UN Climate Disaster Doubling Revisited

Guest essay by Pasi Autio 20th of October 2020

Couple of weeks ago a new United Nations report claimed doubling of natural disasters between periods 1980-1999 and 2000-2019:

Human cost of disasters

An overview of the last 20 years 2000 to 2019

GENEVA, 12 October 2020 – A UN report published to mark the International Day for Disaster Risk Reduction on October 13, confirms how extreme weather events have come to dominate the disaster landscape in the 21st century.

In the period 2000 to 2019, there were 7,348 major recorded disaster events claiming 1.23 million lives, affecting 4.2 billion people (many on more than one occasion) resulting in approximately US$2.97 trillion in global economic losses.

This is a sharp increase over the previous twenty years. Between 1980 and 1999, 4,212 disasters were linked to natural hazards worldwide claiming approximately 1.19 million lives and affecting 3.25 billion people resulting in approximately US$1.63 trillion in economic losses.

Much of the difference is explained by a rise in climate-related disasters including extreme weather events: from 3,656 climate-related events (1980-1999) to 6,681 climate-related disasters in the period 2000-2019. 

The last twenty years has seen the number of major floods more than double, from 1,389 to 3,254, while the incidence of storms grew from 1,457 to 2,034. Floods and storms were the most prevalent events.”

Doubling of natural disaster events – this is bad. The conclusion was based on EM-DAT International Disaster Database (1) consisting of 22000 mass disasters in the world between 1900 and 2020.

The claim is so bold that this requires more thorough analysis of the original data. Everyone can access the EM-DAT database by registering as EM-DAT user. Access is free for non-profit organizations, so let’s do it.

After approval as an EM-DAT database user, I was able to download the whole database of disasters between 1900 and 2020. By just selecting natural disasters I got the Excel file of 15564 individual disasters, each as a line in the Excel. Data contains up to 43 different columns of information for each disaster – for most disasters only less than half of the columns has any data.

Let’s start by plotting a diagram of number of disasters each year between 1900 and 2020.

Figure: All natural disasters worldwide in EM-DAT database

It’s easy to see where the UN claim comes from. Indeed the natural disasters have increased considerably during years 2000-2019 compared to years 1980-1999. But at the same time we see the first hint of what is wrong: There is a gradual increase of natural disasters from 1940s level to today’s level. If you see this kind of data the first question you have is the reporting: How does the improved reporting and data collection affect the dataset?

Let’s plot the number of countries reporting at least one natural disaster per year between 1900-2020:

Figure: Countries reporting at least one natural disaster each year

From the list of countries reporting at least one disaster we can see huge increase of reporting countries during the years. Also the shape of the data is very similar to the total number of natural disasters plotted earlier.

During 1901, for example, there was only two countries in the database (Japan and Uganda) whereas during 2000-2019 the number of countries reporting was about 120 each year. Even the new data has a lot less than countries than there is countries in the world, so it is likely that even the years 2000-2019 are underreporting the actual natural disasters quite heavily.

To analyze further the results in the UN paper the years 1980-2019 are of specific interest. You can see significant increase of reporting countries between 1980 and 2000. After that the number is quite steady. You might think this is due changes like Soviet Union breaking to many different countries, so let’s analyze Soviet Union in more depth:

Figure: Natural disasters in Soviet Union for each year between 1900-1991

We can see that the Soviet Union was not the most open reporter of natural disasters to the EM-DAT database and the increase of countries reporting natural disaster during 1990s is not the result of Soviet Union breakup. And what is the likelihood of country with the size of Soviet Union having only 1-2 natural disasters per year? You can easily conclude that this database contains only small portion of real natural disasters what happened in Soviet Union. Just the forest fires should have tens of reported incidents every year.

And by the way, it seems that the same “issue” of not reporting disasters seems to affect all former Eastern Bloc countries; almost all of them has a huge increase of natural disasters in the database after breakup of Soviet Bloc at 1991. Yugoslavia, for example, reported 0-2 natural disasters per year, but the just Serbia alone is reporting more than that. Clearly the entry criteria is not comparable between pre- and post-Yugoslavia era.

