Understanding Climate Risk

Science, policy and decision-making

Provoking comment and framing risk

with one comment

A couple of my recent forays into the media have provoked comment below the articles themselves and in emails sent querying particular points. They are worth unpacking because they reflect on the different between the straight communication of science and framing risk.

One was in reference to a recent op-ed in The Age. In it, I said:

If people accept the 0.0038 and 0.02 degree benefits as valid then they also accept the science behind a 5.3 degrees warming for business as usual (As in the emission scenario created by Treasury for the 2008 Garnaut Review). Who wants to live in a world warming by 5 degrees or more? Major food crops could not be grown in many parts of the world, projected sea level rise would be tens of metres, most of the shelled species in the ocean would not survive, ecosystems would be disrupted as the pace of change outstripped their ability to adapt and millions to billions of people would lose environmental security leading to mass migrations never before seen.

That prompted an email from an earth scientist wanting to know what peer-reviewed reference I was using for the projected tens of metres of sea level rise. I sent back this now famous diagram and a note saying that I wasn’t putting it on a timetable. He then replied suggesting that people could be misled into thinking that the date was 2100 (because that was tied to the two temperature measures) and that I was being alarmist. Because it would take thousands of years to be realised.

Mean global temperature and sea level with respect to observations (WGBU after Archer 2006)

This comes back to the framing of risk. Those who wish to discount the risk of large-scale sea level rise would prefer it be said this way: “Well yes, it could happen, but thousands of years into the future.” This brings up the issue of hyperbolic discounting – an uncertain risk is moved way into the future. Not to be explicit about the time frame is alarmist.

But is it? Let’s accelerate the cause and effect sequence. If I decide to push a grand piano off a tall building, when does that become a risk? When I decide to do so, when the piano is pushed off the building or when it is about to hit the ground? Quite clearly there is a risk when I decide to push the piano off the building but whether I will do so  is uncertain.  When the piano leaves the building it is committed to hitting the ground. If I was to wait until the piano was about to hit before warning people (setting aside the fact that I pushed the thing off), I would clearly be negligent.

Greenhouse gas emissions establish a risk. Sufficient emissions to produce 5 degrees warming by 2100 would commit large ice sheets to melting, ultimately leading to tends of metres of sea level rise. This could be avoided only if we could induce sufficient cooling to remove heat from the polar oceans around Greenland and Antarctica. If Earth warmed by 5 degrees I seriously doubt whether humanity would have the capacity to apply the technology required to reverse the process. Why not? Because all of the other damage to the planetary, social and economic support systems required to sustain billions of people. Natural processes by themselves would take many thousands of years, operating on a geological time scale.

Alarmist? I think not.

Looking back at my word choice, it may have better to have said sea level rise committed to would be tens of metres. I did think about this, but was not sure whether the term committed would be understood, and did not have the word limit to explain it (it was probably okay to have done so).

Another set of commentators on The Conversation were also provoked by some of my risk statements. In response to this comment:

Since opponents of the carbon tax often like to try reduce the issue to economic terms, why not contrast the $10 billion figure (a spurious one, as you point out) with the net $ benefit to Australia of a .0038 degree reduction in 2100. This would include the marginally reduced cost to Australia of dealing with weather-related catastrophes due to a warming climate (such as megafires, droughts, floods, sea level rise) not just in 2100 but in all years leading up to it and for say 100 years after. I suspect you’d end up with a figure much larger than 10 billion (setting aside for the moment the ethically thorny issue of discount rates).

I said this:

…very difficult to do, and I could have added some stuff on the economic benefits of avoided damages which would have made the story a bit more complex. Existing cost curves are very likely to be underestimates because they don’t do climate extremes very well. Even on those, the $23 per tonne (levy price) comes out as a much larger benefit with social discount rates. This is just preliminary – doing it properly would give a precise estimate. I use both Nordhaus and Tol’s curves. So the real cost is lower than $23. Therefore the benefit of reducing each tonne of CO2 has a lasting benefit for generations.

Costs of recent climate events in Australia are:
2011 Queensland floods $7 billion – human contribution likely
2011 Victorian floods $1 billion – human contribution likely
2009 Victorian fires $4 billion – human contribution almost certain
2011 Victorian hailstorm $680 million insurance payouts – human contribution possible

We definitely need more research in this area

This, of course, caused all sorts of harumphing, with comments like this:

You must surely be joking to suggest that you alone, apparently, can assuredly connect the floods in Queensland, fires, floods and hailstorms in Victoria to Climate Change. I believe there was a report from some people associated with the IPCC who, after about two years work, said that there was only a very low probability that extreme weather could be linked to Global Warming.


