialization hour (06 or 12).
> > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> >
> > I just made an example plot on METviewer for:
> > - month = aug 2018
> > - vx_mask = NWC
> > - init_hour = 06
> > - forecast = day 1
> >
> > And I named it *cmaq_aug2018_NWC_init06_day1*. So if you look in
the
> > history, you'll find it there. In METviewer 2.9, you have the
option of
> > giving a plot a name using the "Job title" box.
> >
> > But this is just one permutation of what you want. If you only
want a
> > couple of ones, you could easily go make them by hand using the
GUI. But
> > you could also use the batch engine to automate this. Perhaps you
want
> > separate plots for:
> > - 12 months of the yera
> > - 2 initialization hours (06 and 12)
> > - 2 forecast lead time groups (day 1 and day 2)
> > - 14 subregions
> >
> > That would be 672 plots... way too many to make by hand. But we
could do
> > them all using a single XML file through the METviewer batch
engine.
> >
> > Make sense?
> >
> > Thanks,
> > John
> >
> > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA Affiliate
via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > >
> > > Hi, John:
> > >
> > > Yes, I can set up a teleconference if I still have confusion on
the
> data
> > > included.
> > >
> > > Ho-Chun
> > >
> > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA Affiliate <
> > > ho-chun.huang at noaa.gov> wrote:
> > >
> > > > Hi, John:
> > > >
> > > > From your email I think I should remove the valid hour list
entirely.
> > I
> > > > simply has trouble to imagine the relationship between
FCST_INI_BEG
> and
> > > > FCST_LEAD.I do not know if it only evaluated
> > > >
> > > > at one hour = FCST hr 24
> > > > or
> > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours
25-48)
> > > > performance.
> > > >
> > > > I really do not know why there is 54 (only Perry will know).
> > > >
> > > > What I want is to compare hourly from FCST hr 1 to FCST hr 24
between
> > > > model and observed and has a combined one CSI score for Day 1
> > performance
> > > > (statistic include all 24 hours). Then again for Day 2 CSI
score
> (from
> > > > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK +
ether 06z
> or
> > > 12Z
> > > > run].
> > > >
> > > > If you look at plot_20190424_1340030, what is your
interpretation of
> > the
> > > > intention of the plot?
> > > >
> > > > I actually want to evaluate DAY2 PM25 CSI score for 12Z runs,
in Aug
> > > 2018,
> > > > for 3 model, for NWC, Does it match with your understanding ?
> > > >
> > > > Ho-Chun
> > > >
> > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > >> Ho Chun,
> > > >>
> > > >> Sorry for the delay.
> > > >>
> > > >> So you had written about plot # 20190422_194456, and asked...
> > > >> If I change FCST LEAD to 1, then it is DAY one mean CSI
score,
> > > correct?
> > > >>
> > > >> I believe the answer is no. When I look at the available
FCST_LEAD
> > > values
> > > >> for your data through METviewer, I see 1, 2, 3, 4, 5 ..., 48,
and
> > then a
> > > >> jump to 54.
> > > >>
> > > >> I am assuming these FCST_LEAD values are in hours. So a day
ONE
> would
> > > >> either be just FCST_LEAD = 24... or maybe you'd prefer to
define day
> > one
> > > >> as
> > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> > > >>
> > > >> I also see that you're setting VALID_HOUR = 00, 01, 02... 23
in the
> > > "Fixed
> > > >> Values" section. This really isn't necessary since that's
the full
> > set
> > > of
> > > >> valid hours that are possible. So specifying this will just
make
> the
> > > >> query
> > > >> a little bit slower but have no impact on the result.
> > > >>
> > > >> I see that you have two daily initializations for this data,
at 06
> and
> > > 12
> > > >> each day. Unless you tell METviewer to plot them separately,
they
> > will
> > > be
> > > >> combined.
> > > >>
> > > >> There's just so many ways you can slice/dice the data, it
isn't
> clear
> > to
> > > >> me
> > > >> exactly what you're looking for.
> > > >>
> > > >> I see that you're interested in the 3 models you listed. And
you
> want
> > > to
> > > >> plot the CSI statistic versus the PM2.5 threshold. But how
many
> lines
> > > do
> > > >> you want on the plot? And how do you want to define the
timing info
> > > that
> > > >> goes into each? What lead times should be used (just 24
hours or
> > 01-24
> > > >> hours)? What initialization hours do you want to include?
And it
> > looks
> > > >> like you want to limit this to August 2018. Is that correct?
> > > >>
> > > >> Once we know exactly what you want we can help you with the
> METviewer
> > > >> configuration.
> > > >>
> > > >> Or if it's easier, we could do a telecon and work on this
together.
> > > >>
> > > >> Thanks,
> > > >> John
> > > >>
> > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA
Affiliate via
> RT
> > <
> > > >> met_help at ucar.edu> wrote:
> > > >>
> > > >> >
> > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > >> >
> > > >> > Hi,
> > > >> >
> > > >> > I got the response from John and Tatiana, thanks.
> > > >> >
> > > >> > I still have question on John's XML: 20190422_17230.
> > > >> >
> > > >> > Mostly, I am not sure how to use the combination of
FCST_LEAD (24)
> > and
> > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > >> >
> > > >> > Does it mean the statistic includes all evaluation that are
> > validated
> > > at
> > > >> > 06Z and 12Z of each day? Also, what is the meaning of
FCST_LEAD =
> > 1,
> > > >> > evaluation of all FCST hours? or only at 07Z and 13Z of
each day
> > > >> (combine
> > > >> > usage of FCST_INIT_BEGabove)?
> > > >> >
> > > >> > Ho-Chun
> > > >> >
> > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA
Affiliate <
> > > >> > ho-chun.huang at noaa.gov> wrote:
> > > >> >
> > > >> > > John,
> > > >> > >
> > > >> > > Thanks. Since you are on duty today, I am going to take
> advantage
> > > of
> > > >> it.
> > > >> > >
> > > >> > > Say, I want to know the performance of the day1 (FCST 1-
24) and
> > > Day 2
> > > >> > > (FCST 25-48), i.e., verify each hours for day1 and day2
period
> and
> > > >> > produce
> > > >> > > a daily mean CSI. Then average over Aug 2018 for
monthly CSI
> > > score.
> > > >> > How
> > > >> > > should I do it?
> > > >> > >
> > > >> > > Does 20190422_182543 plot what I need?
> > > >> > >
> > > >> > > Ho-Chun
> > > >> > >
> > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via RT
<
> > > >> > > met_help at ucar.edu> wrote:
> > > >> > >
> > > >> > >> Ho-Chun,
> > > >> > >>
> > > >> > >> I see you have a question about using METviewer to limit
the
> data
> > > >> going
> > > >> > >> into a plot to 1 month. Thanks for pointing me to your
example
> > > XML.
> > > >> > >>
> > > >> > >> The answer is yes, of course. You can control what is
held
> > > constant
> > > >> on
> > > >> > >> the
> > > >> > >> plot using the "Fixed Values" section. Your plot
already has
> > > >> > "FCST_LEAD =
> > > >> > >> 24" included in that spot. So I just added a new Fixed
Values
> > > entry
> > > >> for
> > > >> > >> "FCST_INIT_BEG" to specify which model initialization
times
> > should
> > > be
> > > >> > >> included. On the right side, there's a little both with
check
> > mark
> > > >> in
> > > >> > it
> > > >> > >> which is the "date picker". I clicked on that and
selected
> > "Start
> > > =
> > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00" and
"By: 6
> > > >> > hours"...
> > > >> > >> and then clicked "Select Values". And that picked all
the
> > > >> > initializations
> > > >> > >> for Aug 2018.
> > > >> > >>
> > > >> > >> Granted, you have a long list of dates... so an easier
time
> > picker
> > > >> would
> > > >> > >> be
> > > >> > >> nice.
> > > >> > >>
> > > >> > >> I updated the title and then remade the plot. If you
look at
> the
> > > >> > history,
> > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > >> > >>
> > > >> > >> Hopefully that helps.
> > > >> > >>
> > > >> > >> Thanks,
> > > >> > >> John
> > > >> > >>
> > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA
Affiliate
> > via
> > > >> RT <
> > > >> > >> met_help at ucar.edu> wrote:
> > > >> > >>
> > > >> > >> >
> > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted
upon.
> > > >> > >> > Transaction: Ticket created by ho-chun.huang at noaa.gov
> > > >> > >> > Queue: met_help
> > > >> > >> > Subject: Metviewer CSI plot without date range
selection
> > > >> > >> > Owner: Nobody
> > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > >> > >> > Status: new
> > > >> > >> > Ticket <URL:
> > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > >> > >
> > > >> > >> >
> > > >> > >> >
> > > >> > >> > To Whom It May Concern:
> > > >> > >> >
> > > >> > >> > Case XML : 20190422_151527
> > > >> > >> >
> > > >> > >> > I want to compare CSI score for August 2018 between
different
> > AQM
> > > >> > model
> > > >> > >> > output. But I recall previously I can select the time
range
> > for
> > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this
time I
> > have
> > > no
> > > >> > >> option
> > > >> > >> > for selecting the time range.
> > > >> > >> >
> > > >> > >> > My Question is : Can CSI be computed for a specific
time
> range?
> > > >> or it
> > > >> > >> just
> > > >> > >> > use the wntire data in the database? IS it a problem
in
> > > pareparing
> > > >> > the
> > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > >> > >> >
> > > >> > >> > Thank you.
> > > >> > >> >
> > > >> > >> > Ho-Chun
> > > >> > >> >
> > > >> > >> >
> > > >> > >>
> > > >> > >>
> > > >> >
> > > >> >
> > > >>
> > > >>
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon Apr 29 11:08:44 2019
Hi, John:
Yes definitely, a sample steps to get the batch working is helpful
(especially).
Actually, CSI is the first I tackled,other type of figures I need to
figure
out is time series between Model and obs Pm concentration, mean bias
(or
ME), and RMSE. So that add another variable for batch engine.
For now, yes, please use your current xml for CSI and combination of
(1) 06Z and 12Z
(2) Day 1 and Day 2
(3) East US, West Us, NWC, SWC, NMT+NPL, NEC+SEC
Let me ask to clear one thing, can I run the batch engine through web
interface Metviewer?
Ho-Chun
On Mon, Apr 29, 2019 at 12:33 PM John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ho-Chun,
>
> Sorry to hear about the demise of your machine! I do think that
this use
> case could serve as a nice example for learning how to use the batch
engine
> in METviewer.
>
> I asked Tatiana about the AWS instance, and she said that batch
isn't
> currently set up. The tricky part is figuring out who can log on to
what
> machines and how on AWS.
>
> But I could probably adapt your XML for running through batch on the
> existing instance if that'd be helpful. Perhaps we could step
through the
> details at the next METplus/NOAA telecon. Do you think that'd be
useful?
>
> Thanks,
> John
>
> On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> > Hi, John:
> >
> > Thanks for the help. My desktop dead this morning and I need to
wait for
> > the new machine to get back into the Metviewer to try. All my
save xml
> are
> > on the dead machine.
> >
> > Will get back to you and try again on your setting. Yes, the next
thing
> is
> > figuring out how to use the batch machine. Perry expressed maybe
to wait
> > till Amazon Cloud is available to us.
> >
> > Ho-Chun
> >
> > On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Ho Chun,
> > >
> > > OK, I think I understand what you want.
> > >
> > > You want 3 lines on the plot... for 3 different models. You
want to
> > limit
> > > the analysis to August, 2018. And you want separate plots for
each
> > > combination of:
> > > - VX_MASK region (NWC, plus other ones too probably)
> > > - Initialization hour (06 or 12).
> > > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> > >
> > > I just made an example plot on METviewer for:
> > > - month = aug 2018
> > > - vx_mask = NWC
> > > - init_hour = 06
> > > - forecast = day 1
> > >
> > > And I named it *cmaq_aug2018_NWC_init06_day1*. So if you look
in the
> > > history, you'll find it there. In METviewer 2.9, you have the
option
> of
> > > giving a plot a name using the "Job title" box.
