Combined disaster scenario frequencies 2

[geotechguy1]

Sorry for the awkward title wording.

Does anyone know of any research or publications into frequencies to assess for combined disasters events

For example, logically assuming a 1/2500 year earthquake occurs at the same time as a 1/250 year flood is probably unreasonable but a 10 year flood occurring at the same time as a 1/250 or 1/500 year earthquake might be more reasonable as an alternate scenario for a ULS event.

 

[EireChch]

I think the most common occurrence would be hurricanes and flooding as opposed to EQ and something else.

But I have never looked into it....

 

[1503-44]

For independent events, you multiply the two probabilities together. 1/2500 x 1/250 = very tiny.

Fukushima happened because they thought earthquake and flood were independent events, but when earthquake triggers tsunami, they are not independent. Hence the greatest probability is (at least) the largest of the two independent values. If you subscribe to the "Where's there's smoke, there's fire" increase that probability even more.

 

[geotechguy1]

Thanks for the comments, especially dependent - independent variables. I had it in my head that you just multiplied them but couldn't remember if there were any qualifiers to that (statistics was not a university strongpoint of mine)

I work in an area where liquefaction assessments are common and the consequences / triggering are driven heavily by the assumed groundwater level. After the recent flooding in the lower mainland / Vancouver area back in Canada I got to thinking about what the consequences of an SLS level earthquake occurring at the same time as a flood would be and whether anyone bothers checking (I don't think they do).

 

[EireChch]

This probably needs more thought, for liquefaction assessment in areas where groundwater is near surface i.e. within say 3m, it is common to just take it at ground level. That may be considered a flood level groundwater level but I have seen it done many times.

 

[geotechguy1]

In large parts of NZ that would probably mean outright dismantling the cities and giving up and moving to Australia so we don't really have the luxury of that

Engineers are equally negligent by being absurdly conservative as they are by being negligent in practice / wrong.

 

[EireChch]

We did this when I worked in Christchurch in 2013

 

[1503-44]

Be careful with calculating the theoretical probabilities of extremely unlikely events, otherwise known as black swans, especially when they are of high consequence. When the data set of the occurrence of those type of events is sparse, meaning that when one happens and its probability is recalculated based on the new data set, it rises, it shows that its probability was originally underestimated. That could mean really big trouble, if it is of high consequence.

 

[GeoPaveTraffic]

With respect to levees and earthquakes, what I see done most often is more of an impact study. Assume that the design earthquake occurs. Determine the damage and/or loss of levee height and the length of affected levee. Then consider how long, number of flood seasons or years, it would take to repair the levee. Review the flood hydrograph/flood elevation curve for the river to the number of years to determine the relative risk to the levee protected area. Based on this risk an informed decision can be made about the need for ground improvement, additional levee section, or other options to reduce the risk.

Mike Lambert

 

[1503-44]

It would seem that the number of years it takes to repair the levee would not preempt a requirement to design for a flood of a given frequency, whether the levee is in a damaged condition, or not.

I mean, if the levee was designed for a requirement of a 100yr flood level, then entered a damaged state for a supposed length of time of 3yrs to repair, you would then still have to consider the consequences of a 100yr flood on the damaged levee, not what might happen with a 3yr frequency flood level. Your repair period would have to last longer than the original design frequency requirement in order to change the original design flood level, but then it would be even higher. In other words, if your levee gets damaged by earthquake, you must still consider the effects of a 100yr flood, because a 100yr flood can occur in any given year with equal probability to the next. What changes in that equation is that you will have an increased probability of actually getting a 100yr flood, because you will be exposed to risk for three 3 years it could take to repair it. "What is the probability of getting a 100yr flood at any time within a three year repair period?", is not the same as, "What is the probability of getting a 100yr flood during any one year period". If you consider a repair period, the probability will be higher, but for the same 100yr frequency event.

 

[cvg]

levees are designed for stability under flood conditions, rapid drawdown and seismic. they need to have freeboard required for the design flood conditions. they need to pass this analysis in order to be certified. if they don't, than no certification and no accreditation by FEMA. if damaged resulting in loss of the required freeboard, then they need to be repaired and if not might be de-certified

 

[1503-44]

That makes sense.

 

[geotechguy1]

Eirch:

"Care should be taken to understand the effect of
compounding conservative assumptions when selecting
groundwater levels for the liquefaction assessment.
For example, there is a low likelihood (much less than
1 in 500) that a 500-year earthquake would occur at the
same time as the groundwater level is at its highest
seasonal level. Accordingly, it is often appropriate to
assume an average (median) groundwater level as the
primary analysis case for liquefaction categorisation. "

Copied straight from the MBIE Liquefaction Guidance. A sensitivity analysis might be conducted by looking at the effects of groundwater at seasonal high and seasonal low but assuming that a groundwater table within 3m of the surface is at the surface is absurdly conservative and I don't think many or any consultants would assess it that way. The issue with GW levels in NZ is more that it seems they have a historical preference for not installing piezo's or doing any monitoring.

 

[EireChch]

MBIE Liquefaction Guidance was not released by the time I was gone from Christchurch. The guidance seems appropriate, but as you have said, how do you know what the potential seasonal variation without standpipes etc. Literature I suppose, or I imagine there is a good handle of groundwater nowadays with all the investigations being done.

We were measuring groundwater from hand auger boreholes, so while it is accurate, its a tiny snapshot in time. If we came back and drilled a borehole in winter it could be 1.5m higher for example. We and other consultants adopted this approach of ground water level at surface or near surface. We were definitely conservative but I worked there shortly after the last big EQ so conservatism was only not an issue but nearly desirable.

 

[Milliontown]

As other have said it is as simple as multiplying the probabilities. Coincident loading can also be calculated using the US Army Corps of Engineers Risk Management Center spreadsheets. They have a lot of risk related calculation spreadsheets for use, including coincident loading, or other loadings like gate failure.

The Joint Loading spreadsheet will calculate the probability of various joint loadings.
[URL unfurl="true"]https://www.rmc.usace.army.mil/Software/RMC-Toolboxes/Risk-Calculations-Suite/[/url]

The Seismic Hazard Suite can also calculated various seismic failure modes such as site classification, seismic hazard curves, and liquefaction risk.
[URL unfurl="true"]https://www.rmc.usace.army.mil/Software/RMC-Toolboxes/Seismic-Hazard-Suite/[/url]

 

[geotechguy1]

The USACE Risk Management Spreadsheets look superb - thanks for sharing!