Map showing earthquake hazard for Bhutan, with 2 percent probability in the next 50 years (similar probability as rolling 12 with two dice). Any point in yellow or above can experience shaking such that objects are lifted from the ground. Any point in green or above has chance of house collapse. The light blue areas have the lowest risk, but these are almost uninhabited. (For specialists: the colour show Peak Ground Acceleration, that is, the strongest expected ground shaking. It is shown in units of g, 1g being the free fall of an object due to the Earth’s gravitational attraction.)

A very detailed investigation of earthquakes and related damages that could occur in Bhutan reveals that the situation could be more critical than previously thought.

Predicting future is difficult. However, when it comes to describing earthquake hazard, the probability of earthquake occurrence and earthquake risk, which is the level of expected damage, one has to tackle the difficult task and consider all the uncertainties.

An example of probability in everyday life can be described by rolling dice. If you roll one, normal, cube-shaped dice, the probability of obtaining a 1 is 1 in 6, that is 16.7 percent. This probability is the same for every other number from 1 to 6. If you roll two dice, there is some maths to do: obtaining 2 or 12 has only 2.8 percent chance (1 in 36), but obtaining 7 has the highest chance at 16.7 percent (6 in 36). If you roll a lot of dice, there is more maths to do, but we know the equations.

The exercise of rolling many dice becomes more difficult in the following two cases. First: if the dice are not normal and have different number of sides. Second: if the dice are biased, meaning that the material is weighted different on some sides, which always affects the result and therefore some results happen more often than others. This is illegal in games, but is completely normal in nature. In such difficult cases, the problem cannot be solved with equations, and we need to rely on sophisticated algorithms and computers. 

In a recent project funded by the World Food Programme and carried out by Tom Robinson (UK) in consultation with the Department of Disaster Management, Bhutan’s earthquake hazard was analysed by looking at a few scenarios. This means the author assumed a few particular situations and analysed them. The worst-case scenario shows 9,000 fatalities, 10,000 seriously injured, and 40,000 people who need to leave their homes, which is about 5 percent of the population.

In parallel to this study, our team of five researchers from five different countries (Bhutan; South Africa; UK; France; and Switzerland) have combined expertise to compute Bhutan’s probabilistic earthquake hazard and risk. This means that we identified numerous aspects describing how often, where, and how strong future earthquakes can be. We also used recent soil stiffness, population and building data in Bhutan, as well as how vulnerable the different types of buildings are to ground shaking. For each of these parameters, we rolled dice that are known to be biased and to have different number of sides. With the help of modern computer programs, we rolled many dice, many-many times, and we combined all these probabilities to obtain the results.

The findings show that the maximum imaginable earthquake magnitude is probably 8.9, which means this is a possible but remote. The largest fault in the area, the Main Himalayan Thrust fault, which starts at the surface at Bhutan’s southern border and lies beneath most of the country at 10-15km depth, is the greatest source of hazard. For the next 50 years, there is 2 percent chance that the ground above this fault will shake so strongly that objects can be lifted from the ground. This chance is slightly less than rolling 12 with two dice! In this case, the peak ground acceleration can exceed 1g, and even reach 2.1g (see the map). This is larger than the value used in the current building code in Bhutan for the entire country (0.36g). Our map also shows that the earthquake hazard varies across the country. Even when considering 10 percent chance in 50 years (a bit less probable than throwing 6 with one dice), the expected ground shaking is two times as strong as what is in the current building code. 

Our results also show that the densely populated areas such as Thimphu, Phuentsholing and Paro are at the highest risk. The probability of having earthquake-related casualties in these cities together is 66-79 people out of 10,000 per year. The expected annual economic loss due to earthquake-induced building damage is about 35 million dollars (2.6 billion Ngultrums). Landslides and soil liquefaction can worsen the local damage and are not included in our calculations. Should Bhutan face the worst-case scenario, the damage would exceed what is currently imagined. In case of an earthquake similar to the one in the year 1714, up to 18 percent of Bhutan’s population (approximately 150,000 people) could be affected.

These results come from hypothetical calculations. The worst-case scenario is possible in everyone’s lifetime: it can happen, but it is not the most likely one to happen. However, establishing and enforcing appropriate building codes, conducting public awareness programs, and inclusion of regular earthquake education at school would clearly help to improve the preparedness of people. Moreover, if buildings have seismic retrofitting, they would be less likely to fall during earthquakes. This will not only save many lives but also money over the long term as retrofitting is cheaper than building a new house.

The full details of our study, with all figures and numbers are published in the journal Natural Hazards.

Contributed by 

Dowchu Drukpa, Victoria L. Stevens, Raffaele De Risi, Romain Le Roux-Mallouf, György Hetényi