TransDEM Forum

I've been using TD now for about a month. It is a great program and I have managed several nice routes using either the Map Tile Servers, or Google Earth as the source for 3D tiles. But, in the course of creating some of these routes, I have wondered about why rivers, deep gullies, embankments, and such aren't shown, or are in such slight rises/depressions as to be almost invisible. Is there a setting I can use that will enhance the vertical resolution to let me see, for instance, a river that is 250M wide and nearly 3M deep? At present, I cannot see it and only know it is there by the tile depicting it. (It is not under the tile either).

I have run across a route that, while trying to create it, I am told repeatedly that certain necessary tiles couldn't be created and were "skipped" because the "source" couldn't be found. I am assuming this "source" is the one you get when you click "File name info SRTM .hgt." If this is true, then I have located and added all the ZIP files it said were necessary, but still cannot create two thirds of the route tiles due to this error. Unfortunately, the route covers 4 SRTM ZIP files, so I realize that it will be tricky. In one of my attempts, I drew the masking rectangle many miles outside what I really needed for the route tiles, making sure I included all of the SRTM square. Still, I got the "can't do it" error.

So far, just this one route is doing this. I created the route using the Map Tile Servers, and it worked just fine. Unfortunately, this area (remote Montana countryside) is virtually a blank, white area with one highway paralleling the railroad. Not very good for placing anything. My GE images are much better, but I cannot create them.

Any help would be appreciated.

BIll

G'day hiballer,

There is probably only one reason or cause, Bill, for your observation of that which I will call "a lack of terrain resolution" (..."...rivers, deep gullies, embankments, and such aren't shown, or are in such slight rises/depressions as to be almost invisible..."...) and that is that you have, somehow (contrary to the recommendations in these pages), managed to make use of a DEM dataset, that, firstly, has a low resolution and secondly, that is not "hydrograhically' (I think that's the right terminology) modified. If, as I suspect is the case, you have followed the instructions and you are using the "bog standard" SRTM 30 meter data, which has a "resolution" of "1 arcsec" (1 point of data in every 30 SQUARE meter plot of the earth's surface), I would expect the results you describe. As you appear to be working on a North American based route, the TransDEM Community recommends that, at the very least, you should be sourcing the DEM terrain data from the USGS 'official' 1/3 arcsec dataset, which has a "resolution" of "10 meters" (or 1 point of data in every 10 SQUARE meters), which is perfect to match the internal 'grid' provided in Trainz. This "level of detail" WILL give you the river banks amd embankments you seek but there is a price for this. As I know you have a programming background, you will clearly see that to get this amount of detail, the 'dataset' for the 1/3 arcsec terrain must, needs be, 90 times larger than the same area covered by the 1 arcsec dataset, Download file sizes are subsequently proportionately larger, as well. Actually, this is not the best resolution available. For some areas of the continental United States (and the area covered is constantly updated), there is available, a 1/9 arcsec dataset, which takes the terrain resolution tight down to 1 meter but the use of this dataset for Trainz purposes is not recommended, mainly because the file sizes are too large and it would need to be 'reduced' to 5 meters resolution, although there is built-in TransDEM functionality to achieve this. The second of the recommendations simply means that the radar data has been manually modified to make it 'flat' or level (at the appropriate elevations) where there are known bodies of water - the USGS 1/3 dataset is one such example.

As for your second problem, I can only surmise that there might be issues with the 'quality' of the download for that 'tile' and that those tiles that cannot be sourced have suffered some form of corruption during their download...

...let me know if you need any help with the USGS...

Jerker {:)}

Thanks, Jerker, for the explanation. I figured that the input DEM dataset would have given me this effect of "flattening" most terrain features, but I am only using the suggested SRTM site given in the "Building your first route" section of the help manual (dds.cr.usgs.gov/srtm/). Where do I find these better resolution SRTM datasets?

If I use the new datasets, I can see where I would have to use the 5M Trainz baseboard setting and that will definitely up the size of the route files. That I can deal with by breaking up the route into smaller chunks. The particular area I am working on is only around 30 Miles in length, so I don't think the file size would get too large.

My second problem might be bad input files, but I doubt it. I made a one-time test using a portion of the proper UTM tile area and a straight-line route across it, running around 5 miles long. When I check the "File name info SRTM box, the same one giving me problems was present. When I tried to create the route, I still didn't get the tile created. I can only surmise that the original ZIP file had, in fact, become damaged. I sent a note to the owner of the site and asked if they could check it out. I don't hold much hope for that.

The same, exact, technique on another area of Montana, using GE images, worked very well indeed, so that would seem to support the possibility that the other ZIP file is damaged. And, since those files contain such low-resolution DEM information, I might not want to use them anyway.

Bill

I'm on holidays in Italy right now, with intermittent internet access, so only a rather brief comment from my side.

As Garry has already pointed out, DEM resolution is down to the DEM source. SRTM works for most of the planet, but it's very limited in detail, or, to be more precise, there is no detail at all. I can only underline that for the US you should use NED 1/3 arc sec. Some areas are covered by 1/9 arc sec, but the amount of data will explode. You download NED DEMs from the USGS web site with "The National Map Viewer". We should have a number of threads and postings about it here.

For UTM tiles, the location for anchor points - every single tile needs one - must already exist in the route, i.e. there must be a baseboard for it. For 1000m tiles, this is the southwest corner. For 3D UTM tiles, the terrain to shape the tile is also taken from the route and not from the DEM, and again, the baseboards must already exist.

A remark about map tiles: You do have ortho-imagery (aerial images like GE) data there as well, from Google, Microsoft and for some parts of the world also from the Russian provider Yandex. Yandex uses a slightly different model of the globe, and the maps/images appear to be slightly off when seen in the preview window but will match nonetheless.

Google tends to change the version number in the web address (URL) quite frequently. You may have to manually update it (there are threads on this in this forum. "202" should still work). Google also limits the number of clippings you can download in one go. Microsoft should be fine.

Okay. Armed with that information, I'll try again. My test routes are, in fact, in the US, but my actual routes are in Northern Japan and, as such, I'm having trouble locating anything but the SRTM hgt ZIP files so I will have to live with that. Armed with the GE images, (if I can get them to generate, that is), I know where the rivers/berms/raised road & rail, and such are located.

BIll

G'day geophil,

Thanks, Roland, for popping in there and giving me a hand..

G'day hiballer,

...unless I'm grossly mistaken (and I could all too easily be), there ought to be Aster GDEM available for Japan, this should be their designated Ver 2, which, although it is nominally only 1 aecsec resolution, unlike the SRTM data, it is "hydrographically" modified AND it should also incorporate some modifications from known/existing terrain contours, which should be an improvement (albeit a small one) over the SRTM stuff. You can get the Aster GDEM data from the LPDAAC and the USGS data from here. This link is to their home page. From there, you need to click on the link in the GIS panel that takes you to the "Download GIS Data" option, which opens the National Map Viewer to which Roland refers above (I have found that if you open the TNMV directly, sometimes - more often than not - it doesn't always 'open' properly), calling it from the home page gets it open without any issues. Once you've got the viewer open, all you need to do is navigate around the map on the right to find the area you're "working" in (zoom in or out accordingly) and then click on the "Find Products" button (the blue one in the top right hand corner of the Dataset panel). This will provide a list of all of the relevant 1 x 1 degree tiles (if you have that option selected from the options above the map) and all you have to do is choose the "download" option from the appropriate tiles (as indicated by TransDEM). This download is 'direct'. no e-mail required and no waiting (well, very little, anyway, depending upon the size of the file, which is typically a zipped file at about 500 MBytes and the 'speed' of your connection). obviously, you will need to extract this file before you can use it and this si where it gets a bit complicated. Amongst a number of other files and folders (which can all be deleted with impunity), there is a folder labelled "grd", followed by the 'name' of the tile you have downloaded. If you 'open' this folder using the "Open DEM" option in the File Menu of TransDEM, you will see a number of *.adf files in the list of available files. You can select any one of these (though I typically choose the one labelled w001001) and the procedure is then the same for any other DEM download in TransDEM. As Roland has noted, these instructions are available in greater detail elsewhere in these pages but if you run into any issues, give me a yell...