What about China? China was specifically mentioned in the UN report as a country with significant increase of natural disasters:

Figure: Natural disasters in China for each year between 1900-2020

You can see that for years before 1980 there is no meaningful data in the EM-DAT database. The reporting increases gradually after 1980 and reaching “steady state” around year 2000.

For 1980 the database contains just 5 disasters in China: Four floods and one tropical cyclone (there where several cyclones making onshore at China during 1980). The peak year 2013 contains total of 43 disasters: Droughts, Extreme temperatures, Storms, Floods, Earthquakes. The effect of better reporting year-by-year can be easily seen. Communist China was not the most open reporter of natural incidents either.

How about USA? Modern western civilization must have really good disaster reporting already in 1980s, right?

Figure: Natural disasters in USA for each year between 1900-2020

For 1980 the database contains just 8 natural disasters in USA: 4 floods, 3 storms, one volcanic activity (St. Helens) and one heat wave.

No tornadoes are present in the database for 1980, but a little study from Wikipedia tells us that year 1980 was below average tornado year with 28 tornado deaths and several bad outbreaks such as 1980 Kalamazoo tornado outbreak and 1980 Grand Island tornado outbreak. In total the season had 866 reported tornadoes. So, the EM-DAT database is just missing all these events.

Speaking of tornadoes, let’s see how much tornado data is available in the EM-DAT database:

Figure: Number of tornadoes in USA for each year between 1900-2020

It seems that while tornadoes have been the issue in the USA all the time, only after end of 1980s there has been some level of data collection of tornadoes to the EM-DAT database. And even now only small number of tornadoes end up into EM-DAT database. I took year 2000 as an example: Based on EM-DAT there was 30 deaths due to tornadoes. But according to the Tornado season 2000 wiki page it should be 41.

According to the Wikipedia, the tornado year 2012 was about twice as harsh as year 2000 in number of tornadoes and death count was 69 during 2012 season. But the EM-data says year 2000 had more tornadoes fulfilling the entry criteria.

Also no wildfires are present in 1980 data for USA. More in-depth study would reveal some more wildfires, but at least Panorama Fire (1980) is missing from the database. In this fire 28,800 acres burned, destroying 310 homes and 67 structures, killing four people, and injuring 77 in north San Bernardino. Clearly an incident, which should be in database according to the entry criteria.

In USA it seems that only from 1990s forward the database has some level of credibility. Earlier data is simply too lacking to draw any meaningful conclusions about the increase or decrease of natural disasters. But even for recent data, you should not make any conclusions about the number of natural disasters in USA.

The same issue of underreporting on early years seems to affect almost every country I look into. The country I live in, Finland, has only three natural disasters in total in the database; two storms (1990) and one flood (2005). Living in Finland I can assure all readers that we have floods every year (especially with rivers in north we have flooding after every winter) causing material damage almost every year. We also have storms affecting tens of thousands every year usually causing material damage in the forests and lot of damages to the electricity distribution.

What is exactly the entry criteria for EM-DAT database:

Entry criteria: The reason for recording the disaster event into EM-DAT. At least one of the following criteria must be fulfilled in order for an event to be entered into the database:

  • Deaths: 10 or more people deaths
  • Affected: 100 or more people affected/injured/homeless.
  • Declaration/international appeal: Declaration by the country of a state of emergency and/or an appeal for international assistance

As you can see the entry criteria quite relaxed: With this criteria almost every F3 or higher tornado, for example, should be in database assuming it happened on populated area.

Thus, we can conclude that the database really does not have any credibility even today, but even less during 1980s and 1990s. I don’t know how the data collection has been organized, but for scientific analysis of natural disaster trends the EM-DAT database has no scientific value. Therefore also the conclusions in the UN report have not merit. All claims that UN made about the increase of natural disasters should be retracted.

