Today a paper on the New Zealand temperature record (NZTR) was accepted by the journal Environmental Modeling & Assessment. Submitted in 2013, we can only imagine the colossal peer-review hurdles that had to be overcome in gaining acceptance for a paper that refutes the national temperature record in a developed country. The mere fact of acceptance attests to a fundamental shift in scientific attitudes to climate change, but expect strident opposition to this paper.
The authors present first a concise observational history of the NZTR, remarking that the established national record was a product of early methodology, then reconstruct an homogenised dataset using the peer-reviewed adjustment standards of Rhoades & Salinger, 1993 (RS93).
A Reanalysis of Long-Term Surface Air Temperature Trends in New Zealand was produced by principal author C.R. de Freitas with M.O. Dedekind and B.E. Brill.
Abstract
Detecting trends in climate is important in assessments of global change based on regional long-term data. Equally important is the reliability of the results that are widely used as a major input for a large number of societal design and planning purposes. New Zealand provides a rare long temperature time series in the Southern Hemisphere, and it is one of the longest continuous climate series available in the Southern Hemisphere Pacific. It is therefore important that this temperature dataset meets the highest quality control standards. New Zealand’s national record for the period 1909 to 2009 is analysed and the data homogenized. 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.
The authors describe their intention as: “… to derive a modernized New Zealand Temperature Record (NZTR) providing a 100-year time series of mean monthly land surface temperature anomalies.” They claim to apply the method set out in RS93 exactly as described, “without adjusting it in any way,” noting it has been the adjustment method of choice in New Zealand climatology for over two decades and “we are unaware of any serious criticism or dispute regarding it.”
In the introduction, they make the point that “on the face of it, New Zealand’s long-term mean temperature has remained relatively stable at 12.6 °C over the past 150 years.” Which means that any new conclusion of warming must have solid grounds.
NZCSC media release
The NZ Climate Science Coalition has released a statement under a striking headline:
New paper finds no significant 20th century warming for New Zealand
A research paper on the homogenisation of the temperature record in New Zealand, reducing the current official warming rate of 0.9°C per century to 0.3°C per century, has just been published in the international scientific journal Environmental Modeling & Assessment.
The paper addresses the values of the data adjustments required during 100 years of the Seven Station Series, which is recognised as being representative of New Zealand as a whole. It also considers corrections to station data contaminated by vegetation growth, urbanisation and other factors.
The New Zealand historical temperature trend has not been addressed in the scientific literature since the first Seven Station Series was published by M.J. Salinger in 1980. At about the same time, a paper by J.W.D. Hessell called into question the quality of the New Zealand historical weather data used in the series.
The new paper builds on both viewpoints by applying modern techniques to correct sub-optimal raw data and to recalculate the 1980 adjustments. The method used for recalculations was that described in the leading New Zealand paper by Rhoades & Salinger (1993).
Lead author Chris de Freitas commented: “Regional and national temperature trends are widely used for a large number of societal design and planning purposes and it is important that they should be as reliable as modern methods allow.”
He added: “New Zealand provides one of the longest continuous climate series in the Pacific Ocean as well as one of the longest in the Southern Hemisphere. This means our trends are of ongoing interest to a wide audience of scientists.”
The paper finds that New Zealand warmed over the 20th century by 0.3°C, which, allowing for accepted margins of error, means that there has been no significant warming.
CO2 Science review
Dr Craig Idso, Director of CO2 Science, has reviewed a pre-release copy of the paper. This is his report.
In introducing their recent study of the subject, de Freitas et al. (2014) write that “a homogenized New Zealand national temperature record has only once appeared in the literature,” citing the one produced by Salinger (1980). They report, however, that it was based on a measurement technique that was significantly improved by its author and a colleague (Rhoades and Salinger, 1993) over a decade later. And although they state that “applying that improvement could have a significant effect on trends,” they indicate that such an improved trend for New Zealand “has never previously been published.”
The three New Zealand researchers thus set out to fill this void by applying the measurement technique described by Rhoades and Salinger to data for the period 1909-2009. And they did it, in their words, “exactly as they [Rhoades and Salinger] describe, without adjusting it in any way,” although they say they corrected for “the contamination of raw data identified in the refereed literature (Hessell, 1980)” and for “shelter-contaminated data.” So what was the final result?