Jerker {:)}

Thanks for all that information, Garry. I've printed it out and will be delving into it soon.

I am totally frustrated in my attempts to use GE images as 3D baseboards. The whole process seems to be hit-or-miss whether the tiles will be created or "skipped" because TD "can't find the source for tile xxxxxx". I've created the route, and it is properly set up with the terrain, but when I try to overlay it with GE tiles, I only get a few of them. What is stranger yet is that using the same, exact outlining box I used for the route, I can export the tiles, but if I do it repeatedly, never moving the bounding box, I get a different number of tiles. On some runs, I get nothing at all, on others I get 20 out of a possible 50 or 60 tiles. I have to be doing something wrong, but I cannot for the life of me figure out what it is.

Creating a small, 6 GE image segment of my desired route went just fine and all of the tiles created properly. I expanded the GE images with another "ring" around the original 6 images and not a one of the new images was turned into a tile. All of them were 'skipped" because of no source. I can't figure that one out.

My explanations probably don't make any sense because I'm still trying to get a handle on TD itself. I go through tutorial after tutorial, and they work well, but if I try to go off and do one of my own routes using the same procedures, they mess up.I think I'll go back to writing for a while and let this whole GE/TD thing simmer.

Bill

For 1000 m UTM tiles you will need their SW corner to exist, which means the baseboard for this must already be present in the route.

I suggest to export a route with the route filter set to something like 7 or 8 baseboards. That will give you a very wide route but a route with enough terrain to play with for UTM tile generation.

The limiting factors are (as far as I am aware of here in Milan):

• rectangular mask
• actual map/image data
• baseboard for this location in the route the UTM tiles are exported to

Evey condition must be fulfilled. (Exceptions might exist but may actually be a bug.)

Aha! That makes all the sense in the world, Roland. I had completely missed the requirement of the marker in the Southwest corner. When I created my test route, I put the GE marker in the SouthEAST corner. The resulting missing tiles were always to the West of the westmost correct tile. I can see now what that happened.

I will do a small test route and make sure all the preconditions are met. And, if I add more tiles, make sure I put the marker in the SW corner and not the SE.

Bill

Well, that didn't go well. I recreated all my 21 GE images, making sure I had the marker in the SW corner. Then I gathered all the images into a master georef file after converting to UTM and saving. for some reason, when I checked for File Name into SRTM hgt, I was only given ONE zip file, but the imaging covered TWO DEM hgt files, so I loaded the second hgt file and saved as a DEM. Both DEMs were loaded and saved as a master route DEM. I made sure the route bounding box covered what it should have covered, adding a margin around the images to give them a baseboard to anchor on.

The route crated properly (when viewed in Surveyor later, all of it was present) However the tile generating process only made 22 tiles, when it should have been around 65. The error box told me a massive amount of tiles could be created because of "no source" and skipped.

I have to be doing something wrong, but I don't know what. The tiles that are generated are located along the southern border of the route and only 2 or 3 tiles deep vertically and 7 tiles wide. Barely covering a tiny bit of the route.

Physically deleting the route and Scenery folders and trying again with a different-sized bounding box produced either a smaller amount of tiles or no tiles at all. Increasing the northern border of the route bounding box to around 15 miles above the northernmost actual rail line still produced the "skipped" tiles generation notice.

For now, I'm just going to leave it alone. I've gone through the tutorials one-by-one for the Nth time, and they work just fine. I try the exact same thing using my own route and they fail. I have to be missing something but I cannot figure out what.

Bill