> > >
> > > But this is just one permutation of what you want. If you only
want a
> > > couple of ones, you could easily go make them by hand using the
GUI.
> But
> > > you could also use the batch engine to automate this. Perhaps
you want
> > > separate plots for:
> > > - 12 months of the yera
> > > - 2 initialization hours (06 and 12)
> > > - 2 forecast lead time groups (day 1 and day 2)
> > > - 14 subregions
> > >
> > > That would be 672 plots... way too many to make by hand. But we
could
> do
> > > them all using a single XML file through the METviewer batch
engine.
> > >
> > > Make sense?
> > >
> > > Thanks,
> > > John
> > >
> > > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA Affiliate
via RT
> <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
>
> > > >
> > > > Hi, John:
> > > >
> > > > Yes, I can set up a teleconference if I still have confusion
on the
> > data
> > > > included.
> > > >
> > > > Ho-Chun
> > > >
> > > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA Affiliate
<
> > > > ho-chun.huang at noaa.gov> wrote:
> > > >
> > > > > Hi, John:
> > > > >
> > > > > From your email I think I should remove the valid hour list
> entirely.
> > > I
> > > > > simply has trouble to imagine the relationship between
FCST_INI_BEG
> > and
> > > > > FCST_LEAD.I do not know if it only evaluated
> > > > >
> > > > > at one hour = FCST hr 24
> > > > > or
> > > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2 (hours
25-48)
> > > > > performance.
> > > > >
> > > > > I really do not know why there is 54 (only Perry will know).
> > > > >
> > > > > What I want is to compare hourly from FCST hr 1 to FCST hr
24
> between
> > > > > model and observed and has a combined one CSI score for Day
1
> > > performance
> > > > > (statistic include all 24 hours). Then again for Day 2 CSI
score
> > (from
> > > > > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK +
ether 06z
> > or
> > > > 12Z
> > > > > run].
> > > > >
> > > > > If you look at plot_20190424_1340030, what is your
interpretation
> of
> > > the
> > > > > intention of the plot?
> > > > >
> > > > > I actually want to evaluate DAY2 PM25 CSI score for 12Z
runs, in
> Aug
> > > > 2018,
> > > > > for 3 model, for NWC, Does it match with your understanding
?
> > > > >
> > > > > Ho-Chun
> > > > >
> > > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > >> Ho Chun,
> > > > >>
> > > > >> Sorry for the delay.
> > > > >>
> > > > >> So you had written about plot # 20190422_194456, and
asked...
> > > > >> If I change FCST LEAD to 1, then it is DAY one mean CSI
score,
> > > > correct?
> > > > >>
> > > > >> I believe the answer is no. When I look at the available
> FCST_LEAD
> > > > values
> > > > >> for your data through METviewer, I see 1, 2, 3, 4, 5 ...,
48, and
> > > then a
> > > > >> jump to 54.
> > > > >>
> > > > >> I am assuming these FCST_LEAD values are in hours. So a
day ONE
> > would
> > > > >> either be just FCST_LEAD = 24... or maybe you'd prefer to
define
> day
> > > one
> > > > >> as
> > > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> > > > >>
> > > > >> I also see that you're setting VALID_HOUR = 00, 01, 02...
23 in
> the
> > > > "Fixed
> > > > >> Values" section. This really isn't necessary since that's
the
> full
> > > set
> > > > of
> > > > >> valid hours that are possible. So specifying this will
just make
> > the
> > > > >> query
> > > > >> a little bit slower but have no impact on the result.
> > > > >>
> > > > >> I see that you have two daily initializations for this
data, at 06
> > and
> > > > 12
> > > > >> each day. Unless you tell METviewer to plot them
separately, they
> > > will
> > > > be
> > > > >> combined.
> > > > >>
> > > > >> There's just so many ways you can slice/dice the data, it
isn't
> > clear
> > > to
> > > > >> me
> > > > >> exactly what you're looking for.
> > > > >>
> > > > >> I see that you're interested in the 3 models you listed.
And you
> > want
> > > > to
> > > > >> plot the CSI statistic versus the PM2.5 threshold. But how
many
> > lines
> > > > do
> > > > >> you want on the plot? And how do you want to define the
timing
> info
> > > > that
> > > > >> goes into each? What lead times should be used (just 24
hours or
> > > 01-24
> > > > >> hours)? What initialization hours do you want to include?
And it
> > > looks
> > > > >> like you want to limit this to August 2018. Is that
correct?
> > > > >>
> > > > >> Once we know exactly what you want we can help you with the
> > METviewer
> > > > >> configuration.
> > > > >>
> > > > >> Or if it's easier, we could do a telecon and work on this
> together.
> > > > >>
> > > > >> Thanks,
> > > > >> John
> > > > >>
> > > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA
Affiliate via
> > RT
> > > <
> > > > >> met_help at ucar.edu> wrote:
> > > > >>
> > > > >> >
> > > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > > >> >
> > > > >> > Hi,
> > > > >> >
> > > > >> > I got the response from John and Tatiana, thanks.
> > > > >> >
> > > > >> > I still have question on John's XML: 20190422_17230.
> > > > >> >
> > > > >> > Mostly, I am not sure how to use the combination of
FCST_LEAD
> (24)
> > > and
> > > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > > >> >
> > > > >> > Does it mean the statistic includes all evaluation that
are
> > > validated
> > > > at
> > > > >> > 06Z and 12Z of each day? Also, what is the meaning of
> FCST_LEAD =
> > > 1,
> > > > >> > evaluation of all FCST hours? or only at 07Z and 13Z of
each day
> > > > >> (combine
> > > > >> > usage of FCST_INIT_BEGabove)?
> > > > >> >
> > > > >> > Ho-Chun
> > > > >> >
> > > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA
Affiliate <
> > > > >> > ho-chun.huang at noaa.gov> wrote:
> > > > >> >
> > > > >> > > John,
> > > > >> > >
> > > > >> > > Thanks. Since you are on duty today, I am going to
take
> > advantage
> > > > of
> > > > >> it.
> > > > >> > >
> > > > >> > > Say, I want to know the performance of the day1 (FCST
1-24)
> and
> > > > Day 2
> > > > >> > > (FCST 25-48), i.e., verify each hours for day1 and day2
period
> > and
> > > > >> > produce
> > > > >> > > a daily mean CSI. Then average over Aug 2018 for
monthly
> CSI
> > > > score.
> > > > >> > How
> > > > >> > > should I do it?
> > > > >> > >
> > > > >> > > Does 20190422_182543 plot what I need?
> > > > >> > >
> > > > >> > > Ho-Chun
> > > > >> > >
> > > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway via
RT <
> > > > >> > > met_help at ucar.edu> wrote:
> > > > >> > >
> > > > >> > >> Ho-Chun,
> > > > >> > >>
> > > > >> > >> I see you have a question about using METviewer to
limit the
> > data
> > > > >> going
> > > > >> > >> into a plot to 1 month. Thanks for pointing me to
your
> example
> > > > XML.
> > > > >> > >>
> > > > >> > >> The answer is yes, of course. You can control what is
held
> > > > constant
> > > > >> on
> > > > >> > >> the
> > > > >> > >> plot using the "Fixed Values" section. Your plot
already has
> > > > >> > "FCST_LEAD =
> > > > >> > >> 24" included in that spot. So I just added a new
Fixed
> Values
> > > > entry
> > > > >> for
> > > > >> > >> "FCST_INIT_BEG" to specify which model initialization
times
> > > should
> > > > be
> > > > >> > >> included. On the right side, there's a little both
with
> check
> > > mark
> > > > >> in
> > > > >> > it
> > > > >> > >> which is the "date picker". I clicked on that and
selected
> > > "Start
> > > > =
> > > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00"
and
> "By: 6
> > > > >> > hours"...
> > > > >> > >> and then clicked "Select Values". And that picked all
the
> > > > >> > initializations
> > > > >> > >> for Aug 2018.
> > > > >> > >>
> > > > >> > >> Granted, you have a long list of dates... so an easier
time
> > > picker
> > > > >> would
> > > > >> > >> be
> > > > >> > >> nice.
> > > > >> > >>
> > > > >> > >> I updated the title and then remade the plot. If you
look at
> > the
> > > > >> > history,
> > > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > > >> > >>
> > > > >> > >> Hopefully that helps.
> > > > >> > >>
> > > > >> > >> Thanks,
> > > > >> > >> John
> > > > >> > >>
> > > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA
> Affiliate
> > > via
> > > > >> RT <
> > > > >> > >> met_help at ucar.edu> wrote:
> > > > >> > >>
> > > > >> > >> >
> > > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted
upon.
> > > > >> > >> > Transaction: Ticket created by ho-
chun.huang at noaa.gov
> > > > >> > >> > Queue: met_help
> > > > >> > >> > Subject: Metviewer CSI plot without date range
> selection
> > > > >> > >> > Owner: Nobody
> > > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > > >> > >> > Status: new
> > > > >> > >> > Ticket <URL:
> > > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > >> > >
> > > > >> > >> >
> > > > >> > >> >
> > > > >> > >> > To Whom It May Concern:
> > > > >> > >> >
> > > > >> > >> > Case XML : 20190422_151527
> > > > >> > >> >
> > > > >> > >> > I want to compare CSI score for August 2018 between
> different
> > > AQM
> > > > >> > model
> > > > >> > >> > output. But I recall previously I can select the
time
> range
> > > for
> > > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But this
time I
> > > have
> > > > no
> > > > >> > >> option
> > > > >> > >> > for selecting the time range.
> > > > >> > >> >
> > > > >> > >> > My Question is : Can CSI be computed for a specific
time
> > range?
> > > > >> or it
> > > > >> > >> just
> > > > >> > >> > use the wntire data in the database? IS it a
problem in
> > > > pareparing
> > > > >> > the
> > > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > > >> > >> >
> > > > >> > >> > Thank you.
> > > > >> > >> >
> > > > >> > >> > Ho-Chun
> > > > >> > >> >
> > > > >> > >> >
> > > > >> > >>
> > > > >> > >>
> > > > >> >
> > > > >> >
> > > > >>
> > > > >>
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Mon Apr 29 11:23:32 2019
Ho Chun,
No, the batch engine cannot be run via the web interface. You have to
log
on to the METviewer machine directly to run it. You set up the XML
which
describes the one or more plots you'd like to generate and then you
run the
script "mv_batch.sh" passing the XML as input on the command line.
Typically, the batch engine is used to automate the generation of
plots.
Let's say that every day or week or month you want to re-generate
several
plots for the "recent" data, you could set up the batch script to run
through the crontab and have them automatically generated.
I'll try to work up the example batch XML by our next telecon.
Thanks,
John
On Mon, Apr 29, 2019 at 11:09 AM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
> Hi, John:
>
> Yes definitely, a sample steps to get the batch working is helpful
> (especially).
>
> Actually, CSI is the first I tackled,other type of figures I need to
figure
> out is time series between Model and obs Pm concentration, mean bias
(or
> ME), and RMSE. So that add another variable for batch engine.
>
> For now, yes, please use your current xml for CSI and combination of
>
> (1) 06Z and 12Z
> (2) Day 1 and Day 2
> (3) East US, West Us, NWC, SWC, NMT+NPL, NEC+SEC
>
> Let me ask to clear one thing, can I run the batch engine through
web
> interface Metviewer?
>
> Ho-Chun
>
> On Mon, Apr 29, 2019 at 12:33 PM John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Ho-Chun,
> >
> > Sorry to hear about the demise of your machine! I do think that
this use
> > case could serve as a nice example for learning how to use the
batch
> engine
> > in METviewer.
> >
> > I asked Tatiana about the AWS instance, and she said that batch
isn't
> > currently set up. The tricky part is figuring out who can log on
to what
> > machines and how on AWS.
