Skip to content

Corinne Le Quere Flooding Claims Rebuttal

January 15, 2014

By Paul Homewood

 

image

http://www.telegraph.co.uk/earth/environment/climatechange/10573767/What-if-man-made-climate-change-is-loading-the-dice-on-floods-in-the-UK.html

 

Corinne Le Quere is Director of the Tyndall Centre, at the University of East Anglia. She recently wrote a piece for the Financial Times, which linked recent heavy rains in the UK to “climate change”, and has now reacted badly to those who have had the temerity to criticise her.

Writing in the Telegraph, she says:

I am a physicist of 20 years’ experience, and climate change research is a science, not a faith. That means it is based on observations and on understanding of how the world works. It is the same kind of science that provides the tides, currents and weather forecasts. It’s not perfect science, but science, and knowing the weather, has taken us a long way in making our everyday life a lot more comfortable.

Mr Thomas is ignorant of the fact that heavy precipitation in winter has increased over the past 45 years in all regions of the UK. That’s not just stories told by people based upon their own experience, it is a lot of data collected and analysed all over the UK.

Mr Thomas is ignorant of the fact that that heavy precipitation is an anticipated consequence of a warming climate in wet regions of the world, such as the UK. It is simple physics: the planet warms, water evaporates more, more moisture is available in the atmosphere for individual storms, therefore more heavy precipitation. Storms are made by the weather, but climate change puts more moisture into the atmosphere that makes the rainfall heavier.

 

The rest of her article is remarkably fact free, just a few suppositions and a couple of snidey remarks about one of her critics.

So, basically, her case rests on her claim that “heavy precipitation in winter” has been increasing in recent years. This claim is derived from the UK Climate Projections Report, itself based on a UEA study by Tim Osborn

Crucially, this study only looked at data from 1961 – 2006. A quick look at the Met Office data for the whole of the period since 1910 highlights a fundamental flaw in this approach.

 

image

http://www.metoffice.gov.uk/climate/uk/summaries/actualmonthly

Figure 1

 

As can be clearly seen, the period from about 1960 to 1980 was an unusually dry one, certainly compared to the first half of the 20thC. Not only that, but terminating the exercise in 2006 means that the run of dry winters in the last few years is also excluded.

This unusual dry period was noted by Professor Stuart Lane, of Durham University, in a 2008 study. He comments:

More than three-quarters of our flood records start in the flood-poor period that begins in the 1960s. This matters because we set our flood protection in terms of return periods – the average number of years between floods of a given size. We have probably under-estimated the frequency of flooding, which is now happening, as it did before the 1960s, much more often that we are used to. “The problem is that many of our decisions over what development to allow and what defences to build rely upon a good estimate of these return periods.

 

It is also relevant that the 1960 – 90 period coincided with the cold phase of the Atlantic Multidecadal Oscillation, that is well recognised as leading to drier weather in NW Europe.

 

Figure 2 shows the longer term England & Wales series, which dates back to 1767. There is certainly nothing “abnormal” about precipitation in recent winters, and the long term trend is consistent with most of the first half of the 20thC, and indeed even the latter part of the 19thC. Furthermore, there is no sign that this trend is increasing; in fact over the last few years it has been decreasing.

 

image

http://www.metoffice.gov.uk/hadobs/hadukp/

Figure 2

 

I appreciate that the specific claim relates to the contribution of heavy rainfall events to total rainfall, rather than simply overall rainfall. But, there is definitely a correlation between high rainfall years and heavy rainfall events –  years with higher totals of rain are likely to have “heavier raindays”  (more on this below).

First though, let’s look at the average rainfall for the days when it does rain, (defined as days when there is more than 1mm).

 

image

Figure 3

 

Unfortunately, the Met Office only keep rainday data back to 1961, so we cannot replicate the earlier part of the century. Certainly we can identify a lower average in the first couple of decades or so, but it is very notable that, since 1990, the numbers have stabilised.

Even the winter of 2012/13, which has the highest average in the last few years, is only the 10th highest since 1990.

If the claim about links to global warming are correct, should we not be seeing an increasing average since 1990?

 

As promised above, I wanted to see if there is any correlation between high rainfall years and heavy rainfall. The suspicion must be that comparison of heavy rainfall events with the “flood dry” period of the 1960’s and 70’s is skewing the results, and not giving a meaningful long term assessment. Would we really see any increase in heavy rainfall events if we looked at data for the whole of the 20thC?

It is certainly logical that wet years tend to have a disproportionate number of heavy rainfall events. After all, if this were not true, it would automatically follow that, not only were there more days with heavy rainfall in these wet years, but also above average numbers of dry, or extremely light rain, days as well. I am not sure if there is any proven mechanism as to why this should be so!

But we can actually go one better, and prove this correlation.

First let’s examine the four years on the above chart with the highest rainfall per rainday average, along with the four lowest. The years with the highest rainday averages also fall into the very wettest years, just as the lowest fit into the driest years.

The winter total rainfalls can be seen in Figure 4.

 

Year Rain/Rainday
mm
Total Winter
Rainfall
mm
Total Rainfall
Ranking
1990 9.0 470 2nd
2008 8.8 386 7th
2007 8.6 437 3rd
1995 8.5 485 1st
       
1963 6.1 168 51st
1971 6.1 229 48th
2010 5.8 255 42nd
1964 4.9 119 52nd

 

image

Figure 4

 

And we can even go one better than that. Figure 5 shows the distribution of total precipitation against the average per rainday. There is a startling correlation between the two – years with low average raindays also have low overall rainfall, whilst the high averages nearly all congregate at the top end of the total rainfall stats.

 

image

Figure 5

 

The evidence could not be clearer – the dry period, which effectively starts in 1961, when the UEA analysis begins, is giving an abnormally low rainday average rainfall. As this period was not typical of most of the 20thC, and it had been much wetter in earlier decades, the claims that “heavy precipitation has increased in the last 45 years”, is not only meaningless, but positively misleading.

As such, the implication that heavy precipitation will “continue to increase” is utterly without foundation. Indeed, all the data points in the opposite direction.

 

Concluding Remarks

1) It is worth restating. Corinne’s claim that “heavy precipitation in winter has increased over the past 45 years in all regions of the UK” is dependent on the fact that the 1960’s and 70’s were unusually dry ones.

If any meaningful conclusions are to be drawn about “heavy rainfall”, then the data needs to be examined back to at least 1910, and preferably back into the 19thC.

2) This whole episode seems to be a classic case of “group think”. Corinne has quoted figures and conclusions published in the Environment Agency’s UK Climate Projections Report, that were in turn provided by the UEA.

It seems it never occurred to her to ask the same questions, and make the same checks, that I have done. One has to ask why?

Does she assume that, because other climate scientists have put the numbers together, they must be right? Does she think they must be correct, because they give the “right results”, the ones that her theories predict?

3) Tim Osborn of UEA, who did the original work, must have been fully aware that the 1960’s/70’s were unusually dry decades. Did it not occur to him that it was not possible to draw meaningful conclusions, without including earlier data? Was it simply too convenient not to have to delve back into earlier data, that was not so readily available? Or did his results give him what he wanted?

4) If Corinne’s little theory about the Arctic is correct, and increasing levels of winter rainfall are a result of a warmer Arctic, then surely it follows that the wetter years in the early years of the 20thC also suggest a warmer Arctic then as well?

5) It is, of course, easy to persuade people that floods, and indeed any type of bad weather, these days is worse than anything before. Most people quite simply lack the longer term, personal perspective to argue against. They also lack the knowledge how to counter such claims.

So, I offer a very simple challenge to Corinne. Analyse the data back to 1910 and tell us all whether this still backs up your contention that “heavy precipitation in winter has increased”.

If it does not, then your claim that climate change has increased the risk of flooding is no more than unscientific propaganda.

 

References

1) UK rainfall & rainday data

http://www.metoffice.gov.uk/climate/uk/summaries/datasets

 

2)

http://www.metoffice.gov.uk/hadobs/hadukp/

20 Comments
  1. Joe Public permalink
    January 15, 2014 10:47 pm

    “3) Tim Osborn of UEA, who did the original work, must have been fully aware that the 1960’s/70’s were unusually dry decades. Did it not occur to him that it was not possible to draw meaningful conclusions, without including earlier data? Was it simply too convenient not to have to delve back into earlier data, that was not so readily available? Or did his results give him what he wanted?”

    Or was it not expected that someone else would ‘delve back into earlier data’.?

  2. miked1947 permalink
    January 15, 2014 11:08 pm

    600 years worth of reliable data is required to better understand the variable nature of climate and set the boundaries of natural variation.
    I base that claim on the know weather patterns such as AMO and PDO that affect regional weather. Those patterns last between 40 and 80 years, guessed average is 60 years. That requires 10 repetitions of the patterns to have sufficient supporting information to base a preliminary conclusion about climate patterns.
    In closing I would say out friend from UAE is about as qualified as Katherine to discuss climate no matter what she claims. She is FOS!

  3. Streetcred permalink
    January 16, 2014 1:07 am

    “[ … ] climate change research is a science, not a faith. That means it is based on observations and on understanding of how the world works.”

