Hi Kurt.
Thanks for the reply.
You are correct that mbgrid outputs a GMT netCDF file. I was running
gdal_translate to overcome a weirdness that may no longer exist but
previously caused the output to be reflected when displayed in GIS
tools. This was due to a difference between GDAL utilities and the GMT
tools in indexing the data (as I recall).
Geotiffs are a nice alternative. I was just used to using netCDF.
Joaquim Luis explained the problem as due to gdal using an old GMT3
format (below)
>>
John,
No wonder the headers are different. The gdal_translate -of GMT converts
the grid into the old-old-old GMT3 version netCDF that was not CF
compliant. And if you want to do that, no need to call gdal_translate,
all you need is to append '=cf' to grid's name of mbgrid command (I'm
assuming that mbgrid delegates all grid writing work to GMT).
In summary, I don't see any problem here as it's very unlikely that Arc
will be able to read that old-old-old netCDF variant (unless it lets
GDAL do all the work)
Joaquim
>>
J.
On 5/5/13 5:39 PM, Kurt Schwehr wrote:
Hi John,
- It's Sunday so the brain isn't really engaged, but doesn't mbgrid
output a gmt grd file? (your dump shows that) Why are you running
gdal_translate when your issue was with qgis/gdal and your destination
is ArcGIS?
- gdal 1.10.0 is just out. Before trying anything else, you might try
upgrading to that.
- Why not write a geotif from gdal? You get the option of LZW
compression then and a lot more tools can read geotif than can read
gmt netcdf grids... it's also a much more tested path through gdal.
Since I never seem to remember how to do compress and I always seem
monstrous amounts of files... for the record, to LZW compress a
geotif, add this:
-co "COMPRESS=LZW"
- are you seeing the same things from gdalinfo on the output as the
input? (e.g. is gdal being self consistent)
- you never mentioned what options you were passing to gdal_translate.
One of your options could be the issue.
-kurt
On May 5, 2013, at 2:04 PM, John Helly <[email protected]
<mailto:[email protected]>> wrote:
Hi.
A few years ago I had a problem with GMT/GDAL interactions such that
the GMT netCDF files were reflected when read into Qgis (for
example). This was worked around by running it through
'gdal_translate -of GMT'. There was also a discussion about this
being a bug in something in GDAL (since Qgis was using GDAL if I
remember correctly) that was supposed to have been fixed in some
version (obviously don't remember the details). At any rate, it
stopped being a problem for me but now something else is causing a
problem in files that have the following workflow:
--> mbgrid (from MB-System) --> gdal_translate -of GMT --> ArcGIS
netCDF reader
Now I find that gdal_translate appears to be changing the header in a
way such that ArcGIS cannot read the netCDF. The first 'ncdump -h'
below is what the header looks like before running through
gdal_translate and the second one is after gdal_translate.
I'm sending this note to MB-System, GMT and GDAL groups since it
seems like an issue of import to all three. Any thoughts on this?
J.
=======================================Output from mbgrid before
gdal_translate ==================================================
ncdump -h MCBCP_BeachModel_UTM11N_Baseline.grdclip.nc
netcdf MCBCP_BeachModel_UTM11N_Baseline.grdclip {
dimensions:
x = 12626 ;
y = 12251 ;
variables:
double x(x) ;
x:long_name = "Easting (meters)" ;
x:actual_range = 444057.827834802, 469309.046567392 ;
double y(y) ;
y:long_name = "Northing (meters)" ;
y:actual_range = 3670183.00468016, 3694684.53612513 ;
float z(y, x) ;
z:long_name = "Topography (m)" ;
z:_FillValue = NaNf ;
z:actual_range = -1995.541015625, 599.818359375 ;
// global attributes:
:Conventions = "COARDS/CF-1.0" ;
:title = "Topography Grid" ;
:history = "grdclip -V
/media/EXT4_Projects_01/SLR_Data/level1/MCBCP/Basemap/MCBCP_BeachModel_UTM11N_Baseline.grd
-G/media/EXT4_Projects_01/SLR_Data/level1/MCBCP/Basemap/MCBCP_BeachModel_UTM11N_Baseline.grdclip.nc
-Sa600/NaN -Sb-2000/NaN" ;
:description = "\n",
"\tProjection: UTM11N\n",
"\tGrid created by mbgrid\n",
"\tMB-system Version 5.3.1982\n",
"\tRun by <hellyj> on <StonestepsLinux> at <Thu Apr 25
19:44:41 2013>" ;
:GMT_version = "4.5.9_r9909 [64-bit]" ;
}
=====================================Output from mbgrid after
gdal_translate ====================================================
ncdump -h MCBCP_BeachModel_UTM11N_Baseline.grdclip.flipped.nc
netcdf MCBCP_BeachModel_UTM11N_Baseline.grdclip.flipped {
dimensions:
side = 2 ;
xysize = 154681126 ;
variables:
double x_range(side) ;
x_range:units = "meters" ;
double y_range(side) ;
y_range:units = "meters" ;
double z_range(side) ;
z_range:units = "meters" ;
double spacing(side) ;
int dimension(side) ;
float z(xysize) ;
z:scale_factor = 1. ;
z:add_offset = 0. ;
z:node_offset = 1 ;
// global attributes:
:title = "" ;
:source = "" ;
}
--
John Helly, University of California, San Diego / San Diego Supercomputer
Center / Scripps Institution of Oceanography / 760 840 8660 mobile / stonesteps
(Skype) / stonesteps7 (iChat) /http://www.sdsc.edu/~hellyj
--
John Helly, University of California, San Diego / San Diego Supercomputer
Center / Scripps Institution of Oceanography / 760 840 8660 mobile / stonesteps
(Skype) / stonesteps7 (iChat) / http://www.sdsc.edu/~hellyj
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
