NZ cranks finally publish an NZ temperature series – but their paper’s stuffed with errors

You can’t teach old dogs new tricks, it seems — certainly not if they’re gnawing a much loved old bone at the time. The lads from the NZ Climate Science Coalition — yes, the same boys who tried to sue NIWA over the New Zealand temperature record and lost, and who then folded a trust to avoid paying court-ordered costs — have finally found a learned journal gullible enough to accept and publish their shonky reworking of NZ’s temperature record. Earlier this month Environmental Modelling and Assessment published A Reanalysis of Long-Term Surface Air Temperature Trends in New Zealand by CR de Freitas & MO Dedekind & BE Brill (DOI 10.1007/s10666-014-9429-z).

My attention was drawn to dFDB 2014 by an NZCSC press release, and yesterday Richard Treadgold, the man who kicked off the whole sad affair five years ago, posted a disingenuous and misleading article about the paper at his blog. As you might expect given the authors, the paper does not call for an upward revision in the amount of warming NZ has experienced over the last century. The abstract concludes with the following:

Current New Zealand century-long climatology based on 1981 methods produces a trend of 0.91 °C per century. Our analysis, which uses updated measurement techniques and corrects for shelter-contaminated data, produces a trend of 0.28 °C per century.

As you might also expect, given the authors and their respective track records, the paper is riddled with schoolboy howlers and outright misrepresentations. It would probably never have seen the light of day without the assistance of Chris “Pal Reviewde Freitas and his undoubted ability to steer tosh to publication.

Here’s a partial list of the errors, misdirections, misrepresentations and shoddy scholarship in the paper, and in the approach taken by de Freitas, Dedekind and Brill (dFDB 2014).

dFDB 2014 repeats the old canard that NIWA’s Seven Station Series (7SS) before the 2010 review was based on the adjustments made in Jim Salinger’s 1981 thesis. This was a key claim in the NZ Climate Science Education Trust‘s evidence to the High Court and so transparently at odds with written reports and papers from 1992 onwards that it was easy for NIWA to refute. As one close observer of the case told me:

Judges may not understand maths, but they are pretty good at English, and take a dim view of litigants who wilfully and perversely misrepresent simple English sentences.

dFDB 2014 derives a warming rate of +0.28ºC per century, by claiming to apply a method published by Rhoades and Salinger in 1993 (RS93). It claims to create a new benchmark record by reapplying an old technique — essentially ignoring all the work done by NIWA in deriving the current 7SS. Unfortunately, the paper is based on a misapplication of the very method it claims to rely on, and includes numerous errors.

The paper as published contains no workings or supplemental material that would allow reproduction of their results, but it appears to be essentially identical to an “audit’ of NIWA’s Seven Station Series conducted by the NZCSC, and which was offered as evidence in their trust’s attempt to sue NIWA.

As such it contains mistakes that were pointed out in NIWA’s evidence to the High Court — evidence which was extensive, thorough and damning, but is not (yet) available in the public domain.

dFDB 2014 claims that RS93 mandates the use of one year and two year periods of comparison data when making adjustments for a station change, but RS93 makes no such claim. RS93 uses four year periods for comparison, in order to ensure statistical significance for changes — and no professional working in the field would use a shorter period.

The choice to limit themselves to one and two year comparisons seems to have been deliberately made in order to limit the number of adjustments made in the reconstructed series. Limiting the comparison periods makes it harder for adjustments to reach statistical significance, leading dFDB 2014 to reject adjustments even in cases where the station records show site moves or changes!

The effect of that is to reduce the warming trend because, as Treadgold’s first venture into this field showed, a naive reconstruction of the raw data shows not much warming.

But perhaps the most critical flaw in dFDB 2014 — one that should have been sufficient to prevent publication in any self-respecting journal operating a credible peer review process — is that their method ignores any assessment of maximum and minimum temperatures in the adjustment process. This was pointed out to the authors in NIWA’s evidence in the High Court. One of these adjustments will almost always be larger than that for the mean, and if that change is significant, then the temperature record will need to be adjusted at that point – it doesn’t matter if the mean temperature adjustment is statistically significant or not.

Silly mistakes in the application of their version of RS93 appeared in the “audit”, were pointed out in NIWA’s evidence to the High Court, but appear to be uncorrected in dFDB 2014. For example, in the “audit”, they infill a month of missing data (May 1920 in the Masterton series) by choosing an unrealistically warm temperature based on an average of years around the adjustment date. This ignores the fact that May 1920 was one of the coldest Mays on record, at all sites involved in the adjustment calculation.

The dFDB 2014 infill has the effect of reducing the statistical significance enough to reject an adjustment — despite the station record clearly showing that an adjustment is required! Any other approach — skipping the month, making a reasonable estimate based on surrounding stations, or even leaving the unrealistically warm guess at the start of the new series but looking at three years instead of limiting it (wrongly) to two years would make an adjustment necessary.

Throughout dFDB 2014, the analytical choices made by the NZCSC team have the effect of reducing the warming trend, and thus minimising the appearance of the very real warming NZ has experienced over the last century. Very convenient choices given their ideological stance on climate change, a cynic might note.

Quite apart from the methodological issues — which are undoubtedly huge — dFDB 2014 makes no reference to the Eleven Station Series (11SS) derived by NIWA from temperature sites that need no adjustments, presumably because it tracks warming at the expected level1 — that is, three times faster than dFDB 2014 finds.

One might speculate that if they had chosen to “audit” the 11SS — which has a strong warming trend in the raw station data2 — they would have been desperate to find adjustments to reduce that trend.

dFDB 2014 fails to acknowledge the existence of or address the issues raised by NIWA scientist Brett Mullan’s 2012 paper in Weather & Climate (the journal of the Meteorological Society of NZ), Applying the Rhoades and Salinger Method to New Zealand’s “Seven Stations” Temperature series (Weather & Climate, 32(1), 24-38), despite it dealing in detail with the method they claim to apply. Perhaps this is because it points out most of the egregious mistakes they made in their “audit”.

dFDB 2014 also fails to make any reference to sea surface temperature records around the country and station records from offshore islands which also support warming at the expected level — as does the well-documented reduction in ice volume in the Southern Alps.

Beyond any doubt, dFDB 2014 is a model of shoddy scholarship. How on earth did it get accepted for publication by Environmental Modelling and Assessment? An earlier version of dFDB 2014 was submitted to a much more relevant journal, Theoretical and Applied Climatology, but was sent back to the authors for substantial revision at least twice before being rejected. One can surmise that in that case peer review was an uncomfortable process for de Freitas, Dedekind and Brill because the peers being consulted were professional climatologists who understand the nitty-gritty of station adjustments.

At EMA, de Freitas seems to have found a more compliant editor and friendlier reviewers — so friendly that they were happy to allow an obviously and critically flawed paper through to publication. A few simple checks by the editors and reviewers should have raised warning flags.

They should have noted that de Freitas presents himself as lead and corresponding author, yet has no publishing track record in climate records and their homogenisation. He acts as front man for Dedekind and Brill — two men with no relevant academic affiliations or any publication track record — effectively prostituting his position at Auckland University to usher yet another rubbish paper through to publication3. If that wasn’t enough, then competent reviewers should have noted the obvious critical flaws and demanded changes.

As an example of ideologically-driven data torture, A Reanalysis of Long-Term Surface Air Temperature Trends in New Zealand is hardly unusual in the world of climate denial. What makes it stand apart is that such a poorly put together and politically-inspired effort has made its way into the peer-reviewed literature. That is a sign of a gross editorial failure by Environmental Modelling and Assessment, and it should be immediately withdrawn. Meanwhile, the NZ temperature record will continue to show what it always has – substantial and highly significant warming over the last 100 years.

  1. The level demonstrated by NIWA’s re-working of the benchmark Seven Station Series, 0.91ºC per century since 1909 []
  2. A powerful argument why there should also be one in any homogenised 7SS. []
  3. See “Pal Review“, and the Maclean, De Freitas & Carter saga for other examples of de Freitas playing fast and loose with the accepted conventions of scientific peer review. []

55 thoughts on “NZ cranks finally publish an NZ temperature series – but their paper’s stuffed with errors”

  1. Hopefully somebody will write to the journal demanding the withdrawal of the rubbish.
    The “benefit” of cause of the denier side having put their rubbish to the curb for all to “look through” is the fact that it so clearly demolishes their own position. 🙂

  2. It’s actually quite amusing to follow the comments at Treadgold’s place. Much huffing and puffing about RS93 (which is explicit about using symmetrical periods before and after an adjustment, and also states that periods longer than 2 years are optimal – as Mullen (2012) demonstrates) but a great deal of avoidance of the elephant in the room: the max/min issue.

  3. Now they have their paper published, let’s see them defend it – should be entertaining!

    BTW, there is an excellent popsci book by Will Storr – The Unpersuadables: Adventures with the Enemies of Science – that I can recommend to anyone seeking insight into the pyschology of denial.

    It includes the following quote:

    ‘And then he said to me – this is honestly true – he said to me, “Well you can prove anything with facts, can’t you?”’

        1. No, but I thought the court case centred on whether NIWA had followed procedures rather than matters of science, which the judge determined that wasn’t for the court to decide. Am I mistaken?

          1. The CSET seemed to want to make it about the methodology of series construction, and submitted their “audit” as evidence. NIWA had no choice but to prepare and submit a detailed rebuttal to that, and a considerable amount of scientist time was spent doing it – time that wasn’t included in the costs judgement against the CSET. But as you say, the judge refused to rule on matters of science – a long-standing legal convention.

