[Met_help] [rt.rap.ucar.edu #97167] History for Bootstrapping FSS in METviewer

John Halley Gotway via RT met_help at ucar.edu
Fri Oct 23 10:43:13 MDT 2020


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

Hello!

I would like to add bootstrap confidence intervals to my aggregate
Fractions Skill Scores (FSS) calculated with MET.
Although Stat-Analysis doesn't have this ability, MET Viewer apparently
does, according to this page from MET-help archive:
https://mailman.ucar.edu/pipermail/met_help/2019-July/003502.html
Do you still recommend we use the MET Viewer? I briefly poked around the
web application, but didn't find any database examples with FSS.
If yes, should I start with the containerized version of MET Viewer, or is
there a dedicated module on NCAR HPCs like cheyenne or casper? I'm
sorry--I'm having trouble finding information on this via the usual
googling.

Dave

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

Subject: Bootstrapping FSS in METviewer
From: John Halley Gotway
Time: Wed Oct 21 12:01:30 2020

David,

I used your latest question to create a new met-help ticket instead of
continuing the older one.

As of right now, there is no instance of METviewer running on
cheyenne, and I don't expect there to be one because of the need for a
database. However, our team is in the process of refactoring/breaking
METviewer down into smaller pieces, including METplotpy and
METcalcpy... for plotting and computations in python, respectively.
The goal is for those packages to be able to run standalone reading
from flat files instead of requiring a database. But that doesn't
actually exist yet. So I don't have an example to share.

So yes, I would recommend running METviewer on a laptop using a Docker
container. In fact, just last week I posted a training video
describing how to do that:

https://dtcenter.github.io/METplus-
Training/modules/METviewer/docker.html

Please LMK if you have any feedback on that training content.

As for an example using existing data to compute bootstrap CI's for
FSS, please go to this instance of METviewer:
http://dtcenter.ucar.edu/met/metviewer/metviewer1.jsp

You can use the attached .png/.xml to upload there and play around
with the example. I created FSS output when comparing CCPA and StageIV
precip datasets to quantify how similar they are.

Thanks,
John



------------------------------------------------
Subject: Bootstrapping FSS in METviewer
From: David Ahijevych
Time: Wed Oct 21 12:06:00 2020

Very nice! This looks like exactly what I need.


On Wed, Oct 21, 2020 at 12:02 PM John Halley Gotway via RT <
met_help at ucar.edu> wrote:

> David,
>
> I used your latest question to create a new met-help ticket instead
of
> continuing the older one.
>
> As of right now, there is no instance of METviewer running on
cheyenne,
> and I don't expect there to be one because of the need for a
database.
> However, our team is in the process of refactoring/breaking
METviewer down
> into smaller pieces, including METplotpy and METcalcpy... for
plotting and
> computations in python, respectively. The goal is for those packages
to be
> able to run standalone reading from flat files instead of requiring
a
> database. But that doesn't actually exist yet. So I don't have an
example
> to share.
>
> So yes, I would recommend running METviewer on a laptop using a
Docker
> container. In fact, just last week I posted a training video
describing how
> to do that:
>
> https://dtcenter.github.io/METplus-
Training/modules/METviewer/docker.html
>
> Please LMK if you have any feedback on that training content.
>
> As for an example using existing data to compute bootstrap CI's for
FSS,
> please go to this instance of METviewer:
> http://dtcenter.ucar.edu/met/metviewer/metviewer1.jsp
>
> You can use the attached .png/.xml to upload there and play around
with
> the example. I created FSS output when comparing CCPA and StageIV
precip
> datasets to quantify how similar they are.
>
> Thanks,
> John
>
>
>
> mohawk mv_hmt_ccpa_st4 ****** ****** mariadb
/usr/local/R/bin/Rscript
> /opt/vxwww/tomcat/webapps/metviewer//R_tmpl
> /opt/vxwww/tomcat/webapps/metviewer//R_work
> /d2/www/dtcenter/met/metviewer_output//plots
> /d2/www/dtcenter/met/metviewer_output//data
> /d2/www/dtcenter/met/metviewer_output//scripts series_plot.R_tmpl
NBR_FSS
> CCPA_vs_ST4 >=2.540 >=6.350 >=12.700 >=25.400 >=50.800 NBRHD 9 25 49
true
> 1000 perc false false median plot_bootstrap_fss_example.data
> plot_bootstrap_fss_example.png plot_bootstrap_fss_example.R
Neighborhood
> Size FSS
> bootstrap_fss_example false list() list() Rscript false false false
false
> true false true false false false false false true png16m 8.5 11 72
in
> c(8,4,5,4) c(1,1,0) 1 2 1.4 -2 0.5 1 -0.75 0.5 0 1 1 1 2 0.5 1 0.5
0.5 1 1
> 1 -2 0.5 3 #cccccc 1 listX 1 1 0.5 0.8 0.8 -0.5 0.5 1 0.5 0.5 1 1 1
0.5 o c(0,
> -.25) 3 0.8 1 #333333 0.8 3 0 0.05
c("boot","boot","boot","boot","boot")
> c(FALSE,FALSE,FALSE,FALSE,FALSE) c(TRUE,TRUE,TRUE,TRUE,TRUE)
> c("#ff0000FF","#8000ffFF","#00ff7fFF","#55ff00FF","#bfff00FF")
> c(20,20,20,20,20) c("b","b","b","b","b") c(1,1,1,1,1) c(1,1,1,1,1)
> c(1,1,1,1,1) c(1,2,3,4,5) c("","","","","") c() c() 0.04 c()

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


More information about the Met_help mailing list