De Freitas et al. report that, whereas the previous analysis yielded a trend of 0.91 ± 0.30°C per century, their analysis—which used updated measurement techniques and corrects for shelter-contaminated data—produces a trend of only 0.28 ± 0.29°C per century, which is a heck of a lot less than what had previously been believed to have been the case.
The significance of de Freitas et al.’s work is two-fold. First, the authors report that the old, contaminated data with the inflated warming trend has been “widely used as inputs for societal design and planning purposes” all across New Zealand. Second, de Freitas et al. note these data are “extensively used in hindcast verifications for regional and local models.” However, as the saying goes, “garbage in equals garbage out.” Therefore, at best, the corrected New Zealand temperature trend, which is three times smaller than the uncorrected version, calls into question all results, findings, conclusions and policies built upon or derived from the old contaminated data record. And at worst, it invalidates them.
Given the great importance of starting with the proper baseline, one would hope that with so much at stake in terms of economics, personal freedoms and governance much greater care and scrutiny would be applied to ensuring the quality and reliability of near-surface air temperature records. But obviously, such has not been the case for New Zealand. And it begs the question as to where else temperature records might be less than par.
Dr Idso notes that the paper finds warming just one-third of the ‘official’ century warming of 0.9°C and claims it “calls into question all results, findings, conclusions and policies built upon or derived from the old contaminated data record. And at worst, it invalidates them.” (emphasis mine). But I’m really not sure this applies to policy.
Our national policy decisions have been guided by IPCC recommendations, which are formed from global average temperatures, not local temperatures. One very good reason for this is that nobody knows reliably how to translate a degree of warming into weather effects, so forecasts are based on generalisations and covered all over with ‘might,’ ‘perhaps’ and ‘could’. However, I’m advised by well-informed and shrewd people that NIWA does foretell the future on its own account based on magically matching its national databases to IPCC forecasts, so if local temperatures have been overstated by 300%, errors will occur. Then territorial authorities will make mistakes in over-estimating future risks.
Anyone familiar with local policy and its origination would be most welcome to provide elaboration of this point.
———————
I spoke with the authors; they won’t want me to say this, but their humility strikes me mightily (though of course they make no show of it). They exhibit no hint of swagger in doing this work, simply telling me that they see it as “just one more paper for the open-minded to consider.”
That is admirable.
Views: 377
Well that’s bound to rip a few nighties at NIWA.
Hmm … he who laughs last, laughs longest.
NIWA uses a regional model which is downscaled from a group of CMIP5 models approved by the IPCC. Then it tweaks and changes the model until it can produce a reasonable hindcast of past New Zealand temperatures. The past temperatures are as set out in the 7SS.
Once the NIWA regional model has shown it can hindcast something close to 0.91°C for the 7SS over the last century, it is regarded as ready to start foretelling the future. It is then used all over the country predicting floods, droughts, sea levels, storms, etc and having a major effect on regional planning.
We now know that NIWA’s NZ model is calibrated to run far too hot. Like three times as hot as it should be. All those Regional plans are wrong.
Nice summary, thanks. Where does NIWA provide details of all this?
Apparently, too, we have to grasp the fact that Australia and New Zealand’s long, fairly lonely records heavily influence Pacific trends, which loom large in the global picture. So this reanalysis might have a broad reach.
>”NIWA uses a regional model which is downscaled from a group of CMIP5 models”
That’s a statistical downscaling:
https://www.niwa.co.nz/climate/research-projects/regional-modelling-of-new-zealand-climate
Their Regional Climate Model (RCM) is UKMO’s PRECIS:
http://www.niwa.co.nz/our-science/climate/information-and-resources/clivar/impacts
“The model requires atmospheric inputs from a version of the United Kingdom Met Office GCM called HadAM3P, or from reanalyses such as ERA-40.”
From ‘Regional modelling of New Zealand climate’ linked above:
First, detailed projections and data sets of future climate change are produced from regional climate models and from statistically downscaled global models. These projections can then be used to drive other environmental models that address issues relevant to water resources, tourism, and urban and coastal infrastructure.
The key steps involve:
#1 validating and improving NIWA’s regional climate model (RCM), and better quantifying natural variability of climate;
#2 generating a range of RCM projections of future New Zealand climate under different emissions scenarios, forced by different global climate models;
#3 improving statistical downscaling of projected New Zealand changes from a large set of IPCC Fourth Assessment global models. This will place the more limited sample of the regional model that are runs in to a broader probabilistic framework;
# 4 [omitted]
# 5 driving a range of environmental models with climate scenario data, in order to improve knowledge of how the cryosphere, rivers and sea-level could change under global warming this century; and
#6 widely disseminating information and data sets of climate change to the end-user community, with a best-practice example being developed in collaboration with Auckland Regional Council.
https://www.niwa.co.nz/climate/research-projects/regional-modelling-of-new-zealand-climate
Australis has mixed #1 and #3 but those are quite different exercises. However, “validating” the RCM in #1 could include what Australis describes possibly, although I’d like to see verification of that. Keep in mind that an atmospheric model is of the troposphere, not just the surface. Validation would have to be against radiosondes and satellites. If those disagree with the surface series then that’s problematic for validation. Satellites don’t go back to the contentious era but radiosondes (HadAT) go back to1958:
http://junksciencearchive.com/MSU_Temps/HadAT2.png
But if NIWA do actually calibrate their RCM against 7SS at the surface (and 11SS as per CRUTEM4) as Australis asserts, then their steps #5 and #6 are “garbage out”, as Craig Idso puts it.
[Thanks, RC, this is useful information. I knew it would be technical, so I wouldn’t understand it, and indeed I don’t, but I’ll keep reading it and something will sink in. It usually does. Cheers – RT]
>”But if NIWA do actually calibrate their RCM against 7SS at the surface (and 11SS as per CRUTEM4) as Australis asserts,”
They don’t from what I can gather. Refer:
‘Regional climate modelling in New Zealand: Comparison to gridded and satellite observations’
Ackerley, Dean, Sood and Mullan (2011)
Abstract [abbrev.]
Climate statistics from the model for 1980-1999 are compared with gridded observations. The geographical distribution of maximum and minimum surface air temperatures compare well with the gridded observational data (spatial correlation values >0.9) with low temperatures in upland areas and higher temperatures in lowland and northern areas. However, temperature biases are also evident, with maximum surface air temperature being too low and minimum surface air temperatures too high
2.2 Experimental setup
Two experiments were undertaken with the regional climate model:
1. European Centre for Medium Range Weather Forecasts Re-Analysis, lateral boundary condition data (ERA-40): The RCM was forced at its boundaries by ERA-40 data (see Uppala et al., 2005 for more information on ERA-40) to give the best estimate of the past atmospheric circulation across New Zealand. This run will be referred to subsequently as RCM1. The RCM1 run did not use the chemistry scheme and had the sulfur cycle switched off, as the ERA-40 boundary data did not contain the necessary sulfate aerosol. Values for the domain Sea Surface Temperatures (SSTs) were taken from the HadISST1.1 data set (Rayner et al., 2003). This simulation used a Gregorian calendar (including leap years).
2. GCM lateral boundary conditions: In this experiment the global GCM was run with observed SSTs (HadISST1.1) with the atmosphere free to respond to those SSTs. Lateral boundary conditions were generated by the GCM (HadAM3P) for the RCM in this run, and will be different from those in RCM1. This run will be referred to in the text as RCM2. The RCM2 run included the chemistry scheme, which represents the sulfur cycle (see Ackerley et al., 2009 and Jones et al., 2001). The use of the chemistry scheme improves the radiative transfer processes of the model. This simulation used a 360-day calendar (12 months of 30 days).
Both RCM1 and RCM2 were initialized on 1st December 1969 and run until December 1st 2000. Averages were made between the years 1980 to 1999 to follow the convention used in the IPCC AR4 report (Solomon et al., 2007) and to avoid the early years of the simulation where errors may occur due to the initial ‘spinning up’ of the model. […]
2.3 Observational data
2.3.1 Virtual Climate Station Network (VCSN) data. In this study, we use the gridded fields of the Virtual Climate Station Network (VCSN) at NIWA, which was produced from observed station data using the methods described in Tait et al. (2006) and Tait (2008). We compare the daily maximum temperature (Tmax), daily minimum temperature (Tmin) and the daily accumulation of precipitation from the VCSN data with the RCM output over the 1980–1999 period. […]
For temperature, Tait (2008) determined the accuracy of daily interpolations by selecting 20 validation sites which were not included in the stations used for the interpolation. For Tmax the daily mean RMSE was 1.2 °C and for Tmin 1.6 °C. The largest errors occurred at high elevation sites. Errors for seasonal or annual means were not provided, but if the errors are assumed to be random then the values stated could be expected to be reduced in magnitude by the square root of the number of independent days used in any average.
