Switchbacks over Stevens Pass
In the Cascade Mountains east of Seattle the mainline of the Great Northern Railway (now BNSF) heads up the Skykomish River to the headwaters of the Tye River. Just below Stevens Pass it runs through the Cascade Tunnel. It did not always do so. Previously it gained several hundred feet more of elevation to reach Wellington, where it went through an older tunnel. And originally, before that tunnel was built, the tracks switchbacked over Stevens Pass.
Diagram 1 (courtesy of Robert D. West, from his website on Stevens Pass) shows the switchbacks.
On the west side (on the left) the switchbacks fit into a bowl at the head of the valley, giving them their "U" shape. There's one wye on the north side (not counting the branch that eventually went into the tunnel) and three on the south side, plus another one extending up a ravine just below the pass. Each wye has stub track at least a thousand feet long to accommodate the trains. These traverses are now a road, only the the road engineers put in turns. (And these turns suffer from some of the same problems as seen on trail switchbacks, such as being much too steep on the inside of the turn.) There is an odd number of switchbacks, which means that trains would have backed over the pass. But no problem, the three traverses on the east side corrected that. (What looks like a combination wye and turn – next to the "3.5% Grade" – I believe was a reversing wye, and not part of the switchbacks.) Switchbacks got the line over the pass in otherwise impassible terrain, but they were a serious bottleneck in the operation of the railway, and took a lot of resources to operate and maintain. The Great Northern was real glad to complete the tunnel.
It might be wondered: have railways ever built switchbacks with turns? Well, strictly speaking, railways do not consider something a switchback unless they have to "stop, switch, and back". In regular operation trains proceed from point A to point B along an essentially one-dimensional track. How close that track doubles back on it self, or even over itself, is irrelevant. Not, of course, to the railroad surveyor who has to get that track over a mountain range on a suitable grade. If he does not have twenty or thirty miles to do one long traverse (like the Milwaukee Road did, between Cedar Falls and Snoqualmie Pass1) then he may have to fold some of the necessary traverse back onto itself. Which means zig-zags. Since railroaders use "switchback" in a more restrictive sense2, a better form of the question is: have railways ever built zig-zags with turns? This would seem ludicrous, considering the wide radius curves ("you want a landing how many hundreds of feet across?"), but in fact the answer is: yes. That is exactly how the Great Northern went from Scenic (site of the current Cascade tunnel) to Wellington (the old tunnel).
Take a close look at diagram 2, below (also courtesy of Robert D. West).
Contour lines are not shown here, but from the various streams you can get a sense of the overall topography. The old route from Scenic to the Windy Point Tunnel (just above Scenic) clearly doubles back on itself. And just begs for a wye going up Martin Creek. Instead, the the engineers went to some trouble to build two bridges across Martin Creek and a curved tunnel to accommodate a turn. A fine example of what can be done if one carefully considers the possibilities.
Although these two traverses parallel the existing mainline (green line), the physical turn at Scenic does not constitute a turn in the sense of a switchback, as the traverse does not reverse relative to the slope; the turn never crosses the fall line. Ascending to Scenic the route is on a north facing slope, which is up on the right-hand side; past Scenic the slope has turned to south-facing, and the route is still up on the right-hand side.
This area can be visited on the Iron Goat Trail.
Back to Definition of 'switchback'.
Back to Switchback Theory and Principles.
- Also in the Cascade mountains, roughly east of Seattle. The former Milwaukee Road mainline is now the John Wayne Pioneer Trail, paralleling I-90.
- The classic texts on railway engineering, such as Wellington's The economic theory of the location of railways (1905) and Clarke's The American Railway: its construction, development, management, and appliances (1883), extended "switchback" to include any route that doubled back on itself in order gain elevation. The apparent restriction of modern usage reflects an operational viewpoint.
Copyright (C) 2008 by J. Johnson.
Retrieved from the Internet Archive Wayback Machine December 2016.