30 January 2017
The human cost of landslides in 2016
Since 2002 I have been collecting data on the human cost of landslides, allowing individual years to be analysed. This dataset underpinned my paper in 2012 that presented the annual cost of landslides in terms of lives lost (see the accompanying blog piece that I wrote at the time). Melanie and I have a new analysis for the period 2004 to 2015 in review at the moment.
The dataset for 2016 suggests that it was a bad year in terms of the number of landslides, but that the number of fatalities was not exceptional. In total I recorded 444 landslides worldwide that caused loss of life, of which just five were triggered by earthquakes. The human cost of these landslides was 2,250 people, 10 of whom died in landslides triggered by earthquakes. By comparison, in 2015 I recorded 345 fatal landslides, causing 2,376 fatalities.
The graph below shows the cumulative total through the year of the number of landslides that caused loss of life (excluding the earthquake induced events) and the number of fatalities that they caused:
I have added to the graph a simple line between the number of fatalities / landslides at the start and the end of the year to illustrate the patterns. This shows that, as usual, in terms of the number of landslides the year started slowly but accelerated through the northern hemisphere spring and summer, and then declined through the autumn and into the winter. This is the pattern that we see every year because the Asian monsoon dominates the statistics. The fatality graph is a little more unusual though, with the peak period occurring between days 120 and 150 (i.e. in May). This is earlier than in most years, and reflects a particularly active early monsoon in South Asia.
In the graph below I have compared the human cost of landslides in 2016 with that of 2015, showing both the number of landslides and the number of fatalities. 2016 data is in black and grey, 2015 in orange and yellow):-
The differences between the annual patterns are stark. Throughout the year I recorded far more landslides that caused loss of life than in 2015. Only in the first and the last two months were the rates of events (i.e. the gradient of the lines) similar. But the number of fatalities in 2016 are dominated by those events in May, which we didn’t see in 2015. However, later in the calendar year, 2015 saw very major events that caused a big ssteps in the number of fatalities. We didn’t see these events in 2016, meaning that the curve is much smoother, and the final tally was lower.
Of course we now start to worry about the human cost of landslides in 2017. The first month of the year appears to be very similar to both 2015 and 2016, but as the graph above shows, January is a poor guide to what will happen.