Total damages

UN also reported the sizable increase of damages. Pielke Jr has a lot more scientific merits to say anything about the normalized damages (2) during the years, but we can still make some interesting observations about the EM-DAT data for damages:

It seems that only small part of the entries has any damage data in the database. In 1994, for example, a tropical storm hit Osaka in Japan. According to the database 1000 died and 6.5 million was affected. The cost of this event is missing. This is just one of the thousands of missing damage entries. Only about 1/3 of all entries in the database has any kind of damage estimate. How do you draw any conclusions from that?

Affected people

UN claims that during 1980-1999 natural disasters affected 3.25 billion people whereas during 2000-2019 4.2 billion people was affected. But they omit to discuss the population increases. The world population was:

  • 1980: 4.46 billion
  • 2000: 6.14 billion
  • 2019: 7.71 billion

We can make the rough estimation that during 1980-1999 the average population was 5.4 billion and during 2000-2019 the average population was 6.9 billion. More population should mean more people affected (and dead) due to natural disasters. 3.25 x (6.9/5.4) = 4.15 billion – thus the increase of people affected can be explained entirely with the world population increase. Actually even more, since most of the population increases tend to happen on natural disaster-prone areas such as India, Bangladesh and Africa.


But regarding EM-DAT dataset itself, almost 30% of the whole dataset is missing the affected people data. Older the data is, more of it is missing.


Data collection methods

While trying to find some information about the EM-DAT history, I found something interesting (3). The document released 2004 provides interesting insights of EM-DATA data sources during the years and will explain quite well what we saw above: Why the observations will explain the “increase” of natural disasters.


This diagram is from the document released by EM-DAT maintainers

Figure: Reporting sources for EM-DAT database between 1974-2002

Natural disaster reports are provided by a number of sources and there seem to be significant evolution of EM-DAT reporting scheme during the years. Significant “increase” of natural disasters by around 1999 seems to be explained entirely by a new source of data “specialized agencies”. Specialized agencies refer to sources like e UN World Food Programme, the World Health Organization or the US National Oceanic and Atmospheric Administration. Thus, it seems that the data collection during the years has not been stable and based on scientifically stable sources.


Summary

  • EM-DAT database data is of poor quality in general
  • EM-DAT data collection methods have significantly evolved during the years rendering trend analysis totally invalid
  • UN report does not take in account the significant increase of world population, which by itself explains all increase of affected people
  • All conclusions made about the increase of natural disasters based on this database should be retracted

REFERENCES

  1. EM-DAT, CRED / UCLouvain, Brussels, Belgium – www.emdat.be (D. Guha-Sapir)
  2. Pielke, R. (2020). Economic ‘normalisation’ of disaster losses 1998–2020: a literature review and assessment. Environmental Hazards, 1-19.
  3. D. Guha-Sapir D. Hargitt P. Hoyois, Thirty years of natural disasters 1974-2003: the numbers, centre for Research on the Epidemiology https://www.unisdr.org/files/1078_8761.pdf

41 thoughts on “UN Climate Disaster Doubling Revisited

    • And if that doesn’t work then redefine terms … plenty of scope with the word disaster. I mean for some people burning a meal qualifies as a disaster.

  1. Tomorrow night the last debate between President Trump and ex-Vice President Biden will feature 7 topics. You guessed it, one will be “climate Change”. If elected Biden will stop climate change. No idea how. If reelected Trump will attend to other matters.

    • Trump will cite how the US has lead the world in reducing CO2 emissions and that China, by far the largest emitter and growing, is doing nothing. He doesn’t even need to deny a need for reduction. Biden is in bed (literally for Hunter) with China, and will never call them to task if elected.

    • He doesn’t need to. Didn’t Obama state that with his election, the seas began to recede and the heat decline? Obama was in office for 8 years and Trump for just less than 4. With 8 years of receedings and declines compared with 4 years of the opposite, are we not still ahead?

      • Obama read the story of King Cnut and misinterpreted the message as a confirmation of the power of regents over nature. He also read the story of the kings new clothes and understood the message as “if enough people believe it the new clothes will magically appear”. Is there a similar fable that can make a demented, dishonest and frail ex VP into the savior of mankind?