> >
> > But I could probably adapt your XML for running through batch on
the
> > existing instance if that'd be helpful. Perhaps we could step
through
> the
> > details at the next METplus/NOAA telecon. Do you think that'd be
useful?
> >
> > Thanks,
> > John
> >
> > On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > >
> > > Hi, John:
> > >
> > > Thanks for the help. My desktop dead this morning and I need to
wait
> for
> > > the new machine to get back into the Metviewer to try. All my
save xml
> > are
> > > on the dead machine.
> > >
> > > Will get back to you and try again on your setting. Yes, the
next
> thing
> > is
> > > figuring out how to use the batch machine. Perry expressed
maybe to
> wait
> > > till Amazon Cloud is available to us.
> > >
> > > Ho-Chun
> > >
> > > On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT <
> > > met_help at ucar.edu>
> > > wrote:
> > >
> > > > Ho Chun,
> > > >
> > > > OK, I think I understand what you want.
> > > >
> > > > You want 3 lines on the plot... for 3 different models. You
want to
> > > limit
> > > > the analysis to August, 2018. And you want separate plots for
each
> > > > combination of:
> > > > - VX_MASK region (NWC, plus other ones too probably)
> > > > - Initialization hour (06 or 12).
> > > > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> > > >
> > > > I just made an example plot on METviewer for:
> > > > - month = aug 2018
> > > > - vx_mask = NWC
> > > > - init_hour = 06
> > > > - forecast = day 1
> > > >
> > > > And I named it *cmaq_aug2018_NWC_init06_day1*. So if you look
in the
> > > > history, you'll find it there. In METviewer 2.9, you have the
option
> > of
> > > > giving a plot a name using the "Job title" box.
> > > >
> > > > But this is just one permutation of what you want. If you
only want
> a
> > > > couple of ones, you could easily go make them by hand using
the GUI.
> > But
> > > > you could also use the batch engine to automate this. Perhaps
you
> want
> > > > separate plots for:
> > > > - 12 months of the yera
> > > > - 2 initialization hours (06 and 12)
> > > > - 2 forecast lead time groups (day 1 and day 2)
> > > > - 14 subregions
> > > >
> > > > That would be 672 plots... way too many to make by hand. But
we
> could
> > do
> > > > them all using a single XML file through the METviewer batch
engine.
> > > >
> > > > Make sense?
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA
Affiliate via
> RT
> > <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > > >
> > > > > Hi, John:
> > > > >
> > > > > Yes, I can set up a teleconference if I still have confusion
on the
> > > data
> > > > > included.
> > > > >
> > > > > Ho-Chun
> > > > >
> > > > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA
Affiliate <
> > > > > ho-chun.huang at noaa.gov> wrote:
> > > > >
> > > > > > Hi, John:
> > > > > >
> > > > > > From your email I think I should remove the valid hour
list
> > entirely.
> > > > I
> > > > > > simply has trouble to imagine the relationship between
> FCST_INI_BEG
> > > and
> > > > > > FCST_LEAD.I do not know if it only evaluated
> > > > > >
> > > > > > at one hour = FCST hr 24
> > > > > > or
> > > > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2
(hours
> 25-48)
> > > > > > performance.
> > > > > >
> > > > > > I really do not know why there is 54 (only Perry will
know).
> > > > > >
> > > > > > What I want is to compare hourly from FCST hr 1 to FCST hr
24
> > between
> > > > > > model and observed and has a combined one CSI score for
Day 1
> > > > performance
> > > > > > (statistic include all 24 hours). Then again for Day 2
CSI score
> > > (from
> > > > > > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK +
ether
> 06z
> > > or
> > > > > 12Z
> > > > > > run].
> > > > > >
> > > > > > If you look at plot_20190424_1340030, what is your
interpretation
> > of
> > > > the
> > > > > > intention of the plot?
> > > > > >
> > > > > > I actually want to evaluate DAY2 PM25 CSI score for 12Z
runs, in
> > Aug
> > > > > 2018,
> > > > > > for 3 model, for NWC, Does it match with your
understanding ?
> > > > > >
> > > > > > Ho-Chun
> > > > > >
> > > > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > >> Ho Chun,
> > > > > >>
> > > > > >> Sorry for the delay.
> > > > > >>
> > > > > >> So you had written about plot # 20190422_194456, and
asked...
> > > > > >> If I change FCST LEAD to 1, then it is DAY one mean
CSI
> score,
> > > > > correct?
> > > > > >>
> > > > > >> I believe the answer is no. When I look at the available
> > FCST_LEAD
> > > > > values
> > > > > >> for your data through METviewer, I see 1, 2, 3, 4, 5 ...,
48,
> and
> > > > then a
> > > > > >> jump to 54.
> > > > > >>
> > > > > >> I am assuming these FCST_LEAD values are in hours. So a
day ONE
> > > would
> > > > > >> either be just FCST_LEAD = 24... or maybe you'd prefer to
define
> > day
> > > > one
> > > > > >> as
> > > > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> > > > > >>
> > > > > >> I also see that you're setting VALID_HOUR = 00, 01, 02...
23 in
> > the
> > > > > "Fixed
> > > > > >> Values" section. This really isn't necessary since
that's the
> > full
> > > > set
> > > > > of
> > > > > >> valid hours that are possible. So specifying this will
just
> make
> > > the
> > > > > >> query
> > > > > >> a little bit slower but have no impact on the result.
> > > > > >>
> > > > > >> I see that you have two daily initializations for this
data, at
> 06
> > > and
> > > > > 12
> > > > > >> each day. Unless you tell METviewer to plot them
separately,
> they
> > > > will
> > > > > be
> > > > > >> combined.
> > > > > >>
> > > > > >> There's just so many ways you can slice/dice the data, it
isn't
> > > clear
> > > > to
> > > > > >> me
> > > > > >> exactly what you're looking for.
> > > > > >>
> > > > > >> I see that you're interested in the 3 models you listed.
And
> you
> > > want
> > > > > to
> > > > > >> plot the CSI statistic versus the PM2.5 threshold. But
how many
> > > lines
> > > > > do
> > > > > >> you want on the plot? And how do you want to define the
timing
> > info
> > > > > that
> > > > > >> goes into each? What lead times should be used (just 24
hours
> or
> > > > 01-24
> > > > > >> hours)? What initialization hours do you want to
include? And
> it
> > > > looks
> > > > > >> like you want to limit this to August 2018. Is that
correct?
> > > > > >>
> > > > > >> Once we know exactly what you want we can help you with
the
> > > METviewer
> > > > > >> configuration.
> > > > > >>
> > > > > >> Or if it's easier, we could do a telecon and work on this
> > together.
> > > > > >>
> > > > > >> Thanks,
> > > > > >> John
> > > > > >>
> > > > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA
Affiliate
> via
> > > RT
> > > > <
> > > > > >> met_help at ucar.edu> wrote:
> > > > > >>
> > > > > >> >
> > > > > >> > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> >
> > > > > >> >
> > > > > >> > Hi,
> > > > > >> >
> > > > > >> > I got the response from John and Tatiana, thanks.
> > > > > >> >
> > > > > >> > I still have question on John's XML: 20190422_17230.
> > > > > >> >
> > > > > >> > Mostly, I am not sure how to use the combination of
FCST_LEAD
> > (24)
> > > > and
> > > > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > > > >> >
> > > > > >> > Does it mean the statistic includes all evaluation that
are
> > > > validated
> > > > > at
> > > > > >> > 06Z and 12Z of each day? Also, what is the meaning of
> > FCST_LEAD =
> > > > 1,
> > > > > >> > evaluation of all FCST hours? or only at 07Z and 13Z of
each
> day
> > > > > >> (combine
> > > > > >> > usage of FCST_INIT_BEGabove)?
> > > > > >> >
> > > > > >> > Ho-Chun
> > > > > >> >
> > > > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA
> Affiliate <
> > > > > >> > ho-chun.huang at noaa.gov> wrote:
> > > > > >> >
> > > > > >> > > John,
> > > > > >> > >
> > > > > >> > > Thanks. Since you are on duty today, I am going to
take
> > > advantage
> > > > > of
> > > > > >> it.
> > > > > >> > >
> > > > > >> > > Say, I want to know the performance of the day1
(FCST 1-24)
> > and
> > > > > Day 2
> > > > > >> > > (FCST 25-48), i.e., verify each hours for day1 and
day2
> period
> > > and
> > > > > >> > produce
> > > > > >> > > a daily mean CSI. Then average over Aug 2018 for
monthly
> > CSI
> > > > > score.
> > > > > >> > How
> > > > > >> > > should I do it?
> > > > > >> > >
> > > > > >> > > Does 20190422_182543 plot what I need?
> > > > > >> > >
> > > > > >> > > Ho-Chun
> > > > > >> > >
> > > > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway
via RT <
> > > > > >> > > met_help at ucar.edu> wrote:
> > > > > >> > >
> > > > > >> > >> Ho-Chun,
> > > > > >> > >>
> > > > > >> > >> I see you have a question about using METviewer to
limit
> the
> > > data
> > > > > >> going
> > > > > >> > >> into a plot to 1 month. Thanks for pointing me to
your
> > example
> > > > > XML.
> > > > > >> > >>
> > > > > >> > >> The answer is yes, of course. You can control what
is held
> > > > > constant
> > > > > >> on
> > > > > >> > >> the
> > > > > >> > >> plot using the "Fixed Values" section. Your plot
already
> has
> > > > > >> > "FCST_LEAD =
> > > > > >> > >> 24" included in that spot. So I just added a new
Fixed
> > Values
> > > > > entry
> > > > > >> for
> > > > > >> > >> "FCST_INIT_BEG" to specify which model
initialization times
> > > > should
> > > > > be
> > > > > >> > >> included. On the right side, there's a little both
with
> > check
> > > > mark
> > > > > >> in
> > > > > >> > it
> > > > > >> > >> which is the "date picker". I clicked on that and
selected
> > > > "Start
> > > > > =
> > > > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31 12:00:00"
and
> > "By: 6
> > > > > >> > hours"...
> > > > > >> > >> and then clicked "Select Values". And that picked
all the
> > > > > >> > initializations
> > > > > >> > >> for Aug 2018.
> > > > > >> > >>
> > > > > >> > >> Granted, you have a long list of dates... so an
easier time
> > > > picker
> > > > > >> would
> > > > > >> > >> be
> > > > > >> > >> nice.
> > > > > >> > >>
> > > > > >> > >> I updated the title and then remade the plot. If
you look
> at
> > > the
> > > > > >> > history,
> > > > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > > > >> > >>
> > > > > >> > >> Hopefully that helps.
> > > > > >> > >>
> > > > > >> > >> Thanks,
> > > > > >> > >> John
> > > > > >> > >>
> > > > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang - NOAA
> > Affiliate
> > > > via
> > > > > >> RT <
> > > > > >> > >> met_help at ucar.edu> wrote:
> > > > > >> > >>
> > > > > >> > >> >
> > > > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was acted
upon.
> > > > > >> > >> > Transaction: Ticket created by ho-
chun.huang at noaa.gov
> > > > > >> > >> > Queue: met_help
> > > > > >> > >> > Subject: Metviewer CSI plot without date
range
> > selection
> > > > > >> > >> > Owner: Nobody
> > > > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > > > >> > >> > Status: new
> > > > > >> > >> > Ticket <URL:
> > > > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > > >> > >
> > > > > >> > >> >
> > > > > >> > >> >
> > > > > >> > >> > To Whom It May Concern:
> > > > > >> > >> >
> > > > > >> > >> > Case XML : 20190422_151527
> > > > > >> > >> >
> > > > > >> > >> > I want to compare CSI score for August 2018
between
> > different
> > > > AQM
> > > > > >> > model
> > > > > >> > >> > output. But I recall previously I can select the
time
> > range
> > > > for
> > > > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But
this
> time I
> > > > have
> > > > > no
> > > > > >> > >> option
> > > > > >> > >> > for selecting the time range.
> > > > > >> > >> >
> > > > > >> > >> > My Question is : Can CSI be computed for a
specific time
> > > range?
