Once-in-a-hundred-year events every five or ten years? Welcome to a brave new world.

Lately, weather has been the news more often than followed the news. After every major storm or other weather event, meteorologists around the world get asked the same question: “Was this due to climate change?” And all around the world, responsible meteorologists give the standard scientific answer – that one cannot attribute a specific single event, even an exceptionally severe one, to climate change. To most people, that sounds like a “no” which, while accurate, is not exactly helpful.

And herein lies a problem – because by the time these events can with scientific certainty be linked to climate change, it will be far, far too late to do anything about climate change. So we must look at probabilities and what the models predict; basically all climate change models agree that as climate change advances and the planet’s climate goes seeking a new balance, severe weather events will become both more frequent and more severe. So while a “100-year-event” occurring in any given year does not in and of itself indicate worsening climate change nor does it increase or decrease the probability of a similar event occurring again the following year, at some point one has to begin drawing some conclusions.

But when?

Would Australia having had two 100-year droughts in the past 100 years qualify? Perhaps the Brisbane river exceeding its ARI 100 (Average Recurrence Interval) flood levels three times in the past 100 years counts? Or what about the severe Amazon rainforest drought, the second “once in a century” drought in just six years? Or the European heat waves of 2003 and 2006, the former of which killed 52,000 people? Or the fact that the 10 warmest years on record have all been since 1998?

One would be inclined to think so. And some even in the mainstream media think so; here is Mike Carlton on Sydney Morning Herald in his article “Flat-earthers, it’s time for a cold shower”:

    [lists the natural disasters from the last 12 months]
    Given this catalogue of global disaster, would now be a good time for the climate change flat-earthers to shut up and listen, do you think? Just for a day or two, or even five minutes?
    They won’t, of course. The global warming denialists ignore the great body of world scientific opinion. When the Queensland catastrophe leaves the headlines the local lot will be at it again, barfing up their crackpot notions.

And meanwhile in Finland, the Finnish Meteorological Institute is – I can only assume fully consciously – implementing a creeping normalcy / shifting baseline phenomenon by switching to use the period 1981-2010 as their comparison period; a period that is on average some 1.5C warmer than the true long-term average up to e.g. 1990. This is a rather dangerous exercise and will end up belittling the warming in Finland by shifting the comparison baseline to a significantly higher level.

In any case, next time there’s a severe weather event, I suggest asking your meteorologist a different question. Ask them “Is it likely that these events would have occurred at this intensity or frequency if atmospheric carbon dioxide had remained at its pre-industrial level of 280 ppm?” and the responsible answer should quite different and come with more certainty than to the first question.

This entry was posted in Climate, Environment and tagged , , , , , , , , . Bookmark the permalink.

5 Responses to Once-in-a-hundred-year events every five or ten years? Welcome to a brave new world.

  1. Andrew says:

    I accept your general point that when looking at multiple occurrences, trends can be discussed, unlike once-off events. However, your examples are not necessarily good examples of this. To quote from one of the links you provide (http://www.goldcoast.qld.gov.au/t_standard2.aspx?pid=334#6):

    As this [ARI 100] is often referred to as the “100 year flood”, it is tempting to think that it will only happen once in 100 years.

    In fact, floods of that size have been known to occur more frequently than once in 100 years. Kempsey in northern NSW experienced two floods of this size in eight months in 1949 and 1950. During the 1890s the Brisbane River experienced three 100 year ARI floods over a period of five years.

    So, in the Brisbane floods case, there seem to be fewer once-in-100-year floods than occurred 100 years ago. This would not provide much support for the case that recent CO2 emissions have increased their frequency.

  2. sim says:

    Yes, the Brisbane example may not be the best one out there, but it’s at least connected to an event fresh in people’s memories. And of course the links there don’t include the latest one.

    While it’s next to impossible to get reliable figures on them, it would be interesting to know what the flood stats are like from the past 1,000 years. If 100-year-floods have occurred on average, say, 3 times per 100 years over the past 1,000 years, I would be inclined to say the definition is wrong / set too low.

    It is, however, much harder to argue against the global average temperature graph.

  3. Andrew says:

    I think it’s just a case of statistics meaning something different to what many people think they mean. A once-in-100-year event might be defined as having 1% chance of occurring in a particular year, i.e. the “expected” number of such events in a 100 year period is 1.0. However, (assuming my maths is right) the probability of no such events occurring in a 100 year period is 2.6%, and that of only one such event occurring is 9.4%. So, there’s an 88% chance of there being more than one once-in-100-year event in a 100 year period!

  4. Andrew says:

    Oops. No, my maths is definitely wrong. The right figures should be P(no events in 100 years) = 37%, P(only one such event in 100 years) = 37%, P(more than one such event in 100 years) = 26%.

  5. Andrew says:

    You know, I’ve spent long enough on these comments that I think I might as well make my own blog post on it.

Leave a Reply

Your email address will not be published. Required fields are marked *