    • The cyclone activity trend chart from BoM is remarkably accurate, especially since they have been overestimating for years, and started fudging evidence like overrating Debbie, which was barely a rain depression when in reached Proserpine.
      The “Firechart” guide is bunk. All about climate change, nothing much about the real cause which is failure to remove fuel load, cut buffer zones, thin overgrown forests etc.

  2. Some kind of “science” coming from the UN. Definitely not helping the cause of climate alarmism when the political agenda posing as science is this obvious.

  3. Is that true, or did you read it in a UN report?
    A light typhoon season in the Pacific thus year.
    And not one major earthquake hit Michigan.
    I was there all year and would gave noticed.

  4. Every day I am more amazed and more dismayed at the lack of analytical skills and the lack of knowledge of the real world of our so-called “elites” in positions of governmental positions. It’s like they all live in their basements and never even venture out to get food or household supplies. Who enables these people? Their mothers?

    • Tim,

      “…… I am more amazed at the lack of analytical skills and the lack of knowledge of the real world of our so-called “elites”……” is quite deliberate and called “fraud”.

  5. The UN: Doubling down on alarmism by doubling up on “climate disasters”. Double the fun, and double the Stupid.

  6. An oldie but goodie. In MN, they were nearly wiped out by an October surprise snowstorm.

    “U.N. Predicts Disaster if Global Warming Not Checked
    PETER JAMES SPIELMANN
    June 29, 1989
    UNITED NATIONS (AP) _ A senior U.N. environmental official says entire nations could be wiped off the face of the Earth by rising sea levels if the global warming trend is not reversed by the year 2000…”

    https://www.apnews.com/bd45c372caf118ec99964ea547880cd0

    • From that 1989 link:

      “ The most conservative scientific estimate that the Earth’s temperature will rise 1 to 7 degrees in the next 30 years, said Brown.”

      Conservative?
      And note that scientific range of 1-7 degrees
      Basically the same thing they are saying now 31 years later

      Got to love climatology

    • Great article link Derg! So many predictions, so many falsified.

      Love the 10 years to save us – in 1989!

  7. the UN Office of Risk Reduction needs the human induced climate risk to increase otherwise they don’t need such an office.

  8. Are we actually experiencing the final push toward worldwide totalitarianism? The lies and the censorship have never been this extreme in my nearly six decades.

    • @Rich – I’m about the same age, and it does seem to be worse than ever before.

      However… I do wonder if it is actually worse. What if there had been, say, a Fred Watt, or a John Breitbart back in the ’50s and ’60s? (Insert mental image of them furiously cranking away at their mimeograph machines.)

      It is not nearly so easy to get away with fraud as it was when we were whippersnappers, seeing the almost worship in the eyes of our parents when THE Walter Cronkite appeared on the black and white television screen.

      • It’s the censorship and the arbitrary infringement of individual rights that I find alarming. Sure the lies have always been with us, and powerful people have always squashed stories. But now it’s “in your face”.

  9. So a big increase in natural disasters since the invention of television and TV “news hour“…

  10. Its a pretty safe bet that ANYTHING coming out of the UN nowadays is a pack of lies and distortions.

    Its what they do.

    It who they are.

  11. Congrats.

    The U.N. Just plotted the evolution of distributed, networked computing.
    Sheesh, back to bed 😊 folks nothing to see here, except…..
    Who were the fools funding this?
    My guess, it’s my tribe in the U.S..
    We have got to reelect DaDonald!
    Any other countries want the prestige in hosting the freeloaders?
    Don’t think so.
    Bye

  12. When you compare apples with spanners, the answer is always 42.

    ”In the period 2000 to 2019, there were 7,348 major recorded disaster events claiming 1.23 million lives, affecting 4.2 billion people (many on more than one occasion) resulting in approximately US$2.97 trillion in global economic losses.

    This is a sharp increase over the previous twenty years. Between 1980 and 1999, 4,212 disasters were linked to natural hazards worldwide claiming approximately 1.19 million lives and affecting 3.25 billion people resulting in approximately US$1.63 trillion in economic losses.”

