[GRASS-user] Latitude/Longitude vs UTM

GRASS GRASS user
1

Hi Folks,

I have been playing with the seamless maps from the USGS. Specifically, my question is on the effect of Latitude/Longitude vs UTM with respect to hydrology modules such as r.watershed and r.topidx. I was trying to generate watershed maps with r.watershed. I've done this before for my dissertation. They did not turn out the same. They looked strange compared to what I generated before. Moreover, they maps kept falling in the same place in the map set regardless of what I turned off or on. (i.e. some where in the center top, when the map was generated from a dem on the bottom.) I then tried to run r.topidx. This was interesting in that it told me that r.topidx could not be run with latitude and longitude and I had to convert to UTM. I was wondering if this is the answer to the problem and I just had to convert to UTM.

One other question. New Hampshire appears to fall within two UTM zones (19T and 20T) Is there a way for a maps set to contain two UTM zones?

Thanks for any help.

Kurt Springs

2

On Thu, 13 May 2010, Kurt Springs wrote:

This was interesting in that it told me that r.topidx could not be run
with latitude and longitude and I had to convert to UTM. I was wondering
if this is the answer to the problem and I just had to convert to UTM.

Kurt,

   Lat/Long represents geographic coordinates, not a projection of location
on a mathematial model of the earth. UTM is the Universal Transverse
Mercador projection that we see on most printed (or computer displayed) maps
of the earth. There is documentation within the GRASS Web site that provides
a good explanation of the differences. GRASS modules work on geographic
projections, not just coordinates.

   There is a USGS technical report from the mid-1980s that's the standard on
projections. While it is becoming more rare to locatate, see if you can find
a copy.

One other question. New Hampshire appears to fall within two UTM zones
(19T and 20T) Is there a way for a maps set to contain two UTM zones?

   Interesting. NH is a tall, narrow state so one would assume it would be
within a single zone. Regardless, yes there is a way to reproject locations
in one zone on the other, but it's non trivial and I've not done it.

   Oregon is primarily in Zone 10, but the eastern edge (I don't recall the
distance within the state) is in Zone 11. The available DEM and hydrologic
data were reprojected from 11 to 10 by the supplying agency.

Rich

3

On Fri, May 14, 2010 at 6:22 AM, Rich Shepard <rshepard@appl-ecosys.com> wrote:

On Thu, 13 May 2010, Kurt Springs wrote:

This was interesting in that it told me that r.topidx could not be run
with latitude and longitude and I had to convert to UTM. I was wondering
if this is the answer to the problem and I just had to convert to UTM.

Kurt,

Lat/Long represents geographic coordinates, not a projection of location
on a mathematial model of the earth. UTM is the Universal Transverse
Mercador projection that we see on most printed (or computer displayed) maps
of the earth. There is documentation within the GRASS Web site that provides
a good explanation of the differences. GRASS modules work on geographic
projections, not just coordinates.

There is a USGS technical report from the mid-1980s that's the standard on
projections. While it is becoming more rare to locatate, see if you can find
a copy.

I think that Rich is referring to this USGS document, #1395

http://pubs.er.usgs.gov/usgspubs/pp/pp1395

Definitely worth the price if you want to become an expert in map projections.

One other question. New Hampshire appears to fall within two UTM zones
(19T and 20T) Is there a way for a maps set to contain two UTM zones?

Yes. Don't use UTM. In this case use a regional projection that suits
your needs:

1) navigation --> use a conformal projection
2) area statistics --> use an equal-area projection
... etc ...

Variations on the Albers or Lambert (conformal) conic projections work
quite well for large regions that are wider than tall, but for such as
small state should be just fine. We use an Albers equal-area
projection to house soil survey data for the lower 48 states.

Interesting. NH is a tall, narrow state so one would assume it would be
within a single zone. Regardless, yes there is a way to reproject locations
in one zone on the other, but it's non trivial and I've not done it.

I wouldn't recommend it. The desirable properties of the UTM system
(i.e. the fairly good compromise between distortion, preservation of
angles, and preservation of area) only occur within a zone's
boundaries. The farther you move from the central meridian of the UTM
zone, the more distortion you will encounter-- therefore 'projecting'
UTM z10 data into UTM z11 is technically possible, but not a great
idea.

Oregon is primarily in Zone 10, but the eastern edge (I don't recall the
distance within the state) is in Zone 11. The available DEM and hydrologic
data were reprojected from 11 to 10 by the supplying agency.

Hmm...

Dylan

Rich
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

4

Dylan Beaudette wrote:

> There is a USGS technical report from the mid-1980s that's the standard on
> projections. While it is becoming more rare to locatate, see if you can find
> a copy.

I think that Rich is referring to this USGS document, #1395

http://pubs.er.usgs.gov/usgspubs/pp/pp1395

Definitely worth the price if you want to become an expert in map projections.

According to that page, it's no longer available in hardcopy; however,
you can download it as a PDF.

> Interesting. NH is a tall, narrow state so one would assume it would be
> within a single zone. Regardless, yes there is a way to reproject locations
> in one zone on the other, but it's non trivial and I've not done it.

I wouldn't recommend it. The desirable properties of the UTM system
(i.e. the fairly good compromise between distortion, preservation of
angles, and preservation of area) only occur within a zone's
boundaries. The farther you move from the central meridian of the UTM
zone, the more distortion you will encounter-- therefore 'projecting'
UTM z10 data into UTM z11 is technically possible, but not a great
idea.

That doesn't stop Norway doing it :wink:

http://upload.wikimedia.org/wikipedia/commons/e/ed/Utm-zones.jpg

See 32V and 31X/33X/35X/37X.

The SW fragment of NH which is in zone 18 is only around half a
degree; projecting that into zone 19 isn't really an issue. Or you
could just use a custom transverse Mercator projection with the
central meridian at e.g. 71d30'W.

--
Glynn Clements <glynn@gclements.plus.com>

5

On Fri, May 14, 2010 at 5:05 AM, Kurt Springs <ferret_bard@mac.com> wrote:

Hi Folks,

I have been playing with the seamless maps from the USGS. Specifically, my question is on the effect of Latitude/Longitude vs UTM with respect to hydrology modules such as r.watershed and r.topidx. I was trying to generate watershed maps with r.watershed. I've done this before for my dissertation. They did not turn out the same.

Are the region settings the same? Is the resolution of the new DEM
identical to the old DEM? Resolution and extend must be identical if
you want to compare results.

They looked strange compared to what I generated before.

In what way?

Moreover, they maps kept falling in the same place in the map set regardless of what I turned off or on. (i.e. some where in the center top, when the map was generated from a dem on the bottom.)

Was the computational region set to the resolution and extends of the
dem on the bottom?

The current computational region influences heavily the results of any
raster map processing. If current extend and resolution are not
properly set, the results may be astonishing.

Markus M

6

Glynn wrote:

The SW fragment of NH which is in zone 18 is only around half a
degree; projecting that into zone 19 isn't really an issue.
Or you could just use a custom transverse Mercator projection
with the central meridian at e.g. 71d30'W.

... and of course there will be a state plane projection designed
for each state already, for NH they use 71d45'W. EPSG code(s):

# NAD83(HARN) / New Hampshire
<2823> +proj=tmerc +lat_0=42.5 +lon_0=-71.66666666666667 +k=0.999966667 +x_0=300000 +y_0=0 +ellps=GRS80 +units=m +no_defs <>

# NAD83(NSRS2007) / New Hampshire
<3613> +proj=tmerc +lat_0=42.5 +lon_0=-71.66666666666667 +k=0.999966667 +x_0=300000 +y_0=0 +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs <>

# NAD83 / New Hampshire
<32110> +proj=tmerc +lat_0=42.5 +lon_0=-71.66666666666667 +k=0.999966667 +x_0=300000 +y_0=0 +ellps=GRS80 +datum=NAD83 +units=m +no_defs <>

... and several more US survey ft variants.

Hamish