I recently watched a very informative webinar by Munich Re. It was their annual webinar on the world’s natural disasters, for 2012 that is. All Munich Re’s webinars are here. I highly recommend dedicating an hour of your time to watch some of these.

Munich Re is the world’s largest re-insurance company, i.e. it’s an insurance company for insurance companies. Because of the nature of that business, Munich Re needs to pay particular attention to natural hazards, since they generate the largest expenditures for insurance companies.

Munich Re distinguishes between the following types of natural hazards:

1. geophysical hazards: earthquakes, volcanic eruptions and the related tsunamis;
2. meteorological hazards: storms
3. hydrological hazards: floods and landslides (they use the term ‘mass movements’)
4. climatologic hazards: extreme temperatures, drought, fires.

I don’t want to get into the details of semantics here and I don’t want to argue with Munich Re’s categories, even though many landslides are part of the geophysical category and most floods and landslides are the results of storms.

Munich Re maintains an enormous (they claim: the world’s largest) database on these events, going back to 1980. Their figures indicate that all but geophysical hazards are increasing in number. The speakers in the webinar aren’t saying it directly, but they are certainly implying that climate change is the root cause of this increase in globally recorded natural disasters. James Hansen (NASA Goddard Space Centre) has always hypothesized that climate change would go hand in hand with an increase in catastrophic weather events, and this hypothesis is now beginning to be supported with evidence. Canadian figures appear to support the trends reported by Munich Re: the Insurance Bureau of Canada, “which represents the majority of private property and casualty insurers in Canada, says claim payouts from severe weather have doubled every five to 10 years since the 1980s” (CBC article here).

But what exactly is a natural disaster?

As part an undergraduate sedimentology and stratigraphy class, I developed a lab exercise on natural events that shape the earth’s surface and that leave a record in the form of sediments. The objective was to help students gain insight in the frequency and duration of events and then to think about the potential resulting sedimentary rocks.

I asked the students to determine how often an event typically occurred and how long the event would last.

Here is my list of events

1. Asteroid impact (on the scale of the K/T impact)
2. Catchment capture
3. Debris flow
4. Delta lob switching
5. Flood basalt eruption
6. Glacial lake sedimentation
7. Hurricane / major storm
8. Pelagic sedimentation
9. River flood
10. Sea level fall
11. Sea level rise
12. spring/neap tidal sedimentation
13. surf on a beach
14. tsunami
15. turbidity current on an active continental margin
16. volcanic ash fall

There was always lots of discussion about this list because of the uncertainty of geographic scale. Were debris flow to be considered globally? Were we talking about the surf on one beach or on all beaches of the world? The idea was to limit yourself to thinking about these processes within a sedimentary basin (after all, it was a sedimentary geology class), although some processes clearly had to be considered on a global scale (sea level fall and rise, asteroid impacts).

Each event had to be given a frequency on a scale ranging from of hundreds of millions (10⁸) of years to seconds (10^-8 yrs). The recurrence, determined for a specific order of magnitude is the Y-axis of the graph. It’s easy to see that a logarithmic scale with only 16 intervals can be nicely plotted on a (virtual) letter-size piece of paper. In addition, each event had to be given a duration, using the same time scale notation. I then asked the students to rank the events from least frequent to most frequent. The duration would follow.

I gave the students the numbers for glacial lake sedimentation: it occurs once a year (during spring melt) and it lasts maybe a few months. That means that the recurrence is 1, (10⁰) and its duration is 10^-1, as a month is more or less 1/10^th of a year. It is important to think about more or less, because when you play with orders of magnitude on a ¹⁰log basis, then it doesn’t matter if a process lasts for example 3 or 5 years: in both cases it lasts less than 10 years and longer than 1 year, i.e. between two successive intervals.

While there were always some interesting differences between what students came up with, they generally (5 classes, 2 different universities, a 6 year period) came up with a graph that looked more or less like this:

My result for recurrence and duration of natural events. 

On the far left of the diagram is an asteroid impact. A big impact, such as the one that intercepted with Earth at what we now call the end of the Cretaceous era (65.5 million years ago) only occurs every few hundred million years. You can argue about how long they last, but you end up with figures between seconds and weeks, but certainly not more than that. Such events do big damage and several of these have been tided to so-called ‘mass extinctions’. Example: the end of what we now call the Cretaceous era is defined by the demise of large reptiles (“dinosaurs”) to the advantage – subsequently – of mammals who took over some of the dinosaurs’ niches. Likely that asteroid impact (the Chixtulub) caused the mass-extinction (a lot of other species became extinct as well), although the outflow of the massive basalts of the Deccan traps (in India) more or less around the same time may also have had something to do with it (such basalt outflows pump nasty gases in the atmosphere). Because of all the spectacular research around the finding of the end-of-Cretaceous Chixtulub crater in the Yucatan, the public and many scientists from other disciplines became more aware of these types of events, and we now send probes into spaces looking out for them, a sort of space traffic warden (Douglas Adams, rest his soul, would have a ball at that idea).  There was a bit of interesting press about asteroid Apophis last week. Apophis came in our vicinity, but is not expected to be anywhere close prior to 2036 (maybe).

At the other extreme, to the far right, is ‘surf on a beach’. Within one sedimentary basin it can go on for tens of millions of years, but each crashing wave only lasts a few seconds. We don’t think twice about it. In combination with rising and falling sea level, beach surf may eventually yield extremely thick coastal sedimentary sequences in the rock record.

In between are a whole bunch of other events (or processes, if you wish).

Again, you can argue about some of the details, but about 150 students came up with more or less comparable graphs.

What does this graph tell us? Let’s try to group these events.

The events on the left have a recurrence that is far longer than any human life span, with a result that humans don’t really worry about them (Apophis mostly caused of a lot of light-hearted banter on Twitter). The events on the right are also not something anyone ever worries or even thinks about much because they simply don’t cause any damage or risk.

But the events in the middle are the ones that worry us. Floods, ash falls, tsunamis, hurricanes, debris flows, and – even if you didn’t think about them – turbidity currents, because they (triggered by earthquakes) may cause tsunamis. These events, in the middle of a spectrum of natural events (and this list is not complete, you can dream up a whole bunch more, but it doesn’t really change the picture), recur at intervals that make them part of oral history, they become part of the human lifespan. They typically occur at intervals of hundreds to a 1000 years and they generally don’t last long, from seconds to days. Terror, in other words, is what they represent.

So what are natural disasters? Natural disasters are normal events, normal from the perspective of our planet. These processes shape our planet and have shaped it for hundreds, even thousands of millions of years.

Even though they are normal, when and where they occur remains the subject of probability calculations. If you live on the shores of the Gulf of Mexico long enough, you will experience a hurricane (I lived there for 6 years and experienced three hurricanes, fortunately they were relatively mild ones). But you will probably never ever experience an earthquake and associated tsunami. So you buy insurance against hurricane damage, but not against earthquake damage.

What MunichRe is saying is that the recurrence interval of the events that already fall within the spectrum of what we call ‘natural hazards’ is increasing. So there is more risk. Prepare yourself: your insurance premiums will go up.