I feel that you are drawing a pretty long bow to claim that the recent bushfires since about 1950 even, are the result of Climate change, but I will await your comments as to how you can distinguish between the cause of a bushfires in 1850, 1909, 1926 and 2009.

and this:

I find this list quite disingenuous.
Climate change has become the latest culprit in major natural disasters, as a way of absolving authorities from any culpability.

The evidence on the Queensland and Victoria floods is circumstantial. Record high sea surface temperatures occurred across Australia’s north in the summer of 2010–11, the source area for this rainfall. September–January rainfall in eastern Australia has increased by 14% with respect to the El Nino–Southern Oscillation (ENSO) index since 1910, most of it in the past 50 years. Rainfall has increased as a function of diurnal temperature range, the best climatic predictor of rainfall. Average Australian rainfall increased as a statistically significant step change in 1973, the same time as maximum temperature increased over much of the continent. Queensland rainfall stands out as an anomaly, but not quite as large as 1974. People are arguing that this means 2010 could not have been influenced by warming – but it is more likely that 1974 was also influenced by warming, and both years would have been exceedingly wet without added warming. Victoria in 2010–11 (Sep–Jan) stands out as an outlier with respect to the two best predictors, the Indian Ocean Dipole and ENSO index. The only comparable outlier is 1992, when heavy rainfall occurred earlier in the season and on a drier catchment, therefore did not cause widespread flooding.

Likewise with forest fire danger index for Victoria. In the years 1997–2009 it was more than one-third higher than in 1972–1996 (about 38%). The warming component of this index is largely anthropogenic and the rainfall component may be.

There is a contributing element of climate change in all of these events. The nature of damage in such events is that they are strongly non-linear, so when a small dose is added to the climate event, a much stronger damages response is felt. However, the methods needed to pin down better estimates of these additional damages still need to be developed. Because they cannot be quantified does not mean they are not having an impact.

The point is, if we are paying a premium in damages now due to climate change, how much can this escalate into the future? It shows that neither adaptation or mitigation can be ignored. Waiting for the piano to hit the pavement is really poor risk management.


One Response

Subscribe to comments with RSS.

  1. Sorry for the delay in commenting. I thought I’d throw in a couple of links to recent material on sea level rise which support the notion that we need to concentrate on mitigation as well as adaptation, and there is a premium on acting now. It seems to me that your earth scientist correspondent desires a relaxed and comfortable life.Depending on his age he may be granted his wish. If he has grandchildren, however, he should have a thought for his grandchildren’s grandchildren, and indeed their grandchildren.

    The first reference is to a paper by Schaeffer et al which looks at the implications for sea level rise up to 2300 for several scenarios. You need to follow the links to understand what the scenarios mean, but in broad terms RCP4.5 refers to stabilisation at forcing of 4.5W per square metre, or 650ppm CO2e. RCP3-PD means 450ppm CO2e, which is roughly where we are now.

    Their conclusions are disturbing. They find, for example, that a 2 °C warming limit would probably see many metres of sea level rise in oncoming centuries, and would maintain rates of SLR higher than today for many centuries. SLR of 1.5m is virtually unavoidable and could only be limited to that by “large-scale deployment of CO2-removal technologies”.

    The second reference is to an OECD study looking at the exposure of 136 port cities with more than a million population to the risk of a 1 in 100 years storm event. Go to this post by Alex Kasdin and follow the links.

    The study looked at the risk exposure now and in 2070, taking into account changes in population, half a metre of sea level rise, subsidence where relevant, and expected increased storminess. They looked at “exposure” on the assumption that adaptation measures don’t exist. Essentially because it would have been too complicated, and pointing out that such measures may in fact fail as they did in New Orleans.

    Right now $3 trillion worth of property and 40 million people are exposed. By 2070 those numbers become $35 trillion (roughly 9% of forecast world GDP) and 147 million (not 120 million as stated in the post). They point out that there is a 74% chance of one or more of these cities being impacted by a once in a century event every year. I’m not sure whether they take account of the notion that what is now a 1 in 100 event becomes much more frequent with half a metre of SLR. Their “increased storminess” refers to the expected increased intensity of severe storms.

    That’s what is in prospect for today’s children’s grandchildren, or rather just on aspect of what they will probably experience.


    August 3, 2012 at 1:38 am

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: