PDA

View Full Version : Making New Maps from DEMs Successful!!



skrielow
09 Feb 04, 16:09
I have managed to make quite a few maps for ATF from USGS 1:250k DEMs. Here is the link to where I got the US data from:

http://edcsgs9.cr.usgs.gov/glis/hyper/guide/1_dgr_demfig/states.html

In order to use the files, you must download them (obviously! (; ) and then change the .dat extension to .dem at the end of the file. Then in the Power Toolkit, you should be about to import the DEMs without too much trouble. I have successfully done about 10 diffrent areas. The only hangup I have is occasionally the program will crash, but I can just continue where I left off.

I would be more than happy to make maps (within reason) for anyone who asks if this won't work. Just E-mail me at skrielow@hotmail.com.

Thanks,
Stewart

Deltapooh
09 Feb 04, 18:33
This is very interesting. DEMs can be used to represent multiple locations. However, I thought we could only use 1:50K or 1:100K, or is that just for map?

skrielow
09 Feb 04, 19:34
I'm not sure, but I do remember reading somewhere that the files that were supposed to be used have to be in raw format. I think the only theoretical drawback to using the 1:250k DEMs would be the lower resolution, but I made an ATF map of Grand Junction, Colorado and there was so much elevation detail that it would be near impossible for units to move through the terrain. I plan to work on using 1:50K and 1:100K DEMs. I just started with 1:250k because the specified format was 1 degree x 1 degree. The USGS 1:250k is 1 degree x 1 degree straight off the bat. I am a GIS Tech when not in school and I have access to professional GIS Software. I have actually gotten permission from my boss to continue working on this. I will get back to the forum and let you know what I find in the next few days.

Thanks,
Stewart

Check out www.rac.louisiana.edu. This is where I work. You can download GIS data for Louisiana. (FORT POLK!!!)

Pat Proctor
09 Feb 04, 22:51
ATF takes 1:250k USGS format DEM's. As you no doubt know, 1:250k is kind of a misnomer. What ATF REALLY requires is data in 1 degree X 1 degree blocks, at 3 arcsecond resolution (1201 X 1201 datapoints). I am not sure how this assumed the name of 1:250k, but that seems to be standard term for the format.

A 1:50k or 1:100k military PAPER map is required to get the MGRS (military grid reference system) data for the map, such as the number of meters/arcsecond, grid to true-north conversion angle, etc. There are a number of free GIS software packages that will, with a little work, give you the same data (microdem/terrabase comes to mind).

The paper map is also necessary to get the roads, vegitation, and urban terrain data to draw on the map, however.

kbluck
10 Feb 04, 16:26
A 1:50k or 1:100k military PAPER map is required to get the MGRS (military grid reference system) data for the map, such as the number of meters/arcsecond, grid to true-north conversion angle, etc.

If a military map is not available, I believe these parameters can be estimated with acceptable precision. In fact, in many situations they will produce arguably *better* precision.


*************************************

MGRS:

There are free MGRS converters available online. Here's one:

ftp://ftp.mentorsoftwareinc.com/public/freebies/mgrs/mgrscnvrt.exe

MGRS is just a shorthand notation for UTM. The rules are fairly straightforward, except at the ellipsoid boundaries where you'll basically just have to look it up. But you'll probaby prefer to just use a converter.


*************************************

Meters/sec:

You can easily compute this for a sphere using the cosine function. Not perfectly accurate, again due to the oblateness issue, but pretty close, within a couple of meters. If you want to get even closer, use the WGS-84 assumptions and interpolate for the exact latitude:

http://home.online.no/~sigurdhu/Grid_1deg.htm

Either method will probably introduce less error than attempting to measure ticks on the map sheet with a ruler.


*************************************

G-T angle:

The basic issue with mercator projections in general and UTM in particular is that you are projecting a spherical surface onto a flat surface. There is inevitably going to be distortion, but UTM is laid out in such a way that the distortion is minimized and rendered insignificant for typical users of the mapping products it generates, allowing them to pretend the Earth is flat and thus greatly simplifying local navigation.

The key to G-T is the 6x8 degree grid zones. UTM divides the earth into 60 north-south wedge-shaped "zones", each 6 degrees wide, starting at the dateline meridian (180W) and moving East. These zones are further divided into 20 rows which form grid zones, usually every 8 degrees of latitude, but rows are largely irrelevant to G-T angle. Each zone is centered on the half-way point. For example, zone 33, which includes Bihac, runs from the 12E meridian to the 18E meridian, and is centered on the 15E meridian.

Obviously, the edges of each zone are not parallel; they gradually curve inward until they meet at the poles. This is the source of G-T angle. On the central meridian for a zone, the G-T angle will be zero. East of that central meridian it will be positive; west it will be negative. At any point in the zone, the G-T will be approximately the difference between the actual longitude coordinate and the central meridian. The G-T angle will never be greater than about 3 degrees, which will occur at points at the very edge of a zone, since the maximum difference between any point on the planet and the center of its zone is 3 degrees.

Now, this is oversimplified. In fact, this 3 degree max difference actually only occurs at the poles. The edges of the grid zone gradually become more and more parallel as you approach the equator; at points on the equator, in fact, G-T is *always* zero no matter what meridian you are on, since the longitude lines are all parallel at the equator. The sine function comes to our aid. Multiply the difference between the selected longitude and the central meridian by the sine of the selected latitude, and you will arrive at a point very close to the true value. To make it perfectly accurate, you have to take into account the ellipsoidal flattening, but good heavens, we're not writing software to target nuclear missiles here. Close enough.

The important point is that G-T is not a constant value for any given map sheet. It varies continuously as you move around the map in any direction. The best you can do is select a "central" value. Map sheets typically select the value at the center of the sheet. Since scenario maps often are a subset of a sheet, or cross sheets, the mapmaker would be better served by selecting the point at the center of the scenario map instead.

To recap: Select latitude and longitude of desired central point. Subtract grid zone central meridian from the longitude. Multiply difference by sine of selected latitude.

Clear as reasonably translucent mud, I hope.

******************************

You *will* still need some sort of reasonably detailed reference map, in order to determine the terrain cover; urban, vegetation, roads, etc. It doesn't have to be a military map, though, or even topographic; many road maps will get you decent information on this. Even satellite photos available online might serve, if you can accurately line them up. The moral of this story is, don't think you can't use a paper map you might happen to have just because it doesn't include all the cartographic parameters the mapmaker utility asks for.

And, of course, if you're just making a notional map, who cares? Select reasonable values and get on with it.

--- Kevin

skrielow
13 Feb 04, 14:51
Ok here's what I've found. The USGS 1:250k DRGs line up perfectly with the DEMs from the link. I believe they contain enough information to construct the terrain from. I don't have a link to download the entire US data from yet but i will find it.

Stewart

Pat Proctor
13 Feb 04, 23:18
When you do, I would love to take a look!

skrielow
16 Feb 04, 14:08
Here is a jpeg (low quality for size reqs) of a 1:250k DEM overlaid and made transparent onto a 1:250k DRG of the same area. This was done quickly with what I have over here. The area mapped doesn't have a lot of Topo. The lighter the area, the higher the elevation. I have also used this DEM and DRG to make a very basic ATF Map as well. Here is the link I was searching for.

http://www.pipeline.com/~rking/gobb.htm

It contains links to most likely all you would need as far as maps go, to create an ATF map. I hope this helps anyone out!

Thanks
Stewart

P.S.- Thanks Capt Proctor for making such a kick-a$$ simulation!! Especially since it can use real GIS data!!!!!

Deltapooh
17 Feb 04, 14:14
Thanks for the hard work skrielow. I'm really would like to create some maps which simulate various regions around the world (Afghanistan, Sinai, etc). This should help. I'll have to break out my old TI-85 Calculator though to get all the information correct.

Pat Proctor
17 Feb 04, 18:41
Skrielow, I would like to second that hardy thanks. This is an interesting technique I look forward to trying out.

To all:

If ANYONE has rest-of-world DEM's or DTED's, DRG's, or military paper maps they would like to contribute to the cause, we will put them to good use. These materials are expensive and hard to come by on the commercial market, which really puts a damper on the number of maps we can produce for public consumption.

If you are not sure if what you have is useful, contact me at:

pproctor@prosimco.com