How the Kingston ash spilled

I have long thought that the massive ash spill the TVA coal plant in Kingston was the result of the hard freeze on the night of December 22, 2008. After reading the analysis of the spill, I am even more convinced.
Engineers who studied the event found no specific cause for the rupture, but they did not consider my hypothesis. They looked at weather, but only rainfall, which they determined was not heavy enough to trigger the flow. As with a sand, ash that is too wet or dry flows easily, but the right amount of moisture makes it sticky. TVA regulated water content in their waste ash pile to keep it sticky.
Something caused the pile to liquefy that night, and engineers examined seismological records and train schedules looking for some type of disturbance. Once flow starts somewhere in the pile, it propagates. The initial rupture was catastrophic, pushing a house off its foundation, but it involved just a fraction of the mass that spilled. Most of the spill happened gradually over about an hour, filling in the mouth of a creek and pouring ash into an impounded stretch of the Emory River but lacking the force to travel uphill and cause further harm to shoreline properties.
When there is a hard freeze, ice forms on the surface of mud or soil. The expansion of water as it freezes causes a pressure drop that pulls water from below the surface. As the photo above illustrates, this pressure can draw out a considerable amount of water, which freezes into crystals or ribbons.
The rupture in the ash pile started in the northeastern corner of the pile where the curvature maximized the surface area, providing more exposure to the cold air. The wicking pressure would be strongest in the corner, and the flow created as ice formed on the surface of the pile is just the sort of trigger engineers were looking for. Since they failed to consider the cold weather, they were unable to find a cause.
Why does it matter? Just a couple years prior to the spill, an outside consultant warned TVA of this danger and recommended that they lower the water content in the pile during winter months to prevent a freeze-triggered collapse. They did this the previous year, but TVA cycled managers to other plants during a reorganization in 2008. Apparently institutional memory was lost and the pile was not dewatered last winter.