[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[GEMPAK #MGA-118138]: Filtering of the number of points in LINE

> Hi Steve,
> 
> Setting CONTUR=3/3 did result in nicer smoothing.  I want to
> understand better how CONTUR works.  My understanding is that the
> first number indicates the sub-boxes to be divided in each grid cell.
> So in this case, does 3 mean that each grid cell is now sub-divided
> into 9 boxes?

Yes, that would be the case.

> 
> As for the 2nd number, I thought it works like a 0 or 1 switch.  Do
> numbers > 1 do anything extra?

No, the second number on CONTUR is the number of smoothing passes.It
should be an integer. Numbers greater than 1 imply additional smoothing passes.


> 
> I am still looking to see if there is a way to decrease the number of
> points used to compose a line beyond the max=1.0 setting in the last
> entry of LINE.  I guess I am looking for a 2, or something similiar to
> IJSKIP=1, although in this case, it is not to skip i and j but "point
> skip".  I suppose this would require a modification of the code, right?

There is a program that NCEP uses called mkelev which I haven't used- to 
generate
topography contours. A companion program called ccplot is also in the
$GEMPAK/utilities/mkelev directory which you could look at for some ideas.
I haven't attempted to do what you are looking to do, but since you obtain
GEMPAK from NCEP, they are probably the best folks to ask about how to
accomplish what you are trying to do.

Steve Chiswell
Unidata User Support




> 
> Kwan
> 
> ----- Original Message -----
> From: Unidata GEMPAK Support <address@hidden>
> Date: Monday, July 9, 2007 5:24 pm
> Subject: [GEMPAK #MGA-118138]: Filtering of the number of points in
> LINE
> 
> > > Steve,
> > >
> > > Thank you for your help on my problems on SFMOD.
> > >
> > > I have a question on a slightly different subject.  It is about
> > the way
> > > LINE plots contours.  Right now, I am using gdplot2 with
> > DEVICE=vg to
> > > generate objectively analyzed isobars in a VGF file which can be
> > opened> and edited in NMAP.  The goal is to have the isobars
> > smooth enough so
> > > that minimum amount of editing is required.  I've already set a
> > couple> of things to smooth the isobars.  They are:
> > >
> > > 1. GDPFUN = SM5S(PMSL)
> > > 2. CONTUR = 9/1
> > > 3. LINE = 5/1/1-1/2/1
> >
> > Kwan,
> >
> > You might have better luck using a higher number of passes in
> > CONTUR rather than
> > subboxes or SM5S()/SM9S(). I typically find something like
> > CONTUR=3/3 provides a nice
> > amount of smoothing without losing the features of the data. Also
> > note you
> > are misisng the "/" character before the "-1" in LINE above.
> >
> > You can also set IJSKIP to subsample your data by every other point
> > such as IJSKIP=1.
> >
> > Steve Chiswell
> > Unidata User Support
> >
> >
> > >
> > > However, the resulting isobars are still not sufficiently
> > smooth.  I
> > > could use SM5S more than once and get smoother isobars outside
> > of high
> > > and low pressure centers.  However, some smaller closed centers are
> > > filtered away.  I could also increase the # of subboxes to e.g.
> 64/1
> > > and get smoother isobars.  But then they are broken up into smaller
> > > line segments (although they look like continuous lines in
> > NMAP).  The
> > > next thing I do is to set the filter in LINE to 1 (max).
> > Compared with
> > > using 0 for the filter, less points are used to produce the line.
> > > However, there are still too many points used even at 1.  I
> > think I'll
> > > need only about half of these points.  So I want to double the
> > amount> of point filtering.  But I can't set it to 2, right?  If I
> > really want
> > > to do that, I'll have to change the filter program itself, right?
> > >
> > > Kwan
> > >
> > >
> >
> >
> > Ticket Details
> > ===================
> > Ticket ID: MGA-118138
> > Department: Support GEMPAK
> > Priority: Normal
> > Status: Closed
> >
> >
> 
> 


Ticket Details
===================
Ticket ID: MGA-118138
Department: Support GEMPAK
Priority: Normal
Status: Closed