    Eh-gad! What observations would those be … looking at defective model regurgitation ? When ‘physicists’ come up with this line of BS then the standing of their profession must be brought into question … makes secondhand car salesmen respectable.

  4. lapogus permalink
    January 16, 2014 5:21 am

    Paul, excellent analysis. I can confirm anecdotal evidence wrt the 1960s, 70s and 80s being comparatively dry decades; SEPA hydrologists and ‘scientists’ were shocked at the magnitude of the big floods on the River Tay in 1990 and 1993, some even publicly attributed the floods to newly discovered man-made global warming. Of course their carefully calculated flood return periods were based on a river system which had very few if any gauges prior to the 1960s. However, the senior hydro-electric engineers upstream were not surprised at all by the 1990 and 1993 floods – they had rainfall data and flow records going back to the 30s and remembered the wet decades before the dry 60s and 70s.

  5. January 16, 2014 5:59 am

    For constructing dams we would like to know the rainfall and flood records for a minimum of 100 years, 1000 would be better.

    In the case of the Nile River, the records are very long and were the subject of landmark statistical studies.

    The Nile studies provide insight into how long a period is needed to distinguish extreme weather events from climate change based on precipitation and flood records.

    While there is no consensus on how long a period would be needed, the length of time would best be measured in centuries.

  6. Joe Public permalink
    January 16, 2014 9:14 am

    Harrabin has a new article “Back-to-nature flood schemes need ‘government leadership’ “, on Aunty’s website.

    http://www.bbc.co.uk/news/uk-politics-25752320

    In it, he twice quotes Katherine Pygott of the Chartered Institution of Water and Environmental Management (CIWEM):- “Flooding is getting worse with changing weather patterns…..”; and “We obviously don’t want to be wasting government money on the wrong projects but with weather patterns changing…..” – neither of which were challenged by the doom merchant.

    • January 16, 2014 11:03 am

      There’s an element of truth in those statements.
      Short term weather pattern are changing in the UK, as they have always done, but I don’t believe it’s anything to do with C02, which is the real issue.

      • Joe Public permalink
        January 16, 2014 2:36 pm

        It is not so much the ‘flooding’/weather which is worsening, it’s the built-environment which has both encroached on flood plains and also affected natural drainage.

        Fig 1 (above) shows a century’s worth of Met Office records.

  7. January 16, 2014 10:22 am

    It is true that the Met. Office only keeps “rainday” data for the UK since 1961 but the HadUKP dataset includes regional daily figures going back to 1931.

    http://www.metoffice.gov.uk/hadobs/hadukp/data/download.html

    Using these files it is possible to compare annual rainfall with the frequency of heavy rainfall days and I can confirm that there is a correlation between such days and high annual rainfall.

    Defining heavy rainfall as over 10mm, in England and Wales the frequency of such rainfall event rises from approximately 10 for an annual rainfall of 700 mm, to 25 for an annual rainfall of 1200mm. The number of days with 10mm or more reach low of 12.8 over the period 1969 – 1978, which also coincided with a low in annual rainfall of about 869 mm.. The figure had been 18.8 over the period 1951-1960, when annual rainfall averaged 946 mm.

    Using the formula to estimate past frequency of heavy rain days from annual figures, it is likely that the peaked at approximately 20 between 1874 and 1883, which
    is approximately the same level as now.

    I haven’t analysed winter rainfall specifically, but I would expect a similar correlation to exist.

    There is no doubt that Corinne Le Quere is looking at this from too short term a perspective.

    • David permalink
      January 16, 2014 5:46 pm

      QV,

      Do you know if > 10mm/day is the correct definition for ‘heavy precipitation’? I can’t find the term defined anywhere. But I used the figure of > 10mm/day and applied it to all the winter months since 1931 in the HadUKP data for each region.

      This gave a raw figure of ‘winter days per month’ in which rainfall exceeded 10mm, per UK region, between 1931 and 2013. I did the same thing for the last 45 years. I then calculated the linear trend (per decade) for these figures to confirm if there has been any change in frequency.

      While the UK as a whole has seen an increase in winter days in which precipitation has exceeded 10mm/day since both 1931 and over the past 45 years, there are some exceptions in both cases. There is also wide regional variation.

      Across the UK regions, days per winter month on which precipitation has exceeded 10mm have increased at a rate of 0.31/dec since 1931; and at a rate of 0.52/dec over the past 45 years. So there has been some acceleration in the latter part of the period.

      However, Eastern Scotland has seen a slight decline in winter days exceeding 10mm precipitation since 1931 (-0.05 days/dec), and this decline has accelerated over the past 45 years (-0.17 days/dec). Northern Ireland has also seen a slight decline over the past 45 years, though it’s trend is upward since 1931.