> > > > > >> or it
> > > > > >> > >> just
> > > > > >> > >> > use the wntire data in the database? IS it a
problem in
> > > > > pareparing
> > > > > >> > the
> > > > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > > > >> > >> >
> > > > > >> > >> > Thank you.
> > > > > >> > >> >
> > > > > >> > >> > Ho-Chun
> > > > > >> > >> >
> > > > > >> > >> >
> > > > > >> > >>
> > > > > >> > >>
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > > > >>
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon Apr 29 11:47:29 2019
Hi, John:
Thanks for clarifying.
Ho-Chun
On Mon, Apr 29, 2019 at 1:27 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> Ho Chun,
>
> No, the batch engine cannot be run via the web interface. You have
to log
> on to the METviewer machine directly to run it. You set up the XML
which
> describes the one or more plots you'd like to generate and then you
run the
> script "mv_batch.sh" passing the XML as input on the command line.
>
> Typically, the batch engine is used to automate the generation of
plots.
> Let's say that every day or week or month you want to re-generate
several
> plots for the "recent" data, you could set up the batch script to
run
> through the crontab and have them automatically generated.
>
> I'll try to work up the example batch XML by our next telecon.
>
> Thanks,
> John
>
> On Mon, Apr 29, 2019 at 11:09 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> > Hi, John:
> >
> > Yes definitely, a sample steps to get the batch working is helpful
> > (especially).
> >
> > Actually, CSI is the first I tackled,other type of figures I need
to
> figure
> > out is time series between Model and obs Pm concentration, mean
bias (or
> > ME), and RMSE. So that add another variable for batch engine.
> >
> > For now, yes, please use your current xml for CSI and combination
of
> >
> > (1) 06Z and 12Z
> > (2) Day 1 and Day 2
> > (3) East US, West Us, NWC, SWC, NMT+NPL, NEC+SEC
> >
> > Let me ask to clear one thing, can I run the batch engine through
web
> > interface Metviewer?
> >
> > Ho-Chun
> >
> > On Mon, Apr 29, 2019 at 12:33 PM John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Ho-Chun,
> > >
> > > Sorry to hear about the demise of your machine! I do think that
this
> use
> > > case could serve as a nice example for learning how to use the
batch
> > engine
> > > in METviewer.
> > >
> > > I asked Tatiana about the AWS instance, and she said that batch
isn't
> > > currently set up. The tricky part is figuring out who can log
on to
> what
> > > machines and how on AWS.
> > >
> > > But I could probably adapt your XML for running through batch on
the
> > > existing instance if that'd be helpful. Perhaps we could step
through
> > the
> > > details at the next METplus/NOAA telecon. Do you think that'd
be
> useful?
> > >
> > > Thanks,
> > > John
> > >
> > > On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA Affiliate
via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
>
> > > >
> > > > Hi, John:
> > > >
> > > > Thanks for the help. My desktop dead this morning and I need
to wait
> > for
> > > > the new machine to get back into the Metviewer to try. All my
save
> xml
> > > are
> > > > on the dead machine.
> > > >
> > > > Will get back to you and try again on your setting. Yes, the
next
> > thing
> > > is
> > > > figuring out how to use the batch machine. Perry expressed
maybe to
> > wait
> > > > till Amazon Cloud is available to us.
> > > >
> > > > Ho-Chun
> > > >
> > > > On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > > Ho Chun,
> > > > >
> > > > > OK, I think I understand what you want.
> > > > >
> > > > > You want 3 lines on the plot... for 3 different models. You
want
> to
> > > > limit
> > > > > the analysis to August, 2018. And you want separate plots
for each
> > > > > combination of:
> > > > > - VX_MASK region (NWC, plus other ones too probably)
> > > > > - Initialization hour (06 or 12).
> > > > > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> > > > >
> > > > > I just made an example plot on METviewer for:
> > > > > - month = aug 2018
> > > > > - vx_mask = NWC
> > > > > - init_hour = 06
> > > > > - forecast = day 1
> > > > >
> > > > > And I named it *cmaq_aug2018_NWC_init06_day1*. So if you
look in
> the
> > > > > history, you'll find it there. In METviewer 2.9, you have
the
> option
> > > of
> > > > > giving a plot a name using the "Job title" box.
> > > > >
> > > > > But this is just one permutation of what you want. If you
only
> want
> > a
> > > > > couple of ones, you could easily go make them by hand using
the
> GUI.
> > > But
> > > > > you could also use the batch engine to automate this.
Perhaps you
> > want
> > > > > separate plots for:
> > > > > - 12 months of the yera
> > > > > - 2 initialization hours (06 and 12)
> > > > > - 2 forecast lead time groups (day 1 and day 2)
> > > > > - 14 subregions
> > > > >
> > > > > That would be 672 plots... way too many to make by hand.
But we
> > could
> > > do
> > > > > them all using a single XML file through the METviewer batch
> engine.
> > > > >
> > > > > Make sense?
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA
Affiliate via
> > RT
> > > <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > > > >
> > > > > > Hi, John:
> > > > > >
> > > > > > Yes, I can set up a teleconference if I still have
confusion on
> the
> > > > data
> > > > > > included.
> > > > > >
> > > > > > Ho-Chun
> > > > > >
> > > > > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA
Affiliate <
> > > > > > ho-chun.huang at noaa.gov> wrote:
> > > > > >
> > > > > > > Hi, John:
> > > > > > >
> > > > > > > From your email I think I should remove the valid hour
list
> > > entirely.
> > > > > I
> > > > > > > simply has trouble to imagine the relationship between
> > FCST_INI_BEG
> > > > and
> > > > > > > FCST_LEAD.I do not know if it only evaluated
> > > > > > >
> > > > > > > at one hour = FCST hr 24
> > > > > > > or
> > > > > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2
(hours
> > 25-48)
> > > > > > > performance.
> > > > > > >
> > > > > > > I really do not know why there is 54 (only Perry will
know).
> > > > > > >
> > > > > > > What I want is to compare hourly from FCST hr 1 to FCST
hr 24
> > > between
> > > > > > > model and observed and has a combined one CSI score for
Day 1
> > > > > performance
> > > > > > > (statistic include all 24 hours). Then again for Day 2
CSI
> score
> > > > (from
> > > > > > > FCST 25-FCST 48) in separate plot. [+ selected VX_MASK
+ ether
> > 06z
> > > > or
> > > > > > 12Z
> > > > > > > run].
> > > > > > >
> > > > > > > If you look at plot_20190424_1340030, what is your
> interpretation
> > > of
> > > > > the
> > > > > > > intention of the plot?
> > > > > > >
> > > > > > > I actually want to evaluate DAY2 PM25 CSI score for 12Z
runs,
> in
> > > Aug
> > > > > > 2018,
> > > > > > > for 3 model, for NWC, Does it match with your
understanding ?
> > > > > > >
> > > > > > > Ho-Chun
> > > > > > >
> > > > > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via
RT <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > >> Ho Chun,
> > > > > > >>
> > > > > > >> Sorry for the delay.
> > > > > > >>
> > > > > > >> So you had written about plot # 20190422_194456, and
asked...
> > > > > > >> If I change FCST LEAD to 1, then it is DAY one mean
CSI
> > score,
> > > > > > correct?
> > > > > > >>
> > > > > > >> I believe the answer is no. When I look at the
available
> > > FCST_LEAD
> > > > > > values
> > > > > > >> for your data through METviewer, I see 1, 2, 3, 4, 5
..., 48,
> > and
> > > > > then a
> > > > > > >> jump to 54.
> > > > > > >>
> > > > > > >> I am assuming these FCST_LEAD values are in hours. So
a day
> ONE
> > > > would
> > > > > > >> either be just FCST_LEAD = 24... or maybe you'd prefer
to
> define
> > > day
> > > > > one
> > > > > > >> as
> > > > > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated together.
> > > > > > >>
> > > > > > >> I also see that you're setting VALID_HOUR = 00, 01,
02... 23
> in
> > > the
> > > > > > "Fixed
> > > > > > >> Values" section. This really isn't necessary since
that's the
> > > full
> > > > > set
> > > > > > of
> > > > > > >> valid hours that are possible. So specifying this will
just
> > make
> > > > the
> > > > > > >> query
> > > > > > >> a little bit slower but have no impact on the result.
> > > > > > >>
> > > > > > >> I see that you have two daily initializations for this
data,
> at
> > 06
> > > > and
> > > > > > 12
> > > > > > >> each day. Unless you tell METviewer to plot them
separately,
> > they
> > > > > will
> > > > > > be
> > > > > > >> combined.
> > > > > > >>
> > > > > > >> There's just so many ways you can slice/dice the data,
it
> isn't
> > > > clear
> > > > > to
> > > > > > >> me
> > > > > > >> exactly what you're looking for.
> > > > > > >>
> > > > > > >> I see that you're interested in the 3 models you
listed. And
> > you
> > > > want
> > > > > > to
> > > > > > >> plot the CSI statistic versus the PM2.5 threshold. But
how
> many
> > > > lines
> > > > > > do
> > > > > > >> you want on the plot? And how do you want to define
the
> timing
> > > info
> > > > > > that
> > > > > > >> goes into each? What lead times should be used (just
24 hours
> > or
> > > > > 01-24
> > > > > > >> hours)? What initialization hours do you want to
include?
> And
> > it
> > > > > looks
> > > > > > >> like you want to limit this to August 2018. Is that
correct?
> > > > > > >>
> > > > > > >> Once we know exactly what you want we can help you with
the
> > > > METviewer
> > > > > > >> configuration.
> > > > > > >>
> > > > > > >> Or if it's easier, we could do a telecon and work on
this
> > > together.
> > > > > > >>
> > > > > > >> Thanks,
> > > > > > >> John
> > > > > > >>
> > > > > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA
Affiliate
> > via
> > > > RT
> > > > > <
> > > > > > >> met_help at ucar.edu> wrote:
> > > > > > >>
> > > > > > >> >
> > > > > > >> > <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > >
> > > > > > >> >
> > > > > > >> > Hi,
> > > > > > >> >
> > > > > > >> > I got the response from John and Tatiana, thanks.
> > > > > > >> >
> > > > > > >> > I still have question on John's XML: 20190422_17230.
> > > > > > >> >
> > > > > > >> > Mostly, I am not sure how to use the combination of
> FCST_LEAD
> > > (24)
> > > > > and
> > > > > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > > > > >> >
> > > > > > >> > Does it mean the statistic includes all evaluation
that are
> > > > > validated
> > > > > > at
> > > > > > >> > 06Z and 12Z of each day? Also, what is the meaning
of
> > > FCST_LEAD =
> > > > > 1,
> > > > > > >> > evaluation of all FCST hours? or only at 07Z and 13Z
of each
> > day
> > > > > > >> (combine
> > > > > > >> > usage of FCST_INIT_BEGabove)?
> > > > > > >> >
> > > > > > >> > Ho-Chun
> > > > > > >> >
> > > > > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang - NOAA
> > Affiliate <
> > > > > > >> > ho-chun.huang at noaa.gov> wrote:
> > > > > > >> >
> > > > > > >> > > John,
> > > > > > >> > >
> > > > > > >> > > Thanks. Since you are on duty today, I am going to
take
> > > > advantage
> > > > > > of
> > > > > > >> it.
> > > > > > >> > >
> > > > > > >> > > Say, I want to know the performance of the day1
(FCST
> 1-24)
> > > and
> > > > > > Day 2
> > > > > > >> > > (FCST 25-48), i.e., verify each hours for day1 and
day2
> > period
> > > > and
> > > > > > >> > produce
> > > > > > >> > > a daily mean CSI. Then average over Aug 2018 for
> monthly
> > > CSI
> > > > > > score.
> > > > > > >> > How
> > > > > > >> > > should I do it?
> > > > > > >> > >
> > > > > > >> > > Does 20190422_182543 plot what I need?
> > > > > > >> > >
> > > > > > >> > > Ho-Chun
> > > > > > >> > >
> > > > > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley Gotway
via RT
> <
> > > > > > >> > > met_help at ucar.edu> wrote:
> > > > > > >> > >
> > > > > > >> > >> Ho-Chun,
> > > > > > >> > >>
> > > > > > >> > >> I see you have a question about using METviewer to
limit
> > the
> > > > data
> > > > > > >> going
> > > > > > >> > >> into a plot to 1 month. Thanks for pointing me to
your
> > > example
> > > > > > XML.
> > > > > > >> > >>
> > > > > > >> > >> The answer is yes, of course. You can control
what is
> held
> > > > > > constant
> > > > > > >> on
> > > > > > >> > >> the
> > > > > > >> > >> plot using the "Fixed Values" section. Your plot
already
> > has
> > > > > > >> > "FCST_LEAD =
> > > > > > >> > >> 24" included in that spot. So I just added a new
Fixed
> > > Values
> > > > > > entry
> > > > > > >> for
> > > > > > >> > >> "FCST_INIT_BEG" to specify which model
initialization
> times
> > > > > should
> > > > > > be
> > > > > > >> > >> included. On the right side, there's a little
both with
> > > check
> > > > > mark
> > > > > > >> in
> > > > > > >> > it
> > > > > > >> > >> which is the "date picker". I clicked on that and
> selected
> > > > > "Start
> > > > > > =
> > > > > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31
12:00:00" and
> > > "By: 6
> > > > > > >> > hours"...
> > > > > > >> > >> and then clicked "Select Values". And that picked
all
> the
> > > > > > >> > initializations
> > > > > > >> > >> for Aug 2018.
> > > > > > >> > >>
> > > > > > >> > >> Granted, you have a long list of dates... so an
easier
> time
> > > > > picker
> > > > > > >> would
> > > > > > >> > >> be
> > > > > > >> > >> nice.
> > > > > > >> > >>
> > > > > > >> > >> I updated the title and then remade the plot. If
you
> look
> > at
> > > > the
> > > > > > >> > history,
> > > > > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > > > > >> > >>
> > > > > > >> > >> Hopefully that helps.
> > > > > > >> > >>
> > > > > > >> > >> Thanks,
> > > > > > >> > >> John
> > > > > > >> > >>
> > > > > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang -
NOAA
> > > Affiliate
> > > > > via
> > > > > > >> RT <
> > > > > > >> > >> met_help at ucar.edu> wrote:
> > > > > > >> > >>
> > > > > > >> > >> >
> > > > > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was
acted upon.
> > > > > > >> > >> > Transaction: Ticket created by ho-
chun.huang at noaa.gov
> > > > > > >> > >> > Queue: met_help
> > > > > > >> > >> > Subject: Metviewer CSI plot without date
range
> > > selection
> > > > > > >> > >> > Owner: Nobody
> > > > > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > > > > >> > >> > Status: new
> > > > > > >> > >> > Ticket <URL:
> > > > > > >> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > > > >> > >
> > > > > > >> > >> >
> > > > > > >> > >> >
> > > > > > >> > >> > To Whom It May Concern:
> > > > > > >> > >> >
> > > > > > >> > >> > Case XML : 20190422_151527
> > > > > > >> > >> >
> > > > > > >> > >> > I want to compare CSI score for August 2018
between
> > > different
> > > > > AQM
> > > > > > >> > model
> > > > > > >> > >> > output. But I recall previously I can select
the time
> > > range
> > > > > for
> > > > > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018. But
this
> > time I
> > > > > have
> > > > > > no
> > > > > > >> > >> option
> > > > > > >> > >> > for selecting the time range.
> > > > > > >> > >> >
> > > > > > >> > >> > My Question is : Can CSI be computed for a
specific
> time
> > > > range?
> > > > > > >> or it
> > > > > > >> > >> just
> > > > > > >> > >> > use the wntire data in the database? IS it a
problem
> in
> > > > > > pareparing
> > > > > > >> > the
> > > > > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > > > > >> > >> >
> > > > > > >> > >> > Thank you.
> > > > > > >> > >> >
> > > > > > >> > >> > Ho-Chun
> > > > > > >> > >> >
> > > > > > >> > >> >
> > > > > > >> > >>
> > > > > > >> > >>
> > > > > > >> >
> > > > > > >> >
> > > > > > >>
> > > > > > >>
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: John Halley Gotway
Time: Mon May 06 12:51:10 2019
Ho Chun,
I just added the following example steps to the METplus telecon
document
illustrating how to convert a METviewer GUI XML to batch. Hopefully
these
steps help get you going. I'll go ahead and resolve this ticket.
Thanks,
John
Example of converting METviewer GUI XML to batch:
-
Made a plot via the METviewer GUI named plot_20190506_154106
-
Logged onto the vm-lnx-metviewdev-process1.ncep.noaa.gov machine.
-
Located METviewer GUI xml:
ls /common/data/apps/metviewer/dev/web/xml/plot_20190506_154106.xml
-
Made a local copy of it and setup working directories:
mkdir -p /usr1/jhgotway/MET/MET_Help/HoChun_data_20190506
cd /usr1/jhgotway/MET/MET_Help/HoChun_data_20190506
mkdir plots data scripts
-
Edit the XML with the username and password:
<user>SEE GOOGLE DOC</user>
<password>SEE GOOGLE DOC</password>
-
Edit the XML with paths to the METviewer install and local output
directories:
<r_tmpl>/usr1/metviewer/metviewer/R_tmpl</r_tmpl>
<r_work>/usr1/metviewer/metviewer/R_work</r_work>
<plots>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/plots</plots>
<data>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/data</data>
<scripts>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/scripts</scripts>
-
Link in mv_batch.sh:
ln -sf /usr1/metviewer/metviewer/bin/mv_batch.sh .
-
Run batch script:
./mv_batch.sh plot_20190506_154106.xml
-
The new plot now lives in the plots sub-directory:
ls plots
-
Next, update the XML to create separate plots for multiple vx_mask
choices and edit the output file names.
cp plot_20190506_154106.xml plot_csi_by_vx_mask.xml
-
Edit the XML in the <plot_fix> section:
<field equalize="false" name="vx_mask">
<val>G104/NWC</val>
<val>G104/SWC</val>
<val>G104/SEC</val>
<val>G104/NEC</val>
</field>
-
Edit the XML to customize the output file names and title.
<data_file>plot_csi_for_{vx_mask}.data</data_file>
<plot_file>plot_csi_for_{vx_mask}.png</plot_file>
<r_file>plot_csi_for_{vx_mask}.R</r_file>
<title>CMAQ PM2.5 Performance Evaluation - {vx_mask} Aug 2018</title>
-
Rerun and you now get 4 plots instead of 1:
./mv_batch plot_csi_by_vx_mask.xml
-
This is a good start but there are several more details to work
out.
-
Specify init hours 00 and 06.
-
Specify days 1 and 2 lead hours.
-
Specify “groups” of vx_mask values for EAST, WEST, NMT+NPL, and
NEC+SEC.
On Mon, Apr 29, 2019 at 11:47 AM Ho-Chun Huang - NOAA Affiliate via RT
<
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
>
> Hi, John:
>
> Thanks for clarifying.
>
> Ho-Chun
>
> On Mon, Apr 29, 2019 at 1:27 PM John Halley Gotway via RT <
> met_help at ucar.edu>
> wrote:
>
> > Ho Chun,
> >
> > No, the batch engine cannot be run via the web interface. You
have to
> log
> > on to the METviewer machine directly to run it. You set up the
XML which
> > describes the one or more plots you'd like to generate and then
you run
> the
> > script "mv_batch.sh" passing the XML as input on the command line.
> >
> > Typically, the batch engine is used to automate the generation of
plots.
> > Let's say that every day or week or month you want to re-generate
several
> > plots for the "recent" data, you could set up the batch script to
run
> > through the crontab and have them automatically generated.
> >
> > I'll try to work up the example batch XML by our next telecon.
> >
> > Thanks,
> > John
> >
> > On Mon, Apr 29, 2019 at 11:09 AM Ho-Chun Huang - NOAA Affiliate
via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > >
> > > Hi, John:
> > >
> > > Yes definitely, a sample steps to get the batch working is
helpful
> > > (especially).
> > >
> > > Actually, CSI is the first I tackled,other type of figures I
need to
> > figure
> > > out is time series between Model and obs Pm concentration, mean
bias
> (or
> > > ME), and RMSE. So that add another variable for batch engine.
> > >
> > > For now, yes, please use your current xml for CSI and
combination of
> > >
> > > (1) 06Z and 12Z
> > > (2) Day 1 and Day 2
> > > (3) East US, West Us, NWC, SWC, NMT+NPL, NEC+SEC
> > >
> > > Let me ask to clear one thing, can I run the batch engine
through web
> > > interface Metviewer?
> > >
> > > Ho-Chun
> > >
> > > On Mon, Apr 29, 2019 at 12:33 PM John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Ho-Chun,
> > > >
> > > > Sorry to hear about the demise of your machine! I do think
that this
> > use
> > > > case could serve as a nice example for learning how to use the
batch
> > > engine
> > > > in METviewer.
> > > >
> > > > I asked Tatiana about the AWS instance, and she said that
batch isn't
> > > > currently set up. The tricky part is figuring out who can log
on to
> > what
> > > > machines and how on AWS.
> > > >
> > > > But I could probably adapt your XML for running through batch
on the
> > > > existing instance if that'd be helpful. Perhaps we could step
> through
> > > the
> > > > details at the next METplus/NOAA telecon. Do you think that'd
be
> > useful?
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA Affiliate
via
> RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > > >
> > > > > Hi, John:
> > > > >
> > > > > Thanks for the help. My desktop dead this morning and I
need to
> wait
> > > for
> > > > > the new machine to get back into the Metviewer to try. All
my save
> > xml
> > > > are
> > > > > on the dead machine.
> > > > >
> > > > > Will get back to you and try again on your setting. Yes,
the next
> > > thing
> > > > is
> > > > > figuring out how to use the batch machine. Perry expressed
maybe
> to
> > > wait
> > > > > till Amazon Cloud is available to us.
> > > > >
> > > > > Ho-Chun
> > > > >
> > > > > On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT <
> > > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > > Ho Chun,
> > > > > >
> > > > > > OK, I think I understand what you want.
> > > > > >
> > > > > > You want 3 lines on the plot... for 3 different models.
You want
> > to
> > > > > limit
> > > > > > the analysis to August, 2018. And you want separate plots
for
> each
> > > > > > combination of:
> > > > > > - VX_MASK region (NWC, plus other ones too probably)
> > > > > > - Initialization hour (06 or 12).
> > > > > > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> > > > > >
> > > > > > I just made an example plot on METviewer for:
> > > > > > - month = aug 2018
> > > > > > - vx_mask = NWC
> > > > > > - init_hour = 06
> > > > > > - forecast = day 1
> > > > > >
> > > > > > And I named it *cmaq_aug2018_NWC_init06_day1*. So if you
look in
> > the
> > > > > > history, you'll find it there. In METviewer 2.9, you have
the
> > option
> > > > of
> > > > > > giving a plot a name using the "Job title" box.
> > > > > >
> > > > > > But this is just one permutation of what you want. If you
only
> > want
> > > a
> > > > > > couple of ones, you could easily go make them by hand
using the
> > GUI.
> > > > But
> > > > > > you could also use the batch engine to automate this.
Perhaps
> you
> > > want
> > > > > > separate plots for:
> > > > > > - 12 months of the yera
> > > > > > - 2 initialization hours (06 and 12)
> > > > > > - 2 forecast lead time groups (day 1 and day 2)
> > > > > > - 14 subregions
> > > > > >
> > > > > > That would be 672 plots... way too many to make by hand.
But we
> > > could
> > > > do
> > > > > > them all using a single XML file through the METviewer
batch
> > engine.
> > > > > >
> > > > > > Make sense?