  4. The aim of denialists like NZCSET is not to win the scientific argument; they know they lost that decades ago.

    Instead, they aim to confuse the general public and thereby forestall the imposition of effective measures to “green” the economy and challenge the status quo that they themselves benefit from.

    This strategy is laid out, in explicit detail, in a recent presentation to oil and gas executives by fossil fuel lobbyist Rick Berman:

    You don’t have to win people over to your side. You just have to confuse them.

    Often, (Berman) explains, they’ll encounter people who are “overwhelmed by the science and ‘I don’t know who to believe.’ But, if you got enough on your side you get people into a position of paralysis about the issue … you get in people’s mind a tie. They don’t know who is right… the tie basically insures the status quo.”

  5. NZ glaciers seem to be retreating at significant rates, comparable to countries with very significant warming trends. Glacier retreat since 1850 on wikipedia makes interesting reading.

      1. AndyS. Read the article. NZ has similar rates of temperature increases and glacier retreat to a variety of countries in the article including parts of Canada, northern europe and asia and latin america. Not exclusively so, but there is a general similarity that is interesting. Typically 2 – 4 kms of glacier retreat since about 1880.

        Of course there are regional factors, and exceptions but overall patterns are more illuminating and significant. Glaciers are just another reason to be suspicious of the CSC paper.

        1. NZ has similar rates of temperature increases and glacier retreat to a variety of countries in the article including parts of Canada

          Which rates are you using? NIWA, this paper, or something else? That is why I ask the question.

          Anyway, NZ glaciers are quite different to those in Canada. There are glaciers in the Yukon that are 100km long and 20km wide

  6. AndyS. It is obvious I meant the NIWA temperatures.The similarities between NZ and the other countries are obvious in the article, and I pointed out it is not exact.

      1. AndyS. Canada has some long glaciers, but that is not the point. The point is rates of glacier retreat. Rates of retreat of NZ glaciers and those in Canada and Northern Europe and Asia seem comparable, typically 2 – 4 kms per century. Roughly comparable.

        This suggests NZ is warming at similar rates to these northern hemisphere countries. They have robust warming of roughly 1 degree celsius per century.

        My comparisons are obviously rough, but rates of glacier retreat in NZ suggest more than the 0.3 degrees warming the climate science coalition claim.

            1. Yes I was well aware that was the one you were referring to – the one that runs from 1659. Not actually comparable with the NZ series is it? – Bit disingenuous to average from 1659 to the present and compare that with 1880 to the present don’t you think? Oh! but if you take the HadCET from 1880 to the present, I think you will find that its way bigger than 0.3 per century – more like 1 degree of warming over that last Century.
              But that doesn’t fit your “argument” does it andy – far better to lie..
              “There are lies, damned lies, and Statistics” Disraeli said that and he knew what he was talking about.

            2. From Clive Best’s article (with graphs to support)

              The only significant continuous trend is an apparent 0.026C/decade recovery from the little Ice Age over the last 360 years. This has not changed or accelerated. There is no hockey stick! In fact there has been no change whatsoever in UK average temperatures since 1940 !

              So, I make 0.026 degrees per decade at (drum roll please) 0.26 degrees per century

            3. AndyS. What is all this nonsense about hadcet and Clive Best? Stick with the official record like Hadcrut. Britain has had far more than 0.3 degrees celsius over the last 100 years. More like 0.8 degrees.

            4. This nonsense called HadCET is the longest temperature series in the world and is confined to the UK.
              This nonsense is maintained by the Met Office

              HadCRUT is a global temperature series

            5. andy what you say is only half true, and half truths are worse than no truth at all. I showed you above the basic error – If you were to refer back to the original data of HadCET, as depicted in the link I gave, (published by the UK Met office) you could see at a glance that there is WAY MORE than 0.3 degrees warming over the past century! Unless of course you are incapable of reading anomaly graphs.

            6. The Met Office shows a kick upwards of around 1 degree around 2000 then this drops back in the more extended time frame shown in Clive Best’s article
              He has fitted linear and Fourier trends to two graphs.

            7. lol! what a muppet!
              And I thought you were a math major!
              You are aware of the limitations of Fourier analysis in such a time series and the assumptions?

              andy the drum roll is for the deceased when they are buried, as that to which you refer, needs to be.
              Again I repeat, comparing 360 years with 100 does not compute. The hockey stick unfortunately exists.

            8. I look forward to your analysis of HADCET.
              Obviously, just an eyeball of the graph is sufficient.

              This maths stuff is highly overrated.

            9. andy – are you completely incapable of English comprehension? I have already alluded to the fact that treating HadCUT or any other such to Fourier analysis was incorrect for some very basic reasons:
              1.Fourier Series deal with functions that are periodic over a finite interval.
              2.Fourier Series are useful if
              (a) the function really is periodic, or
              (b) we only care about the function in a finite range.

              This is not to deny the value of Fourier analysis. But temperature data such as this are NOT periodic! therefore to fit a sinusoidal curve to it is nonsense. Furthermore Best takes 360 years and then says that the last 100 years are same! Absolute rubbish!
              ps I too have more than a smattering of maths and have worked in the Research branches of Dept of Statistics and The Dairy Board. and taught at tertiary level
              So don’t try your bullshit lines here.

            10. Here you go..

              “The temperature trend for 1900–2004 has been calculated from the annual values in Figure 2 using the restricted maximum likelihood technique (REML; Diggle et al., 1999) to take account of the uncertainties of the annual values and to allow for autocorrelation in the residuals from the fit. The best estimate, 0.077°C per decade, has a ±2σ error range of ±0.040°C per decade and is statistically different from zero at the 1% level, despite the uncertainties in annual mean CET. This is because the uncertainties in CET are not coherent on long time scales. For example, calibration biases vary on time scales of a few years because thermometers are changed; urban warming biases have been compensated for and their uncertainties (Table V) are much smaller than the integrated trend since 1900 (0.80°C). If the REML is applied to the series in Figure 2, but with the annual uncertainties constrained to be zero, then the ±2σ error range of the trend remains ±0.040 ° C”

              NOTE THE TREND FOR THE LAST CENTURY ANDY 0.8 ° C ± 0.040 ° C. NOT 0.3.


  7. Glaciers across the planet have been retreating since the last ice age, but there are retreats and advances all the time. Partly this is due to temperature, but also precipitation.

    For example, from the 1980s to about 2000, New Zealand and Norwegian glaciers advanced.

        1. Jonny Knoxious is a squalid liar, and thus a fitting contributor to Slater’s vile blog, but, Manfred old boy, tell us again of your mighty victories in the High Court against NIWA…

            1. Norway and NZ are similar in the sense that they are both long thin countries with a centre spine of mountains that are very exposed to westerlies that dump a lot of snow. Presumably some of this glacial advance and retreat in both countries can be ascribed to changes in Westerly airflows

            2. AndyS. Some of the short term decade scale changes in glaciers may be due to weather events and westerlies.

              Most of the longer term glacier retreat and thinning in NZ and Canada since about 1920 is attributable to global warming. According to the people who know. I refer you back to the article “glacier retreat since 1850”

  8. As an aside, from the source of your graph, note that over 2011/13, there has been a net gain at Rolleston glacier, one of the two monitored NZ glaciers. The other lacks 2013 data.
    Rolleston -429 +740 T. Kerr and colleagues (NIWA)

    Net mass balance gain of 311 mm w.e.

        1. But, think what you’re missing out on, old boy – for starters, I can help you with those “tricky powers of 10” that stuffed up your NIWA critique… or with your belief that the Earth only conducts heat in one direction!

          Also, think of the fun we can have, in proving that a new Ice Age is imminent…

            1. Yep, Bob, best you head back to Treadgold’s Swamp, where the locals wouldn’t know Shi’ite from Sunni…

    1. Rolleston is quite high on Mt Rolleston (Arthur’s pass) and very exposed to westerly snows. Temple Basin ski area, just across the valley, gets huge dumps

      Contrast the Tasman Glacier. This is experiencing rapid retreat, but is west of the divide, so gets less snow than the Franz and Fox
      It is also long and much of the lower part of the glacier is quite flat and also covered in moraine, which must have a heat sink effect and accelerate ice melt.

      Not that I am a glaciologist, but I note these things on my travels, and these two glaciers have had my paw prints on them

      1. Actually, rock cover on flat glacier sections acts as an insulator, slowing down ice melt. At Fox you used to pass large lumps of remnant rock-covered ice on the walk from the carpark to the glacier snout. Haven’t been there for a good few years, so can’t swear they’re still there…

        What’s really driving the rapid retreat of the Tasman is the pro-glacial lake. It’s eating away at the base of the ice, and short of an immediate return to the depths of the last ice age and a huge increase in snowfall up in the névé, the lake is going to extend back up the valley as far as the Ball Hut — sooner rather than later.

        1. On the moraine issue, I think you are correct, Gareth.
          GNS have quite an instructive page on glaciers:


          Because of their darker colour, rocks absorb more of the sun’s radiation than the ice and contribute to faster warming and melting. However if the rocks form a complete layer more than about 10cm thick, then they insulate the ice from the sunlight and slow down the melting process.

          I am pretty sure that most of the moraine on the Tasman is a lot thicker than 10cm

          The GNS page also concurs that the terminal lake accelerates the melt

          I’m not sure what the rock feature you describe is called. It might be an “erratic”

Leave a Reply