How do I make grades on a DEM?

I am using the Sand-White DEM for a route, and at one part there is a very steep, uneven grade where the track needs to go. How do I make the grade stay realistic, and what is the maximum realistic shortline grade? I think it is around 1.5%.
 
I lay the track and at specific points, I measure the height in Google Earth application (not the Google Maps on the web). I then type in the height in those locations I measured into Trainz Surveyor spline height and apply the height. Once that's done, I smooth the ground under the spline in the normal fashion. The reason I do it this way is there is still a lot of noise on a DEM regardless of its source, and what might be a slight bump can sometimes be exaggerated or at other times a nearby rock or outcropping, or even buildings and bridge abutments, will actually alter the height of the track. By measuring these specific points, it removes the sudden up and down grade changes found on the DEM and makes things more true.

You will notice as the track is laid like this, the line will actually make sense as it rolls every so slightly between the up and low grades, and is less like waves on the ocean.

Short grades are usually in that range at the maximum on main lines especially back here in the east. The B&M had no less than a 1% grade from Deerfield to North Adams, including the Hoosac Tunnel. On some industrial branches, they can be a bit steeper. Keep in mind the kind of consist that's going to run on the short grades. If it's strictly passenger commuter trains, like New York subways, then the grade can be quite steep, even more than 1.5%! If the line is a short industrial spur, it can be high as well, and usually so because only one boxcar or other freight car was moved on to the spur, and not a whole train trying to make it over the mountain. Short lines, more likely you mean narrow gauge lines, can have substantial grades more like 2-4.5% in some places, or even more.

The best way to get the most accurate grades is to measure them and/or to use a track chart for the route you are building. To measure a grade, it's essentially the old slope formula of Rise (H) / Run (Length). You can get an approximate distance in Google and use that on the height, which I found to be quite accurate compared to the DEMs I have been working with. The heights within some points on my DEMs has been within a couple of meters in most places, and the difference, unless there's an anomaly in the data image, is usually due to the resolution of the Trainz grid.

John
 
Lay straights where the straight-aways go ... and lay curves where the curves would be.

Then sink the spline points by a little less than a meter deep into the soil (just barely visible, or slightly below).

DO NOT press the "smooth spline" tool button, until the entire grading is done, SEVERAL MONTHS FROM NOW !
As you may be changing your gradients many times before you get it exactly right. And pressing the smooth spline tool button will permanently deform the terrain.

Unfortuneately a DEM terrain can be way off by +/-20m or more in all the x-y-z areas, due to trees and schwubbery. And too the lines drawn on the terrain may have river channels way up on hillsides, and railroads and roads way down in river channels. These lines are just a rough guestimate, and can be way off by 100m in places.

On a long straight lay a ruler from one spline point to the other ... now lay another 2 rulers from the center out to each end spline point(s) ... continue doing this until you have @ 1/4 mile equal distances between all the straight track spline points. Now continuing sinking the spline points till they are just barely under the surface.

A gradient should never be over 2.00%, and should be less than 1.75% at maximum (unless it is a rare backwoods Shay line). Most of my gradients are well under 1.50%, and most often they are only 0.10% to 0.20%, with just slight leaps and dips. I try not to lay anything over 0.75% to 1.00%.

When I lay a 0.00% bottom of the grade section, then go uphill at a +0.10%, then my next will be +0.20%, followed by a +0.30% ... then a short 0.00% section, then the reverse downhill at -0.10, -0.20%, -0.30% ... etc ... etc ...

I spent hundreds of thousands of hours, grading a 4 track mainline that is only @ 45 miles long, and changed the spline point height many thousands of times, before pressing the "Smooth Spline" tool button one year later. Trust me ... Don't rush cutting the cuts, and filling the fills. You will get used to driving trains through sand dunes that block the track.

Sometimes grading a 5 mile section will be a one night breeze, only to be fouled up, and thrown way off by the next 5 mile section of track, that takes 5 days to grade.

When laying a gradient, you can type in the gradient numeral, and keep on going ahead to the next section ... but when you type in 3 gradients, fall backwards down the track, and check the the last 3 spline points were not moved by the ones ahead. And this DOES HAPPEN almost all the time !

Keep checking and re-checking, the gradients constantly, making sure they are not thrown way off.
 
Last edited:
what is the maximum realistic shortline grade? I think it is around 1.5%.
Depends on the era, country/region, and even perhaps the railroad itself.

Ruling Grades (the longest, steepest portion of a climb) over 2.0% are rare, especially on modern railroads where the line has been surveyed and relocated with time.

In the Rocky Mountains of the US, many railroads raced through during their initial construction (going for the golden spike...), resulting in steep grades as high as 4%. These however were mostly rebuilt in the following decade using grades of 2% to 2.5%.

MRL's Bozeman Pass in Montana is 1.9% as it runs today, though it originally challenged trains with +2.5% grades and a switchback when it was first built.

The Chicago, Milwaukee, St Paul & Pacific had ruling grades of 2.1% and 2.2% on their passes in Montana, Idaho, and Washington until the road's abandonment of it's cross country mainline in the late 70's.

The Durango and Rio Grande's Tennessee Pass in Colorado had a westbound grade of 1.4%, while the climb for eastbound trains was 3%. That line was downgraded to secondary status with the D&RG's merger with UP and friends, and was later abandoned in the sometime in the 1990s.

There are also rare exceptions in the case of logging and mining railroads that tracks climbed remarkably steep grades, such s +6%. These lines tended to just geared steam locomotives like the Shay, Climax or Heisler.
 
I am using the Sand-White DEM for a route, and at one part there is a very steep, uneven grade where the track needs to go. How do I make the grade stay realistic, and what is the maximum realistic shortline grade? I think it is around 1.5%.

There were many sections of Class I tracks here in Central Appalachia with grades greater than 1.5 percent. For example, the Virginian Railway had a 13.4 mile long grade on its mainline called the Clark's Gap, on which the grade for the first 2-1/4 miles was 1.21 percent and from there to the summit at Clark's Gap the grade increased to 2.07 percent. There were also numerous branch lines built by the Class I railroads that had grades up to 4 percent. The "common carrier" short line railroad that ran through the town were I live, that operated until 1940, had grades up the 2 percent in sections. On the B&O's mainline through the Allegheny Mountains of Pennsylvania and Maryland, the 100 mile long Sand Patch Grade has grades of .8 to 1.98 percent.
 
Last edited:
I am pretty sure my grade is ok, when I look at the train moving I can't tell it's going uphill, even though there's an increase of 5 meters from where the grade starts. I want to know how long the stretch of track is, but the panel with signals and the odometer won't show up in quick drive. How do I get it to show up?

EDIT: It magically shows up now. The ruling grade now is a short stretch of about 4%, but it will be fixed. The route is about 3 miles so far. Maybe I can just pretend it was a PRR branch line that PC and CR held onto for some reason, and NS spun it off. It now interchanges with NS.
 
Last edited:
Back
Top