Discussion on paper: Unsaturated Free-Standing Mainline Railway Embankments Part 2: An Example of Handling the Awkward Truth

Philip Pells


The Coroner found that the landslip in an embankment on the Alpine Way at Thredbo, which killed all in its way bar Stuart Diver, was caused by ingress of water into a partially saturated fill, from a pipeline which had opened at a joint.
Specifically the Coroner stated:

“A mass of evidence was led before me which established clearly that … the Alpine way fill embankment which ran for approximately 1.3km above Thredbo Village was in a marginally stable state and extremely vulnerable to collapse if saturated by water.
The landslide was triggered when water from the leaking watermain saturated the south-west corner of the landslide in the fill embankment of the Alpine Way setting off the first stage of the landslide. It was a wonderful piece of field mapping by Tim Sullivan that demonstrated beyond doubt that there were two successive parts to the landslide. The first stage impacted upon the eastern wing of Carinya Lodge. Simultaneously the first stage removed the support of land to its east causing that, too, to collapse onto the lodge below.”

My roles in supporting Tim Sullivan’s factual reports to the Coroner were to determine the shear strength properties of the decomposed granite which comprised the fill and undertaking stability analyses to mimic the two stage failure. This work made it clear that the:

  1. Effective stress shear strength parameters were a cohesion of zero and a friction angle which varied with density, and hence depth in the fill, and
  2. The factors which had prevented slip failures in the embankment over a period of about 40 years were negative pore pressures (suction) in the partly saturated fill, and the reinforcing effect of tree roots.

These are exactly the valid factors invoked by Leventhal and Hull in their assessment of the stability of the NSW Mainline railway embankments. The ‘marginally stable state’ referred to by the Thredbo Coroner. The instability at Thredbo was not because the leakage from the watermain saturated the embankment, it was because leakage caused decrease in soil suction which meant loss of shear strength to the point that the first slide occurred from the south-west corner, diagonally across the slope.