On the edge of Richards Bay, Zululand, South Africa is a deposit of very soft muds. Over the centuries, erosion had washed down the clays and silts and deposited them in the saline waters of the bay just off the sandstone cliffs that ring the bay. I tried to sample these clays, but they would flow out of the sampling tubes we had back in the early 1970s. I soon concluded that these clays and silts were not yet really a soil; they were just lots of individual particles in salt water.
Just east of the Columbia River in the territory of the Confederated Tribes of the Colville Reservation, in the state of Washington, is a small flat-bottomed valley. The geologists told me the soft muds that formed the near-horizontal surface of the valley floor were the clays and silts deposited in a lake behind an ice dam that must have existed 10,000 years ago when the Columbia and all that area was glaciated. Again I struggled to get samples of the “soil” that persisted in flowing out of conventional sampling tubes. This was no ordinary soil; it was simply a mix of clay and silt particles hanging around in water at a scary high void ratio. Not yet on the Virgin Compression curve as Professor Jennings would have said.
The Greens Creek Mine is on Admiralty Island, Alaska. I drilled there in the early 1980s. The clays and silts we encountered were very soft. The geologist said they were rocks ground up by the glaciers that coursed up and down the valley and were dumped in lakes formed by ice dams. The geotechnical engineers used terms like meta-stable and thixotropic. Funny stuff it was. It was jelly-like coming out of the sampling tube, but shake it but a tiny bit and it turned to flowing fluid. Again nowhere near that elusive Virgin Compression curve.
All three of these deposits were old. All consisted of ground-up rock and eroded soil. But all three were still essentially fluid, as the lawyer would say. They were just as soft as the 50-year old deposits of diamond tailings I had seen at the De Beers mine. There was a difference though. At De Beers, in a dry climate, the top of the tailings deposit was desiccated, dry and cracked. But scrap away the desiccated layer and beneath were fluid tailings. That dried-out layer was a pretty effective seal to evaporation and further consolidation of the tailings, I concluded.
I had sort of concluded that a crust on the top of a tailings deposit impeded further tailings drying and consolidation when I had helped Professor Jennings examine the Bafokeng slimes dam. Recall this dam had failed, killed thirteen, and the tailings had flowed out of the ring dikes and carried on flowing some fifty miles downstream.
In the last few months on tailings projects from the furthest north of Canada to the tip of South America I have looked at consolidation analyses done to predict how long it will take to consolidate the tailings. The calculated times cluster around fifty to one hundred years.
This is just not the way things happen in nature, is my repeated comment. This is not what has happened to far older clay and silt deposits or younger deposits in far drier climates, I say. There is something wrong with your calcs is my un-expressed thought.
A genius engineer pointed me to the problem. In all the calcs presented to me, the column of soil was divided into equally spaced intervals for the purpose of the calculation. Then they ran their finite elements and finite difference codes that incorporate and use a vast variety of constitutive equations, and they calculated a consolidation time of less than a hundred years.
The genius engineer pointed out to me that the problem is the spacing of the elements in the upper and lower layers. He told a famous consulting company to completely redo their calculations of the consolidation time and to incorporate small elements at the top and the bottom of the consolidating column. I chatted with him and an eminent professor, and we agreed the element should be as small as millimeters even though the column of consolidating tailings was nearly 50 meters.
For you see what happens in practice, and what is not included in the calculations I saw, is that the top and bottom few millimeters quickly consolidate and seal off any outward flow that might, unimpeded, lead to a fifty-year consolidation period. And once that consolidate layer reaches a few centimeters, the consolidation slows down a huge amount. Enough to make my early Holocene deposits still essentially unconsolidated.
So until I see a calculation of the consolidation of a natural old clay deposit or of a new tailings deposit that gives consolidation periods like I know from field experience occurs, I must simply reject the short-time answers. I do not care if they use finite elements or finite differences or even finite strain; nothing counts if they do not begin to replicate reality. For models are a guide to judgment; and when they conflict with observation and judgment we must reject them; or correct them. I am waiting.