> > > > > >
> > > > > > Thanks,
> > > > > > John
> > > > > >
> > > > > > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA
Affiliate
> via
> > > RT
> > > > <
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > >
> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> >
> > > > > > >
> > > > > > > Hi, John:
> > > > > > >
> > > > > > > Yes, I can set up a teleconference if I still have
confusion on
> > the
> > > > > data
> > > > > > > included.
> > > > > > >
> > > > > > > Ho-Chun
> > > > > > >
> > > > > > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA
Affiliate
> <
> > > > > > > ho-chun.huang at noaa.gov> wrote:
> > > > > > >
> > > > > > > > Hi, John:
> > > > > > > >
> > > > > > > > From your email I think I should remove the valid hour
list
> > > > entirely.
> > > > > > I
> > > > > > > > simply has trouble to imagine the relationship between
> > > FCST_INI_BEG
> > > > > and
> > > > > > > > FCST_LEAD.I do not know if it only evaluated
> > > > > > > >
> > > > > > > > at one hour = FCST hr 24
> > > > > > > > or
> > > > > > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day 2
(hours
> > > 25-48)
> > > > > > > > performance.
> > > > > > > >
> > > > > > > > I really do not know why there is 54 (only Perry will
know).
> > > > > > > >
> > > > > > > > What I want is to compare hourly from FCST hr 1 to
FCST hr 24
> > > > between
> > > > > > > > model and observed and has a combined one CSI score
for Day 1
> > > > > > performance
> > > > > > > > (statistic include all 24 hours). Then again for Day
2 CSI
> > score
> > > > > (from
> > > > > > > > FCST 25-FCST 48) in separate plot. [+ selected
VX_MASK +
> ether
> > > 06z
> > > > > or
> > > > > > > 12Z
> > > > > > > > run].
> > > > > > > >
> > > > > > > > If you look at plot_20190424_1340030, what is your
> > interpretation
> > > > of
> > > > > > the
> > > > > > > > intention of the plot?
> > > > > > > >
> > > > > > > > I actually want to evaluate DAY2 PM25 CSI score for
12Z runs,
> > in
> > > > Aug
> > > > > > > 2018,
> > > > > > > > for 3 model, for NWC, Does it match with your
understanding ?
> > > > > > > >
> > > > > > > > Ho-Chun
> > > > > > > >
> > > > > > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway via
RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > >> Ho Chun,
> > > > > > > >>
> > > > > > > >> Sorry for the delay.
> > > > > > > >>
> > > > > > > >> So you had written about plot # 20190422_194456, and
> asked...
> > > > > > > >> If I change FCST LEAD to 1, then it is DAY one
mean CSI
> > > score,
> > > > > > > correct?
> > > > > > > >>
> > > > > > > >> I believe the answer is no. When I look at the
available
> > > > FCST_LEAD
> > > > > > > values
> > > > > > > >> for your data through METviewer, I see 1, 2, 3, 4, 5
...,
> 48,
> > > and
> > > > > > then a
> > > > > > > >> jump to 54.
> > > > > > > >>
> > > > > > > >> I am assuming these FCST_LEAD values are in hours.
So a day
> > ONE
> > > > > would
> > > > > > > >> either be just FCST_LEAD = 24... or maybe you'd
prefer to
> > define
> > > > day
> > > > > > one
> > > > > > > >> as
> > > > > > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated
together.
> > > > > > > >>
> > > > > > > >> I also see that you're setting VALID_HOUR = 00, 01,
02... 23
> > in
> > > > the
> > > > > > > "Fixed
> > > > > > > >> Values" section. This really isn't necessary since
that's
> the
> > > > full
> > > > > > set
> > > > > > > of
> > > > > > > >> valid hours that are possible. So specifying this
will just
> > > make
> > > > > the
> > > > > > > >> query
> > > > > > > >> a little bit slower but have no impact on the result.
> > > > > > > >>
> > > > > > > >> I see that you have two daily initializations for
this data,
> > at
> > > 06
> > > > > and
> > > > > > > 12
> > > > > > > >> each day. Unless you tell METviewer to plot them
> separately,
> > > they
> > > > > > will
> > > > > > > be
> > > > > > > >> combined.
> > > > > > > >>
> > > > > > > >> There's just so many ways you can slice/dice the
data, it
> > isn't
> > > > > clear
> > > > > > to
> > > > > > > >> me
> > > > > > > >> exactly what you're looking for.
> > > > > > > >>
> > > > > > > >> I see that you're interested in the 3 models you
listed.
> And
> > > you
> > > > > want
> > > > > > > to
> > > > > > > >> plot the CSI statistic versus the PM2.5 threshold.
But how
> > many
> > > > > lines
> > > > > > > do
> > > > > > > >> you want on the plot? And how do you want to define
the
> > timing
> > > > info
> > > > > > > that
> > > > > > > >> goes into each? What lead times should be used (just
24
> hours
> > > or
> > > > > > 01-24
> > > > > > > >> hours)? What initialization hours do you want to
include?
> > And
> > > it
> > > > > > looks
> > > > > > > >> like you want to limit this to August 2018. Is that
> correct?
> > > > > > > >>
> > > > > > > >> Once we know exactly what you want we can help you
with the
> > > > > METviewer
> > > > > > > >> configuration.
> > > > > > > >>
> > > > > > > >> Or if it's easier, we could do a telecon and work on
this
> > > > together.
> > > > > > > >>
> > > > > > > >> Thanks,
> > > > > > > >> John
> > > > > > > >>
> > > > > > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang - NOAA
> Affiliate
> > > via
> > > > > RT
> > > > > > <
> > > > > > > >> met_help at ucar.edu> wrote:
> > > > > > > >>
> > > > > > > >> >
> > > > > > > >> > <URL:
> > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > >
> > > > > > > >> >
> > > > > > > >> > Hi,
> > > > > > > >> >
> > > > > > > >> > I got the response from John and Tatiana, thanks.
> > > > > > > >> >
> > > > > > > >> > I still have question on John's XML:
20190422_17230.
> > > > > > > >> >
> > > > > > > >> > Mostly, I am not sure how to use the combination of
> > FCST_LEAD
> > > > (24)
> > > > > > and
> > > > > > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > > > > > >> >
> > > > > > > >> > Does it mean the statistic includes all evaluation
that
> are
> > > > > > validated
> > > > > > > at
> > > > > > > >> > 06Z and 12Z of each day? Also, what is the meaning
of
> > > > FCST_LEAD =
> > > > > > 1,
> > > > > > > >> > evaluation of all FCST hours? or only at 07Z and
13Z of
> each
> > > day
> > > > > > > >> (combine
> > > > > > > >> > usage of FCST_INIT_BEGabove)?
> > > > > > > >> >
> > > > > > > >> > Ho-Chun
> > > > > > > >> >
> > > > > > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang -
NOAA
> > > Affiliate <
> > > > > > > >> > ho-chun.huang at noaa.gov> wrote:
> > > > > > > >> >
> > > > > > > >> > > John,
> > > > > > > >> > >
> > > > > > > >> > > Thanks. Since you are on duty today, I am going
to take
> > > > > advantage
> > > > > > > of
> > > > > > > >> it.
> > > > > > > >> > >
> > > > > > > >> > > Say, I want to know the performance of the day1
(FCST
> > 1-24)
> > > > and
> > > > > > > Day 2
> > > > > > > >> > > (FCST 25-48), i.e., verify each hours for day1
and day2
> > > period
> > > > > and
> > > > > > > >> > produce
> > > > > > > >> > > a daily mean CSI. Then average over Aug 2018
for
> > monthly
> > > > CSI
> > > > > > > score.
> > > > > > > >> > How
> > > > > > > >> > > should I do it?
> > > > > > > >> > >
> > > > > > > >> > > Does 20190422_182543 plot what I need?
> > > > > > > >> > >
> > > > > > > >> > > Ho-Chun
> > > > > > > >> > >
> > > > > > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley
Gotway via
> RT
> > <
> > > > > > > >> > > met_help at ucar.edu> wrote:
> > > > > > > >> > >
> > > > > > > >> > >> Ho-Chun,
> > > > > > > >> > >>
> > > > > > > >> > >> I see you have a question about using METviewer
to
> limit
> > > the
> > > > > data
> > > > > > > >> going
> > > > > > > >> > >> into a plot to 1 month. Thanks for pointing me
to your
> > > > example
> > > > > > > XML.
> > > > > > > >> > >>
> > > > > > > >> > >> The answer is yes, of course. You can control
what is
> > held
> > > > > > > constant
> > > > > > > >> on
> > > > > > > >> > >> the
> > > > > > > >> > >> plot using the "Fixed Values" section. Your
plot
> already
> > > has
> > > > > > > >> > "FCST_LEAD =
> > > > > > > >> > >> 24" included in that spot. So I just added a
new Fixed
> > > > Values
> > > > > > > entry
> > > > > > > >> for
> > > > > > > >> > >> "FCST_INIT_BEG" to specify which model
initialization
> > times
> > > > > > should
> > > > > > > be
> > > > > > > >> > >> included. On the right side, there's a little
both
> with
> > > > check
> > > > > > mark
> > > > > > > >> in
> > > > > > > >> > it
> > > > > > > >> > >> which is the "date picker". I clicked on that
and
> > selected
> > > > > > "Start
> > > > > > > =
> > > > > > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31
12:00:00"
> and
> > > > "By: 6
> > > > > > > >> > hours"...
> > > > > > > >> > >> and then clicked "Select Values". And that
picked all
> > the
> > > > > > > >> > initializations
> > > > > > > >> > >> for Aug 2018.
> > > > > > > >> > >>
> > > > > > > >> > >> Granted, you have a long list of dates... so an
easier
> > time
> > > > > > picker
> > > > > > > >> would
> > > > > > > >> > >> be
> > > > > > > >> > >> nice.
> > > > > > > >> > >>
> > > > > > > >> > >> I updated the title and then remade the plot.
If you
> > look
> > > at
> > > > > the
> > > > > > > >> > history,
> > > > > > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > > > > > >> > >>
> > > > > > > >> > >> Hopefully that helps.
> > > > > > > >> > >>
> > > > > > > >> > >> Thanks,
> > > > > > > >> > >> John
> > > > > > > >> > >>
> > > > > > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang -
NOAA
> > > > Affiliate
> > > > > > via
> > > > > > > >> RT <
> > > > > > > >> > >> met_help at ucar.edu> wrote:
> > > > > > > >> > >>
> > > > > > > >> > >> >
> > > > > > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was
acted
> upon.
> > > > > > > >> > >> > Transaction: Ticket created by
> ho-chun.huang at noaa.gov
> > > > > > > >> > >> > Queue: met_help
> > > > > > > >> > >> > Subject: Metviewer CSI plot without date
range
> > > > selection
> > > > > > > >> > >> > Owner: Nobody
> > > > > > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > > > > > >> > >> > Status: new
> > > > > > > >> > >> > Ticket <URL:
> > > > > > > >>
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > > > > >> > >
> > > > > > > >> > >> >
> > > > > > > >> > >> >
> > > > > > > >> > >> > To Whom It May Concern:
> > > > > > > >> > >> >
> > > > > > > >> > >> > Case XML : 20190422_151527
> > > > > > > >> > >> >
> > > > > > > >> > >> > I want to compare CSI score for August 2018
between
> > > > different
> > > > > > AQM
> > > > > > > >> > model
> > > > > > > >> > >> > output. But I recall previously I can select
the
> time
> > > > range
> > > > > > for
> > > > > > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018.
But this
> > > time I
> > > > > > have
> > > > > > > no
> > > > > > > >> > >> option
> > > > > > > >> > >> > for selecting the time range.
> > > > > > > >> > >> >
> > > > > > > >> > >> > My Question is : Can CSI be computed for a
specific
> > time
> > > > > range?