      Obviously the UKMO will have more precise (possibly hourly?) values that they are applying here, which might also explain why the UK observations data only go back to the 1960s.

      I can post the regional values if you like. Southern Scotland has taken a bit of a battering on both time scales!

      • January 16, 2014 6:15 pm

        “Do you know if > 10mm/day is the correct definition for ‘heavy precipitation’?”
        No, it was just an arbitrary figure i chose.

        I don’t think that changes trends over even 45 years are significant in terms of “man made” climate change.

        Based of the relationship between annual rain fall and daily rain fall, it is likely that we had similar daily rainfall figures in the past, although of course, i can’t prove it!

        The people such as Corinne Le Quere see an increase in rainfall in the short term and assume it’s due to C02 because it suits their purposes, but it doesn’t have to be.

      • January 16, 2014 6:16 pm

        David

        Is the 0.31/dec from 1931 to date? If so, what would the figure be for 1931-60?

        Thanks

        Paul

    • David permalink
      January 16, 2014 7:57 pm

      QV,

      Per above, just to correct a couple of the figures from one part of my earlier reply, paragraph 4 should read:

      ‘Across the UK regions, days per winter month on which precipitation has exceeded 10mm have increased at a rate of 0.27 days/dec since 1931; and at a rate of 0.48 days/dec over the past 45 years. So there has been some acceleration in the latter part of the period.’

      On your point re this being nothing unusual, that could be right, but I fear we’ll never know for certain. I know that NOAA use hourly values for rainfall in a lot of locations and I’m guessing UKMO do too.

      That may be what Prof. LeQuere is basing her figures on. Very unlikely that hourly rainfall data exist in many, if any, UK stations pre the 1960s.

  8. David permalink
    January 16, 2014 1:19 pm

    Perhaps there is a crude way of checking this, back to 1932 at least, using the HadUKP data?

    It’s probably fair to assume that extreme rainfall events can more easily be identified from daily regional records than from monthly country-wide ones. Daily records isolate short term events from persistent but low-key monthly precipitation; and regional records should help isolate localised downpours.

    How about the following as a quick ‘rule-of-thumb’ analysis:

    1. Download the daily HadUKP data (http://www.metoffice.gov.uk/hadobs/hadukp/data/download.html )
    2. Filter out Dec-Feb on Excel or similar
    3. Calculate the daily average rainfall per month
    4. Calculate the trend of these averages in each of the regions.

    If Le Quere is right, then step 4 should reveal a positive trend in each region.

    Alternatively you could calculate the average of the linear trend for each day in each region, which should give similar results but is more complicated because of the different number of days in each month.

    I’m going to try method 1 later and will post results.

    David

    • January 16, 2014 1:46 pm

      Thanks, David.

      I wanted to do something with it, but the format beat me!

      Also, see QV’s comment above.

      • David permalink
        January 16, 2014 4:38 pm

        That’ll teach me to read previous comments!

    • David permalink
      January 16, 2014 7:12 pm

      Alternative method was used @ 3.

      3. Count number of days per month with values > 10.0mm

      Results described above in reply to QV.

      • David permalink
        January 16, 2014 7:47 pm

        Paul,

        Hold on, I’ve made a mistake, but it doesn’t affect the general picture. I loaded the SWEP data for both SWEP and CEP. (I was wondering why both trends looked so similar.) Sorry about that.

        The figure that was 0.31 is now 0.27.

        So 0.27 days/dec is the rate of increase in the number of days per winter month with precipitation greater than 10.0 mm between Dec 1931 and Dec 2013.

        By the same measure, and having corrected CEP, the same figure 1931 to 1960 is 0.10 days/dec.

        I’ll update the post to CV with the (hopefully) correct figures.

      • January 16, 2014 10:42 pm

        Hi David

        Can you email me the files across? I’ll see if we can out some graphs out.

        Let me know if Ok, and I will email you.

        Thanks

        Paul

    • David permalink
      January 17, 2014 7:30 pm

      Paul,

      One more caveat to the above figures that I thought of is that we’re discussing ‘winter only’ months. I multiplied everything by 120, as you would normally do for decadal records. But since winter is only 1/4 of each decade, I think 30 might have been the appropriate multiplier (?).

      Note, this doesn’t affect the proportional differences between the values quoted, just the actual numbers.

      Also, I changed my spreadsheet to reflect decadal values. It might still be of some use though. If you can point me to an email link I’ll forward the spreadsheet on. I’ll explain further in the email.

      David.

Comments are closed.

%d bloggers like this: