DEM -> contours -> OCAD

Using a USGS digital elevation model (DEM) to generate contours for an orienteering map might be described as "garbage in, garbage out." The USGS DEM is crude at best, at least from the viewpoint of an orienteering map.

But what this technique does offer is the ability to generate contours at a different interval than a USGS quadrangle map. For the Chattahoochee Bend map, I need countours at a 3 meter interval. The USGS quad has 20 foot contours. The object is to take the DEM, run it through some utility programs, and generate contours at a 3 meter interval that I can use a a crude base map.

Every contour must be redrawn during field checking to more accurately match the terrain, but there is an advantage to having contours at the correct interval to start from.

So I got a DEM of the area, for free, from the Georgia GIS Data Clearinghouse. This came in NAD27, which is what I was using with the DOQQ and NAIP photographs, so my OCAD map is already using those coordinates. If I were using some other coordinate wywtem I would have to convert so hat all my data sources are on the same page of the hymnal.

I use the GDAL utilities to do coordinate transformations and to generate the contours. These are command-line programs that are available for almost every operating system. I'm using a mac - not that it matters. The program to generate contours from a DEM is gdal_contour. The command looks like this:

gdal_contour -a contour -i 3 e3384d8.dem whitesburg.shp

The default output is a shape file. Unfortunately, the GDAL utilities don't seem to be able to generate DXF files, which is what OCAD needs. So I had to find a program on the net to do the conversion. I used Arcv2cad, which had a demo version. The dxf file is then imported into OCAD and the contours converted to a symbol. I actually copied the contour symbol and changed the color to gray.

So how does the output look? And how does it compare to the USGS quad contours? Here's an excerpt of the map:

[caption id="attachment_26" align="alignnone" width="341" caption="USGS contours and DEM contours"][/caption]

The USGS contours are in brown, and are at a 20 foot interval. The DEM contours are in gray at a 3 meter interval. So there should be about twice as many DEM contours. Note also that the DEM contours are points connected by straight line segements, and the USGS contours, which are taken from the DRG, are smooth lines.

This excerpt is about 850 meters east to west and 600 meters north to south, so you can see that there is not a lot of elevation chage on this map.

So is this useful? I would like to have both in the field, so that I can contrast and compare. The DEM gives me a clue where the extra contour line should go.

This would not be useful in a map with relief detail that could be obtained with photogrammetry. Photogrammetry would be a vast improvement when there is more relief and there is detail to be discovered. But this is such a "gentle" area that the photogrammetry is not going to reveal much, and comprehensive field checking is needed anyway. It is during the field checking that the detail will be added and the contours smoothed.

NAIP photographs

I am wrestling with the decision of whether to have photogrammetry done or not. It's the traditional way to go. But this map lacks intricate contour detail, so it is debatable that photogrammetry will do any good.

What is needed are recent aerial photographs. I found a new program from the department of Agriculture that is trying to take recent photographs, and I bought two photos that cover most of the park from 2007.

The first example below is a portion of the barerock area. There is a swath of rock features that run across the map. The second is of the northern tip of the horseshoe bend, showing some of the old fields and the new pine plantations. The quality and size of the photo excerpts is reduced, of course.

[caption id="attachment_23" align="alignnone" width="256" caption="Bare rock at Chattahoochee Bend"][/caption]

[caption id="attachment_22" align="alignnone" width="256" caption="Horseshoe Bend"][/caption]

The plan is to load these into Ocad as templates, and print them at 1:7500 with grid lines and other info, then take them out into the field and use them for field checking.

Jan 2009 Update

Spoke with my contact at the DNR this morning to get an update on the park. Last year he had told me to call back in January. Here is what I learned:

- Big surprise - development money has been released, despite the budget problems. They are already widening and paving the road leading into the park. Planning will begin to design the facilities now. It will probably take a few more years.

- They had a lot of problems with the hunting season. He would not give me specifics except to say that the surrounding landowners filed a lot of complaints. Because of this there is a chance this will be the last hunting season, although personally I wouldn't count on it - the hunting lobby is pretty strong.

- Logging will continue a little longer. They are thinning all the pine forests.

I now have permission to go out there and actually make a map, but I cannot go until after Turkey season - May 15th, and I have to call before I go and let him know I will be on the property. This is all because of the complaints they have gotten about the hunting and for liability problems. Not good news, but I have to do what the DNR says.
So I have sent an email to the board of GAOC, informing them of what I have found out and that I now have permission from the DNR to go on the property and make a map. Hopefuly we can squeeze in a meeting this weekend at the A meet and discuss it.