Skip to content

Are Natural Disasters On The Rise?

September 16, 2017

By Paul Homewood





The BBC ran a piece on natural disasters yesterday, which I have to say was a bit more balanced than most of their output on climate change.

It discussed claims that the number of natural disasters has been rapidly increasing in recent decades.

They quoted UN Secretary-General Antonio Guterres as saying last week:

The number of natural disasters has nearly quadrupled since 1970.

According to the BBC, similar stories have cropped up across the media, from The Economist to Fox News.

These claims apparently originate from an Oxfam report in 2007, which in turn relied on data from CRED (Centre for Research on the Epidemiology of Disasters).

To be fair to the BBC programme, it did raise question marks over the consistency of the data over time, but failed to really push this home. As even CRED readily admit, many more disasters get to be reported nowadays for all sorts of reasons. Instead the programme claimed that climate change could also be responsible for some of the increase.

So, let’s look at CRED, and how their data stacks up.

In 2004, they published this report:





It included this comment:



And continued:



[EM-DAT,  the Emergency Events Database, was launched by CRED in 1988.]



Their graph makes all of this abundantly clear:


Nobody in their right mind would believe that there were hardly any natural disasters in the first half of the 20thC. Many disasters happened in the past, but which don’t appear in the official stats.

A clue to this is that most of the apparent increase is due to small disasters:




In fact, the criteria for what constitutes a “disaster” is set at a very low level indeed:



Thousands of such small events would have escaped official notice in the past.

There is one more clue in the 2003 report:



While the number of reported disasters has remained pretty much flat from disaster agencies and governments, there was a huge increase from specialised agencies in the 1998 – 2000 period, along with a steady increase from insurance companies.

This is clear evidence that the apparent trend is solely due to how the data is collected.



Now fast forward to their Annual Report for 2006:





Note that CRED have only been publishing annual stats since 1998. Although the EM-DAT was begun in 1988, it would appear that the data can only replied upon since 1998.

We then find this graph:




So although the number of natural disasters appears to have doubled since 1987, in reality there was a big step change between 1997 and 2000.

Coincidence? I think not.

Again, nobody could seriously be expected to believe that the number of actual disasters suddenly shot up in 1998, and then stayed at that level.

We are entitled to be even more suspicious when we examine the number of victims (deaths plus affected):


The linear trend is highly misleading because of the anomalous spike in 2002. In reality, the trend is flat, and certainly does not support the message that the number of disasters is increasing.



If we look at their most recent report for 2015, we can see that the number of disasters has actually been trending downwards since 2000.




In summary, there is absolutely no evidence that natural disasters have become more common since 1970.

EM-DAT was specifically set up to provide accurate data on disasters, something that has improved as time has gone on. Prior to that, aid agencies and the like were too busy on the ground to bother with collating numbers.

EM-DAT may be a worthwhile exercise, but it should not be used for analysis of long term trends.

  1. quaesoveritas permalink
    September 16, 2017 2:25 pm

    It would be interesting to know exactly what the disasters were in 2002, which caused the huge increase in the number of victims in 2002.
    Also, what it the definition of “affected”?
    That could cover a wide range of effects.

    • quaesoveritas permalink
      September 16, 2017 2:31 pm

      A quick search on the internet has revealed the following in January 2002:

      “Jan. 17–18, Goma, Dem. Rep. of Congo: Rivers of molten lava poured from Nyiragongo volcano, wiping out at least a dozen villages and engulfing the city of Goma. An estimated 300,000–500,000 people were displaced, many of whom fled temporarily to neighboring Rwanda. The Red Cross set the death toll at 47, but the number was expected to rise. Experts said the eruption was the worst since 1977, when 2,000 people died after lava burst from fissures on the volcano’s flanks. A few days later (Jan. 21) the hot lava touched off an explosion at a gas station in the center of Goma, killing dozens of people who were looting gasoline.”

      I don’t know if that is included in the CRED figures, but if it was it might account for a large number of the “affected” figure. But it isn’t “climate change”!

      • quaesoveritas permalink
        September 17, 2017 7:38 am

        Looking at the 2006 report it appears that the high number in 2002 was not due to the volcanic eruption in the Congo, but to droughts in India and China (360m), a “wind storm” in China (100m) and “a flood” in China (60m).
        Without further details and verification, I feel that these figures must be highly dubious.

      • quaesoveritas permalink
        September 17, 2017 7:41 am

        It looks like this was the “wind storm” in china:

  2. Jack Broughton permalink
    September 16, 2017 2:35 pm

    This is another example of the use of powerful spread-sheets by people who do not understand the basis of correlation / causation analysis. The hypothesis that natural disasters are increasing should be subject to proper assessment of historical data not just presentation of numbers. As you say the definition of a disaster as 10 deaths would not have even been reported beyond local press a few decades ago in most of the world.

    GIGO rules unfortunately.

  3. NeilC permalink
    September 16, 2017 3:02 pm

    Fig 2 – It’s the hockey stick all over again, forget past information, disasters just didn’t exist.

