[GRASSLIST:3190] bathymetry

GRASS GRASS user
1

Hi all, I have been attempting to import the Smith and Sandwell worldwide
bathymetry map, and have been having a lot of difficulty. The data comes as a
mercator projected file on a sphere with coordinates in latitude and longitude.
I know the number of rows and columns, and the bounding coordinates (in
lat-long). How can I import this entire data set?

I tried to import it in many different locations and projecting among them,
latlong, merc, xy (and then projecting into latlong). All of them end up
giving me an image that is not georeferenced correctly. It is usually close,
but it does not line up well with GTOPO30 data, or digital chart of the world
data. I have been using GRASS for a couple of years now, and am pretty
comfortable with it, but I don't have any idea what I am doing wrong. I am
making sure that the projection parameters are correct (ellipsoid=sphere), but
it seems like a projection problem with the slight offset of the data.

I also tried to use img2mercgrd and grd2xyz in GMT and then tried to import the
data as an ascii file. Again, the data is offset. I have used this data in
conjunction with GTOPO30 data in GMT, so I know that they really do line up.

Has anyone else imported this data into GRASS, and if so, could you tell me how
you did it?

Thanks, Ian

2

Hamish, thanks for the response. I am using GRASS 5.0.0 on a mac (10.2.8). I
am not sure what you mean by datum support. I thought that I specified the
datum for a location when I set it up. Do you mean that I can not project
across datums when using a command like r.proj?

the address for this data is here

http://topex.ucsd.edu/marine_topo/mar_topo.html

-Ian

> Hi all, I have been attempting to import the Smith and Sandwell worldwide
> bathymetry map, and have been having a lot of difficulty. The data comes
as a
> mercator projected file on a sphere with coordinates in latitude and
longitude.
> I know the number of rows and columns, and the bounding coordinates (in
> lat-long). How can I import this entire data set?

Which version of GRASS are you using?

To get correct datum support, you will have to use at least version 5.3.
This will probably only account for an error of a couple hundred meters
though.

Could you post a link to the public data repository? (if there is one)

good luck,
Hamish

3

I am using GRASS 5.0.0 on a mac (10.2.8). I am not sure what you mean
by datum support.

for a better explanation than I could give you, see:
http://biology.usgs.gov/geotech/documents/datum.html

I thought that I specified the datum for a location
when I set it up. Do you mean that I can not project across datums
when using a command like r.proj?

Not with GRASS 5.0.0, maybe there is limmited NAD27 and NAD83 support?
Someone else would have to answer that, but generally no; you need a
newer version of GRASS for r.proj to handle transformations between
different datums correctly.

the address for this data is here

http://topex.ucsd.edu/marine_topo/mar_topo.html

cheers.

Some projection and datum info is here:
ftp://topex.ucsd.edu/pub/global_topo_2min/arcinfo.html

Hamish

4

On Sat, Apr 17, 2004 at 02:16:50PM -0700, Ian Macmillan wrote:

Hamish, thanks for the response. I am using GRASS 5.0.0 on a mac (10.2.8). I
am not sure what you mean by datum support. I thought that I specified the
datum for a location when I set it up. Do you mean that I can not project
across datums when using a command like r.proj?

Right. For datum transformation support you need GRASS 5.3 or 5.7.

Markus

> > Hi all, I have been attempting to import the Smith and Sandwell worldwide
> > bathymetry map, and have been having a lot of difficulty. The data comes
> as a
> > mercator projected file on a sphere with coordinates in latitude and
> longitude.
> > I know the number of rows and columns, and the bounding coordinates (in
> > lat-long). How can I import this entire data set?

>
> Which version of GRASS are you using?
>
>
> To get correct datum support, you will have to use at least version 5.3.
> This will probably only account for an error of a couple hundred meters
> though.
>
> Could you post a link to the public data repository? (if there is one)
>
>
> good luck,
> Hamish
>

--
Markus Neteler <neteler itc it> http://mpa.itc.it
ITC-irst - Centro per la Ricerca Scientifica e Tecnologica
MPBA - Predictive Models for Biol. & Environ. Data Analysis
Via Sommarive, 18 - 38050 Povo (Trento), Italy

5

On Sun, 18 Apr 2004, Hamish wrote:

> I am using GRASS 5.0.0 on a mac (10.2.8). I am not sure what you mean
> by datum support.

for a better explanation than I could give you, see:
http://biology.usgs.gov/geotech/documents/datum.html

> I thought that I specified the datum for a location
> when I set it up. Do you mean that I can not project across datums
> when using a command like r.proj?

Not with GRASS 5.0.0, maybe there is limmited NAD27 and NAD83 support?
Someone else would have to answer that, but generally no; you need a
newer version of GRASS for r.proj to handle transformations between
different datums correctly.

But you can export the data, & run it through the cs2cs program manually.
This is a part of the proj package (the same one GRASS uses to manage
reprojection).

From the man page:

       Cs2cs performs transformation between the source
       and destination cartographic coordinate system
       on a set of input points. The coordinate system
       transformation can include translation between
       projected and geographic coordinates as well as
       the application of datum shifts.

After reprojecting it (with the datum shift) you can create a new GRASS
mapset for the new projection and import the data.

I don't believe you can overlay data from different mapsets on the fly
though.

Brent Wood

6

On Mon, 19 Apr 2004, Brent Wood wrote:

[...]

After reprojecting it (with the datum shift) you can create a new GRASS
mapset for the new projection and import the data.

I don't believe you can overlay data from different mapsets on the fly
though.

Just to correct this for the mailing list record...
You can overlay data from different mapsets within a location (i.e. they all
have the same projection) but not from a different location...the whole
GRASS location concept is based around the fact that locations are
incompatible in this way.

You could use viewproj (src.contrib/LM/viewproj) for viewing data from
different locations overlaid, but I haven't tried it and it is old and doesn't
support datum conversions so maybe isn't a lot of use.

Paul