> > > > > > > >> or it
> > > > > > > >> > >> just
> > > > > > > >> > >> > use the wntire data in the database? IS it a
problem
> > in
> > > > > > > pareparing
> > > > > > > >> > the
> > > > > > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > > > > > >> > >> >
> > > > > > > >> > >> > Thank you.
> > > > > > > >> > >> >
> > > > > > > >> > >> > Ho-Chun
> > > > > > > >> > >> >
> > > > > > > >> > >> >
> > > > > > > >> > >>
> > > > > > > >> > >>
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >>
> > > > > > > >>
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
Subject: Metviewer CSI plot without date range selection
From: Ho-Chun Huang - NOAA Affiliate
Time: Mon May 06 13:46:05 2019
John:
Thanks a lot. Once I can access Metviewer server. I will try.
Please
close the ticket.
Ho-Chun
On Mon, May 6, 2019 at 2:51 PM John Halley Gotway via RT
<met_help at ucar.edu>
wrote:
> Ho Chun,
>
> I just added the following example steps to the METplus telecon
document
> illustrating how to convert a METviewer GUI XML to batch. Hopefully
these
> steps help get you going. I'll go ahead and resolve this ticket.
>
> Thanks,
> John
>
> Example of converting METviewer GUI XML to batch:
>
> -
>
> Made a plot via the METviewer GUI named plot_20190506_154106
> -
>
> Logged onto the vm-lnx-metviewdev-process1.ncep.noaa.gov machine.
> -
>
> Located METviewer GUI xml:
>
> ls /common/data/apps/metviewer/dev/web/xml/plot_20190506_154106.xml
>
> -
>
> Made a local copy of it and setup working directories:
>
> mkdir -p /usr1/jhgotway/MET/MET_Help/HoChun_data_20190506
>
> cd /usr1/jhgotway/MET/MET_Help/HoChun_data_20190506
>
> mkdir plots data scripts
>
> -
>
> Edit the XML with the username and password:
>
> <user>SEE GOOGLE DOC</user>
>
> <password>SEE GOOGLE DOC</password>
>
>
> -
>
> Edit the XML with paths to the METviewer install and local output
> directories:
>
> <r_tmpl>/usr1/metviewer/metviewer/R_tmpl</r_tmpl>
>
> <r_work>/usr1/metviewer/metviewer/R_work</r_work>
>
<plots>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/plots</plots>
>
> <data>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/data</data>
>
>
<scripts>/usr1/jhgotway/MET/MET_Help/HoChun_data_20190506/scripts</scripts>
>
> -
>
> Link in mv_batch.sh:
>
> ln -sf /usr1/metviewer/metviewer/bin/mv_batch.sh .
>
> -
>
> Run batch script:
>
> ./mv_batch.sh plot_20190506_154106.xml
>
> -
>
> The new plot now lives in the plots sub-directory:
>
> ls plots
>
> -
>
> Next, update the XML to create separate plots for multiple
vx_mask
> choices and edit the output file names.
>
> cp plot_20190506_154106.xml plot_csi_by_vx_mask.xml
>
> -
>
> Edit the XML in the <plot_fix> section:
>
> <field equalize="false" name="vx_mask">
>
> <val>G104/NWC</val>
>
> <val>G104/SWC</val>
>
> <val>G104/SEC</val>
>
> <val>G104/NEC</val>
>
> </field>
>
> -
>
> Edit the XML to customize the output file names and title.
>
> <data_file>plot_csi_for_{vx_mask}.data</data_file>
>
> <plot_file>plot_csi_for_{vx_mask}.png</plot_file>
>
> <r_file>plot_csi_for_{vx_mask}.R</r_file>
>
> <title>CMAQ PM2.5 Performance Evaluation - {vx_mask} Aug
2018</title>
>
> -
>
> Rerun and you now get 4 plots instead of 1:
>
> ./mv_batch plot_csi_by_vx_mask.xml
>
> -
>
> This is a good start but there are several more details to work
out.
> -
>
> Specify init hours 00 and 06.
> -
>
> Specify days 1 and 2 lead hours.
> -
>
> Specify “groups” of vx_mask values for EAST, WEST, NMT+NPL,
and
> NEC+SEC.
>
>
> On Mon, Apr 29, 2019 at 11:47 AM Ho-Chun Huang - NOAA Affiliate via
RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> >
> > Hi, John:
> >
> > Thanks for clarifying.
> >
> > Ho-Chun
> >
> > On Mon, Apr 29, 2019 at 1:27 PM John Halley Gotway via RT <
> > met_help at ucar.edu>
> > wrote:
> >
> > > Ho Chun,
> > >
> > > No, the batch engine cannot be run via the web interface. You
have to
> > log
> > > on to the METviewer machine directly to run it. You set up the
XML
> which
> > > describes the one or more plots you'd like to generate and then
you run
> > the
> > > script "mv_batch.sh" passing the XML as input on the command
line.
> > >
> > > Typically, the batch engine is used to automate the generation
of
> plots.
> > > Let's say that every day or week or month you want to re-
generate
> several
> > > plots for the "recent" data, you could set up the batch script
to run
> > > through the crontab and have them automatically generated.
> > >
> > > I'll try to work up the example batch XML by our next telecon.
> > >
> > > Thanks,
> > > John
> > >
> > > On Mon, Apr 29, 2019 at 11:09 AM Ho-Chun Huang - NOAA Affiliate
via RT
> <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
>
> > > >
> > > > Hi, John:
> > > >
> > > > Yes definitely, a sample steps to get the batch working is
helpful
> > > > (especially).
> > > >
> > > > Actually, CSI is the first I tackled,other type of figures I
need to
> > > figure
> > > > out is time series between Model and obs Pm concentration,
mean bias
> > (or
> > > > ME), and RMSE. So that add another variable for batch engine.
> > > >
> > > > For now, yes, please use your current xml for CSI and
combination of
> > > >
> > > > (1) 06Z and 12Z
> > > > (2) Day 1 and Day 2
> > > > (3) East US, West Us, NWC, SWC, NMT+NPL, NEC+SEC
> > > >
> > > > Let me ask to clear one thing, can I run the batch engine
through web
> > > > interface Metviewer?
> > > >
> > > > Ho-Chun
> > > >
> > > > On Mon, Apr 29, 2019 at 12:33 PM John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Ho-Chun,
> > > > >
> > > > > Sorry to hear about the demise of your machine! I do think
that
> this
> > > use
> > > > > case could serve as a nice example for learning how to use
the
> batch
> > > > engine
> > > > > in METviewer.
> > > > >
> > > > > I asked Tatiana about the AWS instance, and she said that
batch
> isn't
> > > > > currently set up. The tricky part is figuring out who can
log on
> to
> > > what
> > > > > machines and how on AWS.
> > > > >
> > > > > But I could probably adapt your XML for running through
batch on
> the
> > > > > existing instance if that'd be helpful. Perhaps we could
step
> > through
> > > > the
> > > > > details at the next METplus/NOAA telecon. Do you think
that'd be
> > > useful?
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Mon, Apr 29, 2019 at 6:50 AM Ho-Chun Huang - NOAA
Affiliate via
> > RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845 >
> > > > > >
> > > > > > Hi, John:
> > > > > >
> > > > > > Thanks for the help. My desktop dead this morning and I
need to
> > wait
> > > > for
> > > > > > the new machine to get back into the Metviewer to try.
All my
> save
> > > xml
> > > > > are
> > > > > > on the dead machine.
> > > > > >
> > > > > > Will get back to you and try again on your setting. Yes,
the
> next
> > > > thing
> > > > > is
> > > > > > figuring out how to use the batch machine. Perry
expressed maybe
> > to
> > > > wait
> > > > > > till Amazon Cloud is available to us.
> > > > > >
> > > > > > Ho-Chun
> > > > > >
> > > > > > On Thu, Apr 25, 2019 at 4:38 PM John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > > Ho Chun,
> > > > > > >
> > > > > > > OK, I think I understand what you want.
> > > > > > >
> > > > > > > You want 3 lines on the plot... for 3 different models.
You
> want
> > > to
> > > > > > limit
> > > > > > > the analysis to August, 2018. And you want separate
plots for
> > each
> > > > > > > combination of:
> > > > > > > - VX_MASK region (NWC, plus other ones too probably)
> > > > > > > - Initialization hour (06 or 12).
> > > > > > > - Day 1 (1-24 hours) or Day 2 forecast (25-48 hours).
> > > > > > >
> > > > > > > I just made an example plot on METviewer for:
> > > > > > > - month = aug 2018
> > > > > > > - vx_mask = NWC
> > > > > > > - init_hour = 06
> > > > > > > - forecast = day 1
> > > > > > >
> > > > > > > And I named it *cmaq_aug2018_NWC_init06_day1*. So if
you look
> in
> > > the
> > > > > > > history, you'll find it there. In METviewer 2.9, you
have the
> > > option
> > > > > of
> > > > > > > giving a plot a name using the "Job title" box.
> > > > > > >
> > > > > > > But this is just one permutation of what you want. If
you only
> > > want
> > > > a
> > > > > > > couple of ones, you could easily go make them by hand
using the
> > > GUI.
> > > > > But
> > > > > > > you could also use the batch engine to automate this.
Perhaps
> > you
> > > > want
> > > > > > > separate plots for:
> > > > > > > - 12 months of the yera
> > > > > > > - 2 initialization hours (06 and 12)
> > > > > > > - 2 forecast lead time groups (day 1 and day 2)
> > > > > > > - 14 subregions
> > > > > > >
> > > > > > > That would be 672 plots... way too many to make by hand.
But
> we
> > > > could
> > > > > do
> > > > > > > them all using a single XML file through the METviewer
batch
> > > engine.
> > > > > > >
> > > > > > > Make sense?
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John
> > > > > > >
> > > > > > > On Wed, Apr 24, 2019 at 12:30 PM Ho-Chun Huang - NOAA
Affiliate
> > via
> > > > RT
> > > > > <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > >
> > > > > > > >
> > > > > > > > Hi, John:
> > > > > > > >
> > > > > > > > Yes, I can set up a teleconference if I still have
confusion
> on
> > > the
> > > > > > data
> > > > > > > > included.
> > > > > > > >
> > > > > > > > Ho-Chun
> > > > > > > >
> > > > > > > > On Wed, Apr 24, 2019 at 2:28 PM Ho-Chun Huang - NOAA
> Affiliate
> > <
> > > > > > > > ho-chun.huang at noaa.gov> wrote:
> > > > > > > >
> > > > > > > > > Hi, John:
> > > > > > > > >
> > > > > > > > > From your email I think I should remove the valid
hour list
> > > > > entirely.
> > > > > > > I
> > > > > > > > > simply has trouble to imagine the relationship
between
> > > > FCST_INI_BEG
> > > > > > and
> > > > > > > > > FCST_LEAD.I do not know if it only evaluated
> > > > > > > > >
> > > > > > > > > at one hour = FCST hr 24
> > > > > > > > > or
> > > > > > > > > Multiple hours for DAY 1 forecast (hr 1-24) and Day
2
> (hours
> > > > 25-48)
> > > > > > > > > performance.
> > > > > > > > >
> > > > > > > > > I really do not know why there is 54 (only Perry
will
> know).
> > > > > > > > >
> > > > > > > > > What I want is to compare hourly from FCST hr 1 to
FCST hr
> 24
> > > > > between
> > > > > > > > > model and observed and has a combined one CSI score
for
> Day 1
> > > > > > > performance
> > > > > > > > > (statistic include all 24 hours). Then again for
Day 2 CSI
> > > score
> > > > > > (from
> > > > > > > > > FCST 25-FCST 48) in separate plot. [+ selected
VX_MASK +
> > ether
> > > > 06z
> > > > > > or
> > > > > > > > 12Z
> > > > > > > > > run].
> > > > > > > > >
> > > > > > > > > If you look at plot_20190424_1340030, what is your
> > > interpretation
> > > > > of
> > > > > > > the
> > > > > > > > > intention of the plot?