    Apart from the poor reporting issue as outlined in the article, You can’t compare two disaster events, and measure them by any meaningful metric. If one event kills more people or causes more economic loss, tragic that may be, it probably tells us more about population and structural vulnerability than it does about the severity of the event. If a flood affected X number of people in 1900 and (somehow) the exact same flood affected >x in 2020, then it’s because the pop. has increased.
    Point, I suppose is, none of the factors remain the same.

    Eamon.

    • That was the point with population increase part of my text: The cyclone that killed one cow and destroyed local farmer’s crops at 1980 might damage 50000 houses in the suburb built to that some place by 2015. First one did not end up to EM-DAT database due to entry criteria not being reached and no-one cared (expect that farmer), but latter did end up into EM-DAT along with the news story in CNN and all other relevant media outlets in the world.

      Funny thing was that the same kind of “increases” of natural disasters, which was shown in the first figure was also evident with disasters like earthquakes. There is no reason whatsoever for earthquakes to increase in numbers except better reporting and improved data collection.

  13. Its quite clear that this is a statistical nonsense. If you take into account some countries but exclude others, or vices versa then you can get whatever result you like.

    Lets take a simple case: If you were to measure the height of school children between the ages of 10 and 16 then in a later analysis include 4-9 year olds, could you assume that the height of school children has decreased ? Of course not. This is what this analysis is doing.

  14. Well, when dealing with global disasters, the first thing the one should do is divide the number of reported disasters by the number of reporting countries. One can get more sophisticated by dividing by either the total area of reporting countries, or because the definition of disasters is keyed to the number of people killed/affected, divide by the total number of people represented by the reporting countries for a commonly used metric of disasters per million people. Either the people compiling the database are incompetent, or they are purposely attempting to misrepresent the situation.

  15. To Mr Pasi Autio, you have quite rightly analysed this ridiculous “disaster landscape” metric by which the UN hope to garner sympathetic responses using the misery of others, with the manipulation of disaster statistics in an unscrupulous attempt to fabricate a story that in reality doesn’t actually exist, “it’s always worse than we thought”.

    This year the Indian and Bangladesh authorities managed a monumental task of evacuating millions of their citizens out of the way of Super Cyclone Amphan, death toll 128 people. A similar cyclone in 2007, Sidr, saw at minimum 3,500 dead and very probably a heck of a lot more. Climate conclusion?

    S Gupta et al. has interesting data about Decadal Frequency of Cyclones in the Bay of Bengal over 140 year time span. (Spoilers – it’s not getting worse)

    https://ibb.co/KK76386

    Here is another one, in a different part of the world, I was reading not so long ago:

    A 66-year tropical cyclone record for south-east Africa: temporal trends in a global context
    Jennifer M. Fitchett and Stefan W. Grab 2013.

    3.3. Trends in tropical cyclone frequency
    Analyses are henceforth made on the most robust 66-year Unisys record. This record indicates an average of
    2.9 (σ = 1.7) and 0.8 (σ = 0.8) cyclones per annum making landfall over Madagascar and Mozambique,
    respectively. A statistically insignificant decreasing number of tropical cyclones have made landfall over Madagascar (−0.2/decade; r = 0.18, p = 0.15) and insignificant increasing number over Mozambique (0.04/decade; r = 0.08, p = 0.50) during the last 66 years. Despite a 0.3 °C sea surface temperature increase over the south Indian Ocean since 1960 (Reason and Keibel, 2004; Mavume et al., 2009), the frequency of tropical cyclone landfalls over south-eastern Africa has not increased.

    The UN must be disbanded.

  16. Years ago I ran the radiochemistry lab at a nuclear power plant. We monitored all our pure water systems for trace contamination – primarily chlorides, fluorides, sulfates, etc. Normality everything was below detection limits (<5 ppb). But when we upgraded instrumentation and could detect down to 0.5 ppb we would sometimes report actual values and suddenly it was “unprecedented” and a “crisis” because we had never had 0.75 ppb chloride in our system. It was very difficult trying to explain detection bias to management.

  17. When economic losses are presented for different time periods w/o adjusting for inflation and GDP, you can stop reading.

Comments are closed.