Thursday 25 August 2016

On earthquake prediction, we’re asking the wrong question!

The recent devastating earthquake in Central Italy. The terrible scenes from Ecuador earlier this year. Nepal last year. And all throughout history.  When earthquakes strike, buildings collapse, and tragically people die, lose their homes, jobs, livelihoods. Their streets will never be the same again.

Inevitably, after larger and newsworthy earthquakes, articles fly around asking, ‘Can we predict earthquakes?’. The most recent I have seen is from WIRED.com which ‘…answers your biggest questions about earthquakes’. The top three are:

1. Can scientists predict earthquakes?
2. But still, given all those computer models and such, why can’t they predict?
3. Are you sure scientists can’t predict earthquakes?

Clearly, questions on earthquake prediction is on our minds...

Furthermore, below is a plot of popularity of google searches for the term ‘predict earthquake’. I’ve highlighted major earthquakes with fatalities immediately prior the major peaks that may have triggered people to google these search terms.


There are obviously other major earthquake where there aren’t peaks in searches, but from this diagram it seems to suggest that peaks in searches do follow deadly earthquakes.

Interestingly, the first peak follows the first major earthquake sequence in Italy following the L’Aquila 2009 event and its surrounding controversial prediction stories. This may be why there was such a large peak in searches. 

Additionally, popular articles following an earthquake may drive these searches significantly – such as the aforementioned from WIRED.com.

Now I understand why this question is important. If we knew when earthquakes were going to happen, we can get out of buildings and avoid injury and fatalities. We might lose our houses, schools, offices, but we escape with our lives and health. But I think there is a much more important line of questioning that is needed:

1. If we could predict earthquakes, what would we do?
2. If we had two months notice of an upcoming earthquake, what would we do?
3. If we had 2, 5, 10, 50 years notice, what would we do?

In the short term, I think we may prepare emergency plans, make arrangements in case our house collapsed, buy insurance if available, etc.

In the longer term, we might assess the vulnerability of buildings, reinforce those that are less safe, make sure that new buildings are built well, etc.

Now, if you live in an area where earthquakes have happened in the past, it is pretty likely that they will happen again. So actually you have your longer term warning. You have your ‘prediction’!

So, let’s get started on preparing properly. Let’s start sorting out our buildings, our construction practices, our emergency plans. Let’s reduce the risk to inevitable future earthquakes, and the devastation and suffering that they bring.

Let’s start answering the right questions!

Monday 8 February 2016

Some initial thoughts from an earthquake engineer on the collapse of the apartment building in Tainan

There has been a lot of coverage of the total collapse of an apartment building observed in Tainan after the earthquake a few days ago. This blog takes a very initial look, using only the information available through news outlets, to try and start to understand what happened to cause this catastrophic collapse.

The structure concerned is the 17 storey Wei Guan Jun Ling apartment complex. Having had a look on Google street view I have found what I think is the building that has collapsed. These are views from the front and back:

View from front of collapsed building (from Google street view)

View from rear of collapsed building (from Google Street view)

It is obvious that the building has overturned in the direction shown in the diagram below. The apartments on the top 14 or so storeys of the building remain (relatively) intact with most of their damage likely to have been caused by the impact of it falling to the ground.

Collapse mechanism of structure
When we come to look at the bottom two storeys of the building, this is where this building's major weakness is clear. The diagram below highlights a possible soft storey (see red box). This is likely to have caused the overturning of the entire structure by either:

1. the collapse of the soft storey, particularly at the front (i.e. the white circled columns); or
2. the difference in stiffness between the apartments above and the soft storey below causing failure at the junction between the two (i.e. at second floor level).
Looking further through photos available online I found this photo (below). The photo appears to show (as far as I can tell) the ceiling that you would have seen if you had been standing on the first floor (on it's side). The walls and columns have ripped off the structure below, leaving steel reinforcement dangling. This gives some weight to the theory of the overturning of the building being initiated by a soft storey failure (see point 2 above). Additionally it is worth noting that the white circled columns above are orientated to offer less support if the building was trying to overturn towards the main road (as it did in the end), which gives weight to point 1 (see above).

http://edition.cnn.com/2016/02/08/asia/taiwan-earthquake/

It is clear that something significant was wrong with this structure. It lies on its side, whilst buildings next door have even managed to keep their windows intact (windows are often the first thing to break in earthquakes). Much of the discussion has focussed on poor construction of the building. The photo below appears to show tin cans in the bottom of a floor slab (although again I can't be sure). Filling concrete with 'junk' obviously reduce the concrete costs on a building site and is obviously a sign of poor construction, possible poor design, and poor building control (i.e. poor checks by government enforcers or designers as to whether the building is being constructed as it was designed). Depending on the extent of this 'junk-filled' concrete, this might also be a contributing factor.

http://edition.cnn.com/2016/02/08/asia/taiwan-earthquake/

It just goes to show that reinforced concrete structures are very complex buildings to design, and construct, and once built they can be very hard to assess for their vulnerability (as all of the weaknesses are hidden inside the concrete).

But the bottom line is, again, that poor construction and lack of enforcement of building regulations continue, unnecessarily, to kill people in earthquakes.