> > > > > > > > >
> > > > > > > > > I actually want to evaluate DAY2 PM25 CSI score for
12Z
> runs,
> > > in
> > > > > Aug
> > > > > > > > 2018,
> > > > > > > > > for 3 model, for NWC, Does it match with your
> understanding ?
> > > > > > > > >
> > > > > > > > > Ho-Chun
> > > > > > > > >
> > > > > > > > > On Wed, Apr 24, 2019 at 2:04 PM John Halley Gotway
via RT <
> > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > >
> > > > > > > > >> Ho Chun,
> > > > > > > > >>
> > > > > > > > >> Sorry for the delay.
> > > > > > > > >>
> > > > > > > > >> So you had written about plot # 20190422_194456,
and
> > asked...
> > > > > > > > >> If I change FCST LEAD to 1, then it is DAY one
mean CSI
> > > > score,
> > > > > > > > correct?
> > > > > > > > >>
> > > > > > > > >> I believe the answer is no. When I look at the
available
> > > > > FCST_LEAD
> > > > > > > > values
> > > > > > > > >> for your data through METviewer, I see 1, 2, 3, 4,
5 ...,
> > 48,
> > > > and
> > > > > > > then a
> > > > > > > > >> jump to 54.
> > > > > > > > >>
> > > > > > > > >> I am assuming these FCST_LEAD values are in hours.
So a
> day
> > > ONE
> > > > > > would
> > > > > > > > >> either be just FCST_LEAD = 24... or maybe you'd
prefer to
> > > define
> > > > > day
> > > > > > > one
> > > > > > > > >> as
> > > > > > > > >> all FCST_LEAD = 01, 02, 03, ... 24 aggregated
together.
> > > > > > > > >>
> > > > > > > > >> I also see that you're setting VALID_HOUR = 00, 01,
02...
> 23
> > > in
> > > > > the
> > > > > > > > "Fixed
> > > > > > > > >> Values" section. This really isn't necessary since
that's
> > the
> > > > > full
> > > > > > > set
> > > > > > > > of
> > > > > > > > >> valid hours that are possible. So specifying this
will
> just
> > > > make
> > > > > > the
> > > > > > > > >> query
> > > > > > > > >> a little bit slower but have no impact on the
result.
> > > > > > > > >>
> > > > > > > > >> I see that you have two daily initializations for
this
> data,
> > > at
> > > > 06
> > > > > > and
> > > > > > > > 12
> > > > > > > > >> each day. Unless you tell METviewer to plot them
> > separately,
> > > > they
> > > > > > > will
> > > > > > > > be
> > > > > > > > >> combined.
> > > > > > > > >>
> > > > > > > > >> There's just so many ways you can slice/dice the
data, it
> > > isn't
> > > > > > clear
> > > > > > > to
> > > > > > > > >> me
> > > > > > > > >> exactly what you're looking for.
> > > > > > > > >>
> > > > > > > > >> I see that you're interested in the 3 models you
listed.
> > And
> > > > you
> > > > > > want
> > > > > > > > to
> > > > > > > > >> plot the CSI statistic versus the PM2.5 threshold.
But
> how
> > > many
> > > > > > lines
> > > > > > > > do
> > > > > > > > >> you want on the plot? And how do you want to
define the
> > > timing
> > > > > info
> > > > > > > > that
> > > > > > > > >> goes into each? What lead times should be used
(just 24
> > hours
> > > > or
> > > > > > > 01-24
> > > > > > > > >> hours)? What initialization hours do you want to
include?
> > > And
> > > > it
> > > > > > > looks
> > > > > > > > >> like you want to limit this to August 2018. Is
that
> > correct?
> > > > > > > > >>
> > > > > > > > >> Once we know exactly what you want we can help you
with
> the
> > > > > > METviewer
> > > > > > > > >> configuration.
> > > > > > > > >>
> > > > > > > > >> Or if it's easier, we could do a telecon and work
on this
> > > > > together.
> > > > > > > > >>
> > > > > > > > >> Thanks,
> > > > > > > > >> John
> > > > > > > > >>
> > > > > > > > >> On Wed, Apr 24, 2019 at 6:33 AM Ho-Chun Huang -
NOAA
> > Affiliate
> > > > via
> > > > > > RT
> > > > > > > <
> > > > > > > > >> met_help at ucar.edu> wrote:
> > > > > > > > >>
> > > > > > > > >> >
> > > > > > > > >> > <URL:
> > > https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > >
> > > > > > > > >> >
> > > > > > > > >> > Hi,
> > > > > > > > >> >
> > > > > > > > >> > I got the response from John and Tatiana, thanks.
> > > > > > > > >> >
> > > > > > > > >> > I still have question on John's XML:
20190422_17230.
> > > > > > > > >> >
> > > > > > > > >> > Mostly, I am not sure how to use the combination
of
> > > FCST_LEAD
> > > > > (24)
> > > > > > > and
> > > > > > > > >> > FCST_INIT_BEG (06Z and 12Z runs)
> > > > > > > > >> >
> > > > > > > > >> > Does it mean the statistic includes all
evaluation that
> > are
> > > > > > > validated
> > > > > > > > at
> > > > > > > > >> > 06Z and 12Z of each day? Also, what is the
meaning of
> > > > > FCST_LEAD =
> > > > > > > 1,
> > > > > > > > >> > evaluation of all FCST hours? or only at 07Z and
13Z of
> > each
> > > > day
> > > > > > > > >> (combine
> > > > > > > > >> > usage of FCST_INIT_BEGabove)?
> > > > > > > > >> >
> > > > > > > > >> > Ho-Chun
> > > > > > > > >> >
> > > > > > > > >> > On Mon, Apr 22, 2019 at 2:29 PM Ho-Chun Huang -
NOAA
> > > > Affiliate <
> > > > > > > > >> > ho-chun.huang at noaa.gov> wrote:
> > > > > > > > >> >
> > > > > > > > >> > > John,
> > > > > > > > >> > >
> > > > > > > > >> > > Thanks. Since you are on duty today, I am
going to
> take
> > > > > > advantage
> > > > > > > > of
> > > > > > > > >> it.
> > > > > > > > >> > >
> > > > > > > > >> > > Say, I want to know the performance of the
day1 (FCST
> > > 1-24)
> > > > > and
> > > > > > > > Day 2
> > > > > > > > >> > > (FCST 25-48), i.e., verify each hours for day1
and
> day2
> > > > period
> > > > > > and
> > > > > > > > >> > produce
> > > > > > > > >> > > a daily mean CSI. Then average over Aug 2018
for
> > > monthly
> > > > > CSI
> > > > > > > > score.
> > > > > > > > >> > How
> > > > > > > > >> > > should I do it?
> > > > > > > > >> > >
> > > > > > > > >> > > Does 20190422_182543 plot what I need?
> > > > > > > > >> > >
> > > > > > > > >> > > Ho-Chun
> > > > > > > > >> > >
> > > > > > > > >> > > On Mon, Apr 22, 2019 at 1:25 PM John Halley
Gotway via
> > RT
> > > <
> > > > > > > > >> > > met_help at ucar.edu> wrote:
> > > > > > > > >> > >
> > > > > > > > >> > >> Ho-Chun,
> > > > > > > > >> > >>
> > > > > > > > >> > >> I see you have a question about using
METviewer to
> > limit
> > > > the
> > > > > > data
> > > > > > > > >> going
> > > > > > > > >> > >> into a plot to 1 month. Thanks for pointing
me to
> your
> > > > > example
> > > > > > > > XML.
> > > > > > > > >> > >>
> > > > > > > > >> > >> The answer is yes, of course. You can control
what
> is
> > > held
> > > > > > > > constant
> > > > > > > > >> on
> > > > > > > > >> > >> the
> > > > > > > > >> > >> plot using the "Fixed Values" section. Your
plot
> > already
> > > > has
> > > > > > > > >> > "FCST_LEAD =
> > > > > > > > >> > >> 24" included in that spot. So I just added a
new
> Fixed
> > > > > Values
> > > > > > > > entry
> > > > > > > > >> for
> > > > > > > > >> > >> "FCST_INIT_BEG" to specify which model
initialization
> > > times
> > > > > > > should
> > > > > > > > be
> > > > > > > > >> > >> included. On the right side, there's a little
both
> > with
> > > > > check
> > > > > > > mark
> > > > > > > > >> in
> > > > > > > > >> > it
> > > > > > > > >> > >> which is the "date picker". I clicked on that
and
> > > selected
> > > > > > > "Start
> > > > > > > > =
> > > > > > > > >> > >> 2018-08-01 06:00:00" and "End = 2018-08-31
12:00:00"
> > and
> > > > > "By: 6
> > > > > > > > >> > hours"...
> > > > > > > > >> > >> and then clicked "Select Values". And that
picked
> all
> > > the
> > > > > > > > >> > initializations
> > > > > > > > >> > >> for Aug 2018.
> > > > > > > > >> > >>
> > > > > > > > >> > >> Granted, you have a long list of dates... so
an
> easier
> > > time
> > > > > > > picker
> > > > > > > > >> would
> > > > > > > > >> > >> be
> > > > > > > > >> > >> nice.
> > > > > > > > >> > >>
> > > > > > > > >> > >> I updated the title and then remade the plot.
If you
> > > look
> > > > at
> > > > > > the
> > > > > > > > >> > history,
> > > > > > > > >> > >> you'll find it dated as XML: 20190422_172306.
> > > > > > > > >> > >>
> > > > > > > > >> > >> Hopefully that helps.
> > > > > > > > >> > >>
> > > > > > > > >> > >> Thanks,
> > > > > > > > >> > >> John
> > > > > > > > >> > >>
> > > > > > > > >> > >> On Mon, Apr 22, 2019 at 9:25 AM Ho-Chun Huang
- NOAA
> > > > > Affiliate
> > > > > > > via
> > > > > > > > >> RT <
> > > > > > > > >> > >> met_help at ucar.edu> wrote:
> > > > > > > > >> > >>
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > Mon Apr 22 09:25:34 2019: Request 89845 was
acted
> > upon.
> > > > > > > > >> > >> > Transaction: Ticket created by
> > ho-chun.huang at noaa.gov
> > > > > > > > >> > >> > Queue: met_help
> > > > > > > > >> > >> > Subject: Metviewer CSI plot without
date range
> > > > > selection
> > > > > > > > >> > >> > Owner: Nobody
> > > > > > > > >> > >> > Requestors: ho-chun.huang at noaa.gov
> > > > > > > > >> > >> > Status: new
> > > > > > > > >> > >> > Ticket <URL:
> > > > > > > > >>
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=89845
> > > > > > > > >> > >
> > > > > > > > >> > >> >
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > To Whom It May Concern:
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > Case XML : 20190422_151527
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > I want to compare CSI score for August 2018
between
> > > > > different
> > > > > > > AQM
> > > > > > > > >> > model
> > > > > > > > >> > >> > output. But I recall previously I can
select the
> > time
> > > > > range
> > > > > > > for
> > > > > > > > >> > >> > comparison, e.g., AUG 2018 versus SEP 2018.
But
> this
> > > > time I
> > > > > > > have
> > > > > > > > no
> > > > > > > > >> > >> option
> > > > > > > > >> > >> > for selecting the time range.
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > My Question is : Can CSI be computed for a
specific
> > > time
> > > > > > range?
> > > > > > > > >> or it
> > > > > > > > >> > >> just
> > > > > > > > >> > >> > use the wntire data in the database? IS it
a
> problem
> > > in
> > > > > > > > pareparing
> > > > > > > > >> > the
> > > > > > > > >> > >> > database ? i.e., mv_emc_cmaq.
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > Thank you.
> > > > > > > > >> > >> >
> > > > > > > > >> > >> > Ho-Chun
> > > > > > > > >> > >> >
> > > > > > > > >> > >> >
> > > > > > > > >> > >>
> > > > > > > > >> > >>
> > > > > > > > >> >
> > > > > > > > >> >
> > > > > > > > >>
> > > > > > > > >>
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>
------------------------------------------------
More information about the Met_help
mailing list