Nancy Map Progress

[sweeteye]

I have just been doing the preliminaries for setting up a game map of the Nancy area. Before I go too far with it I am wondering if somebody might be able to tell me if a 100 meter contour interval is too much or if it really matters? Here is a view of the map I am using for an underlay.

[Lieste]

100m seems high - If the whole map is steep it won't matter at all, but if you have plateau separated by steep escarpments then you'll get too much 'random' relief in the low-areas and lose much definition overall.

It will still work, but it would work much better with an explicit DEM with per-cell heights over an arbritary range (-500m to 7000m should cover it ) replacing the engine derived DEM-from-contours.

In a way, I'd actually like the terrain representation (map) to be a simplified version of the DEM, in the same way that the MAP is a simplified version of the world, and that a perfect representation between the two isn't a requirement - but I guess that would lead to a lot of bug reports [:D]

Particular things to watch out for... be very careful of narrow(ish) stream beds at 50-99m, these will sometimes be interpreted as narrow gorges at 0m height - (not always unreasonable for the 0-50m cases, over even for a few of those in this group), and become impassable. Better to distort the heights by closing the stream bed at 100m as it becomes too narrow to form a good Height Map - example can be seen in the Thermopylia Map in the lower centre (out of most of the action, so unimportant). Also watch the map edges - these often 'droop' more than is desired where the map is 'flat' - not a problem where the course of action doesn't use them much, but could affect operations where the map edge is close to the mission areas.

You seem to find the map making relatively fast and 'easy' so if you fancy having a go at it, by all means try... if nothing else, having the source material prepared would help when redoing the map/scn under new versions in the future.

[sweeteye]

I do have a program to convert DEM's to contours but I am having trouble finding free DEM's of Europe. Perhaps I am looking in the wrong places. I really would prefer more contour lines as it makes the map more interesting to me.

[Lieste]

Ah - is the problem one of height range (ie 1600m height range over 16 contours) or of map data - you only have a map with 100m contours for an area with (say) 400m height range?

I assumed the former, but I'm not sure from your reply if that was correct?

[sweeteye]

It is the latter, 100 meter contour interval from 100 to 400 meters

[Lieste]

If you have suitable software (I use Global Mapper, albeit an earlier version than current), then you could do worse than snaffle the relevant tiles from CGIAR

Each is 5x5 degree geoTiff or Acsii DEM (SRTM based, with void-fill, including relevant input from Viewfinder Panoramas for high relief areas)
srtm_37_03.zip 'Paris' 45-50N, 0-5E
srtm_37_02.zip 'Calais' 50-55N, 0-5E
srtm_38_03.zip 'Karlsruhe' 45-50N, 5-10E - This one includes the area for Nancy, but the others are relevant to late 1944/45 operations in general.
srtm_38_02.zip 'Bremen' 50-55N, 5-10E

General Link to the Search/Download Page: I forget if you are asked to register or not, but not terribly onerous - Commercial use might require a check first, but personal use is fine.
http://srtm.csi.cgiar.org/SELECTION/inputCoord.asp

Otherwise use the 1x1 degree non-void-filled SRTM tiles from USGS, or the individual tiles from Viewfinder Panoramas (if voids are a huge problem in an area he has probably 'done something' to the cell) - also has (albeit with artifacts due to method) data for areas falling outside the SRTM mission (Iceland, Northern Europe/Siberia above 60N)

[Lieste]

Some caveats:

The DEM data is actually a DSM, so includes heights of trees/buildings in already - to some degree, but the edges are smoothed by the ~90mx60m cell size (at this latitude), so I wouldn't worry about the difference, except to be aware of it when looking at the contours generated...

The data has some noise, not usually significant, but do be aware of it in flat valley bottoms, where it can exceed the cell-cell variation in height due to slope. Smooth the contours in the noisy transition in these areas by eye - particularly as CO sometimes does odd things with relatively small areas above/below contours. If in doubt, remove all islands...

If you do need a legitimate depression, then split the higher zones surrounding it, the elevation layers don't support holes, but can be overlapped to form a hole 'outside' their component areas.

[sweeteye]

Excellent Lieste! I believe that is what I am looking for. Your zips will save me some valuable time. Thank you very much.[:)] I did not think USGS has european data. Going to have to spend some time looking at data. I spent a lot of hours making maps for a 3D American Civil war game. Was very frustrating gamemap making and only one out of four turned out very good. I like this 2D game better for map making not to mention I grew up playing the Avalon hill board wargames and am very comfortable with the 2D game and counter type units. I like to get things done as quickly as possible and still be reasonably accurate.

[Lieste]

USGS:
http://edcsns17.cr.usgs.gov/NewEarthExplorer/
Enter "Nancy France" in the address place, or alternatively enter your bounds directly in the co-ordinates window(s)

Now select datasets, and open and select:
Digital Elevation, SRTM

You might also want the water masks (but they are very 'blocky' and only suitable as a guide to identifying the (current) lakes/rivers) - IMO better results can be obtained by searching for zero slope in the DEM, with user input to the flood fill - I can discuss this at more length in a new thread if that is interesting - it is a PITA that no decent water masks exist, and I'm making my own for DEM that I work with frequently - especially since so many interesting theatres include water filled depressions, and sea level ocean, so a single assumed sea level kills so much relevant data - and there is no clean/simple way of also marking the elevated lakes with the same 'water' type:
Digital Line Graphs, SRTM Water Body Data

Additional Criteria allow the choice of 1 and 3" data, but only 3" data is released for non-US mainland, so you can ignore.

[Lieste]

For accuracy, you will want to ensure that you revert any recent hydro developments - set the river width back to 'original values' and redraw the affected/flooded areas using period maps, if possible. Other changes to heights should be minor - large open cast mines and nuclear power cooling ponds are probably the most frequent 'significant' changes.

[sweeteye]

I am downloading the Karlsruhe zip file at the moment. Will take a look at it and see what I can do with it. Should be able to get the area that is like the map I posted above. Between the two I should be ready to roll. Location of orchards, broken ground and swamps are the only real info I should be in need of. Download is going very slowly... nearly to the point of not moving... going to close out all windows and sign out for a bit...Thanks again...will report back.

[Lieste]

yes, it can be 5-10kb per second. Usually they complete though... I should have remembered. I always find a good book/movie/radio show/coffee and it will be waiting when I've done [8|]

[sweeteye]

OK...Download worked fine and I am able to open the .tif file using my 3DEM program. Fell asleep while waiting but coffee is on and I am ready to go with it. Should be able to make a very accurate map. Any contour interval you would prefer?

[Lieste]

Pick the smallest x10 or x20m interval that will fit the entire map - if you intend a series of closely related scenarios in the same general region, then try to plan the interval for all maps to allow them to be joined (ie keep the same common 'base' elevation and intervals). You can cheat the upper limit (see COTA, Olympus map) if the mountains were impassable and fighting on foothills and lower reaches only.

Widely spaced maps can take advantage of smaller intervals (or if you are prepared to make additional contour sets if you want to join additional areas)

I'd suggest making the interval still a x5, x10 or x20m depending on relief - or possibly 50' or similar, so that suitable real-world mapping can be used for any patching you need to do.

I have no preferences myself - but it makes sense to try to find as close to a common set as possible within the 16 layer limit - this makes most scenarios play similarly - if one map uses a masking data set at 2m intervals there will be fewer long range shot opportunities than the same area at 20m, or 100m.

[sweeteye]

I kind of like the idea of either 20 meter or 50 meter intervals. 20 meter intervals would look better to me but 50 meter intervals would take less time. I will debate this for a little while while I am setting the map up for an underlay.

[Lieste]

Remember, you are limited to 15 intervals, so 16 heights - I'm unsure exactly what extents you have for your map, and therefore can't just query the data for the height range, but a 300m range is on the limit for 20m contours.

[sweeteye]

Good point to consider. I was thinking of going with 20 meter intervals but may have to go with 50 meter intervals. The Jeandelaincourt map is at 50 meter intervals so maybe that is a plus as far as joining the two maps together. Am looking at google earth now to get the coordinates.

[Lieste]

You might consider a 25m interval and add contours to the JDC map if you come to join them later.
Your call ultimately, as you are doing the work, but that might be a good compromise if the JDC would fit into the same 25m contour ranges.

[sweeteye]

Here is what I came up with so far. Had to do a lot of trial and error working with the the three programs I have to create contour maps. Biggest problem I have is when I zoom in on the contour map the quality of the lines gets worse. They get very "blocky" looking and would be difficult to trace.With a color filled contour map I may be able to trace the outlines better. Any way I look at it it is going to be a chore to do.


Satellite view


color contours


black and white

[Lieste]

What are your choices for projection/coordinates/contour interval? I'd be happy to see if my software can do any better. (Either as a single multi-contour file, or as 'n' files at selected heights - this might be easier to work with?)

I'd suggest that Lambert I (nord) or UTM would make suitable projections - the former is the IGN mapping standard, the latter a good 'universal' set if no particular source map must be matched directly. Any additional maps can of course be trivially reprojected if they have varying projections, but it simplifies things if you set a reasonable common standard.

Coordinates can be centre/dimension, or corners.