Hi Dan,
This may not be exactly what you are after but here is a short go
script that I sometimes use to plot a variable from my model which has
an icosahedral grid. For this grid the coords are specified using two
1D arrays, lon and lat, like you. The script uses the polygon command
to draw circles at the locations. The downside is you end up with gaps
between the cells. It's a 'quick look' approach if you like. Ideally
what you need is a ferret command to follow the polygon shape defined
by the lat/long coordinates but I don't think that's possible without
some effort. I usually interpolate to a high resolution regular
lat/long grid to do my analysis.
Glenn
--
! start
set mode verify
!
! viewbbxy - plots lat/lon map of bucky variable on bucky grid
!
! Args: $1 - x range for go basemap e.g. x=180:360
! $2 - y range for go basemap e.g. y=0:-40
! $3 - polygon command arguments e.g. /lev=(-2,20,2)/title="pv"
! note that /over/key is not needed as this script adds it.
! $4 - variable to be plotted. e.g. sigma[k=1] (see note below)
! $5 - (optional) scale factor. If absent, set to 0.8
\cancel mode verify
!
! Note
! The polygon script works by using the 'k' axis to produce the
polygons.
! Therefore if you have a region set for the k axis it will interfere
! and polymark won't work. Likewise if you have a region set for the
x axis
! it will interfere with selecting all the gridpoints from the variable.
! It's best not to set a region, but if you do, then only set one for
the
! time axis.
!
! save current region
define region/default save
let/quiet scale = $5%0.8%
go basemap $1 $2
go polymark polygon/over/key$3 lon lat $4 circle scale
! restore settings
set region save
! end
At 3:27 am +0800 26/4/07, Yu, Hao-Cheng wrote:
Hi Steve
Thanks for your reply.
But my problem is not Lack of basemap.
My question is:
Is Ferret can do plot(shade, or vector) with an irregular gridded
data(from finite element model) without Object Analysis.
Such dataset won't have 2-D definition for variables, usually only
1-D definition.
Like
I J K L
Lat 1:10 : : :
Lon 1:10 : : :
Var 1:10 : : :
Variables location are defined as an irregular grid (Lat, Lon).
However, Ferret only do "shade" or "vector" plot for 2-D variables.
Any comments or suggestions?
Sincerely, dan
2007/4/26, Steve Guimond
<<mailto:guimond@xxxxxxxxxxxxxxxxx>guimond@xxxxxxxxxxxxxxxxx>:
Hi,
Try laying down a basemap first, then using the shade command.
go basemap x=...
=======================================================
Stephen R. Guimond
Graduate Research Assistant
Center for Ocean-Atmospheric Prediction Studies (COAPS)
=======================================================
On Thu, 26 Apr 2007, Yu, Hao-Cheng wrote:
Dear ferreters,
I know Ferret can deal with grided data (for finite difference
method) and
it always works perfectly.
However, after searching on the mail archives and UserGuide.
Irregular dataset(like finite element method) seems only can be
plotted by
Regriding process(like objective analysis).
I try with "3-arguments shade (even vector)", and Ferret tells if the
dataset is not define as 2D, this won't work.
file/var=lon,lat,tmp test.dat
shade tmp,lon,lat
test.dat contains 3 columns data which define each element node
location and
the property to plot (shade).
Is it possible to "shade" or "vector" on irregular grid?
Sincerely, Dan