    Volcanic eruptions, earthquakes, wars and terrorism can all be counted as “disasters” with more than 10 deaths, but have absolutely nothing to do with climate change.

    Starting in 1900 Hurricane hit Galveston, deaths 8000-12000, 1914-18, 1939-45 now there are three massive disasters. Surely the graph should be the other way round.

    And exactly how much has the climate changed since 1990? or indeed since 1900, less than 1 Deg C.

    It’s another load of bo££ocks.

    • Jack Broughton permalink
      September 16, 2017 7:19 pm

      like the use of “bo££ocks” in context Neil, and will certainly steal it!

    • September 17, 2017 11:29 am

      Further, the deaths caused by the Galveston Hurricane in 1900 was basically from the lack of the modern prediction methods AND the lack of the internal combustion engine–there were no high-water trucks or the Cajun Navy to rescue people.

  4. September 16, 2017 3:42 pm

    More people, more wealth, better reporting => more ‘disasters’. Nothing to do with CAGW. No different than more EF-1 tornadoes in the US after doppler radar was rolled out nationally in the late 1980’s, more tropical,storms after weather satellites launched that can spot them.

  5. September 16, 2017 3:56 pm

    How many, or what percentage, of these ‘natural’ disasters are in reality strongly related to land use changes?

  6. Ben Vorlich permalink
    September 16, 2017 4:29 pm

    In the Communist era natural disasters didn’t happen in the USSR or China. Many totalitarian regimes didn’t report disasters in the 20th century.

  7. Broadlands permalink
    September 16, 2017 5:01 pm

    “The number of natural disasters has nearly quadrupled since 1970.”
    Even if true, what are we supposed to do about controlling these natural events?

  8. Athelstan permalink
    September 16, 2017 5:16 pm

    “disasters have evolved over time”…………………… oh really?

    ‘disasters evolving’

    FFS, primarily: they need to revisit a dictionary of English.

    In a chaotic and dynamic system, shit happens but NB: disasters can’t “evolve”.

    News24/7, televisual coverage has ‘blossomed’, increased the ‘keep ’em worried, keep em cowed’ and relentless journo’ hyperbole has greatly boomed, the pscience of hazard evaluation has taken on a life of its own and some bods somewhere are no doubt trying to statistically analyse it – quantum mechanics and all..

    We are making advances yet we can’t predict earthquakes because too much government funding is wasted on MMCC or and goes into projection of what if scenarios and btw the merest “lightest” touch can change stuff and if one of those boogers ever hits as one surely will, then we won’t need to worry about ‘evolving’ any longer, it will be a life terminating disaster.

    Me? I trust in the runes, bones cast and tea leaves read, just like the wet office do, it’s no evolution, I’m telling yers.

  9. ben dussan permalink
    September 17, 2017 12:46 am

    Note that the world population has more than quadrupled since 1900: thus as the population increases, it is likely that natural disasters will kill more and definitely affect many more people….

  10. Bitter&twisted permalink
    September 17, 2017 8:28 am


  11. September 17, 2017 1:30 pm

    Reblogged this on Climate Collections and commented:
    In summary, there is absolutely no evidence that natural disasters have become more common since 1970.

    EM-DAT was specifically set up to provide accurate data on disasters, something that has improved as time has gone on. Prior to that, aid agencies and the like were too busy on the ground to bother with collating numbers.

    EM-DAT may be a worthwhile exercise, but it should not be used for analysis of long term trends.

  12. September 17, 2017 10:51 pm

    Hey Paul why the long analysis for something that does not deserve it
    There is no need to stand in the forest counting the trees
    Instead step out of the forest and spot thr forest is BS
    As Jack Broughton says the claim is GIGO

    The standard deceiving by omission and implying stuff
    The Economist graph implies it’s counting “Weather disasters”
    …It’s conditional : Weather disasters, which are reported and have 10 deaths or a charity appeal.
    #1 Golden rule of comparisons is you can’t compare apples and oranges
    Look if you had the same weather event in a valley in 1900 and then in 2010
    It wouldn’t have got counted the first time, cos either news wasn’t reported or no one was living there.
    #2 Another Golden rule define your terms and the economist didn’t
    #3 Another golden rule is “an extraordinary claim” needs “extraordinary evidence”
    and in this case NO extraordinary evidence was presented to back up the claim”WDs have quadrupled” was just “according to a survey”

    #3 Another golden rule is the “All other things being equal” , but they aren’t reporting rates and population change as well as the 2 min variables : Years/count.
    You have to take step to smooth out effect of extra variables.
    So in counting deaths , it is not good enough to just given raw numbers, they need to be expressed in % of world population etc.

    The main take is that the Economist/Oxfam et al chose PR tricks over truth and science.

  13. quaesoveritas permalink
    September 23, 2017 6:49 pm

    An actually quite sceptical discussion of this on the BBC World Service, although I think the programme is “More or Less”, which was first broadcast on R4.

Comments are closed.

%d bloggers like this: