The BNSF derailment at Doon, Iowa

7939 views
341 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    October, 2008
  • From: Calgary
  • 1,704 posts
Posted by cx500 on Wednesday, July 11, 2018 12:15 AM

Murphy Siding
When was the last verified incidense of soil liquefaction causing a train derailment?

I understand it has occurred a couple of times on CPR lines within the past decade, but as far as I know there was no standing water around (maybe some down in the parallel ditch of course).  Probably just deformed subgrade strata retaining elevated moisture levels after a wetter period.  I won't go into geotechnical explanations; a little knowledge is a dangerous thing, as shown by certain amateurs here.

  • Member since
    December, 2017
  • From: I've been everywhere, man
  • 646 posts
Posted by SD70Dude on Wednesday, July 11, 2018 12:41 AM

This fatal accident was caused by a section of saturated subgrade becoming weak and unstable after the water level dropped, due to a downstream beaver dam rupturing:

http://www.bst-tsb.gc.ca/eng/rapports-reports/rail/1992/r92t0183/r92t0183.pdf

Greetings from Alberta

-an Articulate Malcontent

  • Member since
    May, 2003
  • From: US
  • 14,085 posts
Posted by BaltACD on Wednesday, July 11, 2018 6:38 AM

SD70Dude
This fatal accident was caused by a section of saturated subgrade becoming weak and unstable after the water level dropped, due to a downstream beaver dam rupturing:

http://www.bst-tsb.gc.ca/eng/rapports-reports/rail/1992/r92t0183/r92t0183.pdf

If all else fails - blame it on the beavers!  Oops - Sign

         

Never too old to have a happy childhood!

  • Member since
    January, 2014
  • 4,728 posts
Posted by Euclid on Wednesday, July 11, 2018 6:49 AM

Liquefaction requires certain conditions of soil saturation, ground vibration, and soil type.  

https://scholarsmine.mst.edu/cgi/viewcontent.cgi?article=1537&context=icrageesd

 

 

  • Member since
    September, 2003
  • From: Omaha, NE
  • 9,171 posts
Posted by dehusman on Wednesday, July 11, 2018 7:18 AM

Murphy Siding
When was the last verified incidense of soil liquefaction causing a train derailment?

The last one or the last one the NTSB or the TSB investigated?

You probably won't know the last one since it could have been investigated internally by the railroad and cause determined, but the results never released to the public.

Dave H. Painted side goes up.

  • Member since
    May, 2005
  • From: S.E. South Dakota
  • 11,927 posts
Posted by Murphy Siding on Wednesday, July 11, 2018 7:37 AM

dehusman
 
Murphy Siding
When was the last verified incidense of soil liquefaction causing a train derailment?

 

The last one or the last one the NTSB or the TSB investigated?

You probably won't know the last one since it could have been investigated internally by the railroad and cause determined, but the results never released to the public.

 

  I'm just wondering how common it is since that keeps getting thrown out there as a possible reason for the derailment in question. If it's a very rare occurrence then it seems quite plausible that the derailment was caused by something a lot more common and the water was a non-factor.

 

Thanks to Chris / CopCarSS for my avatar.

  • Member since
    November, 2013
  • 902 posts
Posted by VOLKER LANDWEHR on Wednesday, July 11, 2018 8:26 AM

Deleted

  • Member since
    January, 2014
  • 4,728 posts
Posted by Euclid on Wednesday, July 11, 2018 8:28 AM

Murphy Siding

 

 
dehusman
 
Murphy Siding
When was the last verified incidense of soil liquefaction causing a train derailment?

 

The last one or the last one the NTSB or the TSB investigated?

You probably won't know the last one since it could have been investigated internally by the railroad and cause determined, but the results never released to the public.

 

 

 

  I'm just wondering how common it is since that keeps getting thrown out there as a possible reason for the derailment in question. If it's a very rare occurrence then it seems quite plausible that the derailment was caused by something a lot more common and the water was a non-factor.

 

 

Liquefaction is a rare occurrence because the necessary conditions have to be present at one time and place.   Two of the conditions are saturated roadbed and ground vibration.  Certainly, the ideal ground vibration for inducing liquefaction was present.  The third necessary condition is soil type.  This third condition will allow saturation and also directly contribute to a loss of soil particle cohesion.  Saturated roadbed also requires high water inundating the supporting soil, which most certainly was present. 
 
Also, the Doon derailment seems to rule out most forms of soil erosion because that happens before a train arrives, and so the locomotives either derail or at least produce a sensation of support loss.  That sensation may have happened, and we have not been told about it.
 
However, that type of erosion is also visible, and with the constant monitoring that was probably occurring, the erosion would have been detected.  But the conditions for liquefaction are not visible or readily detectable during inspection and monitoring.  Indeed, it would require the train to produce the liquefaction.
 
So not only were the conditions for liquefaction at least somewhat present, but also, it is about the only type of roadbed failure that fits the circumstances.  Therefore, I strongly suspect that the derailment was either caused by liquefaction or by something unrelated to the flood such as a broken rail, broken wheel, broken flange, etc.    
  • Member since
    November, 2013
  • 902 posts
Posted by VOLKER LANDWEHR on Wednesday, July 11, 2018 9:01 AM

dehusman
Before everybody gets too excited about the flood gauge data, a significant detail is that the flood data is not updated on a real time basis, it is generally updated about hourly with a little lag.

The USGS gage data is updated every hour with the current discharge and levels. The delay is about 5 to 10 minutes. Example: Most recent instantaneous value: 1410 07-11-2018   08:15 CDT displayed at 8:21 CDT

As the update is at the same time hour for hour your delay is max. 15 minutes if you know when to look.

dehusman
Since it takes several data points to determine a trend, it could take several hours to even detect that something was shifting in the water levels. If it was a short term event, like a sudden rise and drop, over a short period, lets say 3-4 hours, it could be over by the time the event was detected.

When I read you post I get the impression that this is rocket science. This is observation and simplest mathematics.

Upriver: On June 21, 2018 0:00 am there had been a rise of about 3 ft within the last 10 yours upstream. At about 10 am there was another 5 ft rise. After a few hours pause there was a rise of additionally 3 ft crossing the Major Flood Mark.

Down river: On June 21, 0:00 am there had a rise of 2 ft within 10 hours. In contrast to upriver the level rose constantly without pause by 8 ft within the following 24 hours.

So there was early warning, more than the 3-4 hours you talked about. Everything else is constantly watching.

dehusman
The downstream gauge is a lagging indicator, it tells you what happened, not what is going to happen, plus it lags the upstream gauge by probably an hour or more, being several miles away. By the time you could confirm that there was a difference in the trend and see a change in the downstream gauge, it would be somewhere over 3-4 hours.

the gage is about 4.5 miles downstream but it showed the same tendency at the same time! So there was no delay. For me that indicates that a lot of water came into Rock River after the upriver gage.
Edit: The upriver gage still rose at the time of the accident, so no need to think the situation at Doon was much different than Rock Valley.

dehusman: There are also several tributaries from multiple directions feeding into the main river, in any case you would expect the downstream flow to be larger than then upstream flow, since there aren't gauges on all the tributaries it is pure speculation to assign the majority of it to one tributary.

There are a number of tributaries. Looking at their watershed areas makes my assumption that majority of the balance of the discharge an adjucated guess. 

dehusman: This is all really cool for monitoring an event over a period of days or weeks (done that), but its not real handy for predicting something short term on a real time basis.  After you get 6-12 hours of trend, you can make some pretty good guestimates of where it might go.  They do provide forecasts, but like all weather forecasts they are estimates and they can be off by time and elevation (seen that too).

About 3.5 hours before the accident the Rock Valley gage 4.5 miles south of Doon showed a near record level with a discharge only exceeded twice in 70 years. Do you need more indication that this is not the typical yearly spring flood?

There was ample warning, there was enough time to observe and even a last warning shot 3.5 hours before the accident.

dehusman: Its pretty easy to figure out what caused car crash after the event, but what the railroad's situation is trying to figure out if there will even be a crash, when and where it will happen, before it happens, all while riding in the car.  Higher degree of difficulty.

I have tried to show what information about the flood should have been known at BNSF as basis for their decision how to operate the oil train. I did it afterwards but what I found is available online almost simultanously with event.

I don't try to find the reason for the accident. I think the appropriate care and more Rule 1.1.1 especially with an oil train would have reduced the aftermath. It is an opinion and I might be wrong. I went into this detail as some posters here tried to play down the high water despite the facts.

One can value facts differently, an the BNSF apperently did.

The railroad might have had all the information and come to the conclusion there was no reason to reduce the speed. Hopefully the results of the internal investigation will show up.
Regards, Volker

  • Member since
    May, 2005
  • From: S.E. South Dakota
  • 11,927 posts
Posted by Murphy Siding on Wednesday, July 11, 2018 9:48 AM

Euclid
Therefore, I strongly suspect that the derailment was either caused by liquefaction or by something unrelated to the flood such as a broken rail, broken wheel, broken flange, etc.



     Uff Duh! How could you get any more ambiguous? You're covering all your bases with custard pie or something else unrelated to the discussion.Dinner

     You could just as easily have strongly suspected that the derailment was either caused by aliens from outer space or by something unrelated to the flood such as a broken rail, broken wheel, broken flange, etc. Alien

 

Thanks to Chris / CopCarSS for my avatar.

  • Member since
    January, 2014
  • 4,728 posts
Posted by Euclid on Wednesday, July 11, 2018 10:19 AM

Murphy Siding

 

 
Euclid
Therefore, I strongly suspect that the derailment was either caused by liquefaction or by something unrelated to the flood such as a broken rail, broken wheel, broken flange, etc.

 



 

     Uff Duh! How could you get any more ambiguous? You're covering all your bases with custard pie or something else unrelated to the discussion.Dinner

     You could just as easily have strongly suspected that the derailment was either caused by aliens from outer space or by something unrelated to the flood such as a broken rail, broken wheel, broken flange, etc. Alien

 

 

What in the world do find ambiguous in just two completely objective and practical explanations?  Everybody has been reminding us that the wreck may have been caused by something other than the flood.  That cannot be discounted, so naturally I include it as a possibility.  And yes, that option does include aliens from outer space, but so what?  If it was not caused by the flood, it was caused by something else.  Do I have to list those infinite possibilities of non-flood cause?  

If the cause is flood related, I cannot see any possible cause other than liquefaction.  That is unless the railroad was as dismissive of high water as all their defenders seem to be.  But the defenders tell me that the railroad is just loaded with precautionary monitoring and inspection routines in cases such as this.  So I can't see how they could have been blindsided by the high water; unless of course they were caught off guard by that pesky surge.

  • Member since
    January, 2014
  • 4,728 posts
Posted by Euclid on Wednesday, July 11, 2018 10:22 AM

dehusman: Its pretty easy to figure out what caused car crash after the event, but what the railroad's situation is trying to figure out if there will even be a crash, when and where it will happen, before it happens, all while riding in the car.  Higher degree of difficulty.

VOLKER LANDWEHR
I have tried to show what information about the flood should have been known at BNSF as basis for their decision how to operate the oil train. I did it afterwards but what I found is available online almost simultanously with event. I don't try to find the reason for the accident. I think the appropriate care and more Rule 1.1.1 especially with an oil train would have reduced the aftermath. It is an opinion and I might be wrong. I went into this detail as some posters here tried to play down the high water despite the facts. One can value facts differently, an the BNSF apperently did. The railroad might have had all the information and come to the conclusion there was no reason to reduce the speed. Hopefully the results of the internal investigation will show up. Regards, Volker

Yes Volker, perfectly said. Yes

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy