Looking for help with TransDEM and Google Earth maps.

[gisa]

As a new transDEM user who is a tad dense, I must agree smileyman: this was VERY useful and well-put! Even though Versitek forgot to thank you, allow me to thank you for him and allow ME to THANK YOU very much for writing it. Even though I can get topographic maps for the entire route I want to model, google earth UTM tiles will be very handy in the towns I want to model (especially one that has a sizeable yard in it!).

:wave:

Gisa ^^

 

[cascaderailroad]

I bring up this topic: DEMs are a real pain to lay gradients on, most DEM's are +/-20' wrong in the x-y-z elevations and track and river placements are way off more than that. I am modeling 5 county mountainous area between Huntingdon and Johnstown Pa. Without a DEM it would take me years to make all the mountains. So I am using a DEM.

However: If I were modeling a relatively flat area like Boston, LA, NEC, Kansas, Nebraska, Wyoming...etc...I would choose flat baseboards and use GE and Topo map images overlayed on 1km x 1km Basemaps. Unless you are really good at making DEM's and get highly accurate results...sometimes a DEM is a bit, not needed, and overkill ?

 

[geophil]

Or is it the maps in use which leads to the errors observed?

I have found NED DEMs being very accurate. 1/3 arc sec often even include cuttings and embankments - and right on the spot.

NED is terrestrial as opposed to orbital SRTM. SRTM tends to be slightly wrong on elevations in built-up areas or woodland. Horizontal error for a DEM, however, is a very rare thing. The only ones where I have seen this were some isolated ASTER granules.

Maps are a different story. Maps and DEM must fit. This usually requires a larger scale for the maps. Smaller scale like TIGER (1:160000) or improper georeferencing will produce misalignment. TerraServer USA also has some smaller local areas with georeferencing errors (unless fixed in the meantime).

geophil

 

[gisa]

I bring up this topic: DEMs are a real pain to lay gradients on, most DEM's are +/-20' wrong in the x-y-z elevations and track and river placements are way off more than that. I am modeling 5 county mountainous area between Huntingdon and Johnstown Pa. Without a DEM it would take me years to make all the mountains. So I am using a DEM.

However: If I were modeling a relatively flat area like Boston, LA, NEC, Kansas, Nebraska, Wyoming...etc...I would choose flat baseboards and use GE and Topo map images overlayed on 1km x 1km Basemaps. Unless you are really good at making DEM's and get highly accurate results...sometimes a DEM is a bit, not needed, and overkill ?

You raise a good point. For something very flat, I suppose one could go without using a DEM. But then I think that no land is exactly flat. Gentle hills, ravines, slight depressions...those happen naturally. Trying to model a large hill over a long distance is a PITA in surveyor (i.e. laying several large splines, lining them up, smoothing them and then deleting them) and I'll never go about building anything that's not flat without DEM's again (and that includes if I was to do something fictional).

As well, DEM's are great for accuracy and in general, the hardcore/nitpickers types (I suppose I'm becoming one) opt for that wherever possible.

My 2 cents...

:wave:

Gisa ^^