3. Results
3.1 RCM1 compared to gridded and satellite observations
3.1.1 Maximum surface air temperature (Tmax)
The annual mean Tmax for RCM1 and the observations can be seen in Figures 2(a) – (b). The RCM captures the geographical distribution of the temperatures well (the spatial correlation between VCSN and RCM1 data is 0.96) with the coldest values in mountainous areas (such as the central North Island and the southern and western South Island) and the warmest temperatures in lowlying areas (such as the northwestern North Island and the eastern South Island). Similarly, the summer (DJF, Figures 2(d) – (e)) and winter (JJA, Figures 2(g) – (h)) distribution of Tmax broadly matches those of the observations (spatial correlations of 0.92 and 0.97, respectively), albeit with warmer (cooler) temperatures in DJF (JJA).
However, the annual values of Tmax are generally too low throughout much of New Zealand (see Figure 2(c), which includes the New Zealand mean difference along with Figures 2(i) and (f)), with the coldest biases in the South Island in JJA (Figure 2(i)). The North Island is also generally cooler in the model than the VCSN data, but there are warm biases in northeastern areas which extend throughout much of the lower North Island and the far north of the South Island in DJF (when they are at their maximum, see Figure 2(f)). These biases in temperature appear to be greater than the RMSE error expected for the VCSN gridded data described in Section 2.3.1.
3.1.2 Minimum surface air temperature (Tmin)
The modelled and observed values of annual (Figure 3(a) – (b)), summer (DJF, Figures 3(d) – (e)) and winter (JJA, Figures 3(g) – (h)) mean Tmin both display a similar geographical distribution of temperatures. In a similar manner to Tmax (see Figure 2), the lowest values of Tmin are found in the upland areas (central North Island, southern and western South Island) in JJA and the warmest temperatures can be seen in the lowland areas (northern North Island and eastern South Island) in DJF. The spatial correlations between the VCSN and RCM1 data are also very high (?0.96) for Tmin (see Figure 3).
However, unlike Tmax there is more of a tendency for warmer temperatures in the model, compared to the VCSN data, for the annual mean Tmin (see Figure 2(c)), which implies that the diurnal temperature range is likely to be too small in RCM1. This is particularly evident in the North Island, where Tmin is generally >1oC warmer in the model than the observations (except in the middle of the central North Island during JJA, see Figure 3(i)). The South Island however, has a mixture of cold and warm biases throughout the year. Warm biases dominate many areas of the South Island during DJF (Figure 3(f)), with colder biases more prevalent during JJA (Figure 3(i)).
Figure 5: The annual, New Zealand mean (a) Tmax (oC), (b) Tmin (oC) and (c) daily mean precipitation (mm day-1) from VCSN (black lines), RCM1 (red lines) and RCM2 (yellow lines) data between 1980 – 1999. The correlation values (given in the key for each graph) are between the VCSN time series and each of the model time series separately. Statistically significant correlations (p≤0.05 using a student t-test) are given an asterisk.
References
Tait, A., Henderson, R. and Turner, R. and Zheng, X.G., 2006. Thin plate smoothing spline interpolation of daily rainfall for New Zealand using a climatological rainfall scheme. International Journal of Climatology, 26(14), 2097-2115: DOI:10.1002/joc.1350
Tait, A.B., 2008. Future projections of growing degree days and frost in New Zealand and some implications for grape growing. Weather and Climate, 28, 17-36.
http://www.metsoc.org.nz
2012_321_3-22_ackerley.pdf
Virtual Climate Station Network (VCSN) is this NIWA product:
https://www.niwa.co.nz/climate/our-services/virtual-climate-stations
“Virtual climate station estimates are produced every day, based on the spatial interpolation of actual data observations made at climate stations located around the country.”
# # #
Question is: how does “observed station data using the methods described in Tait et al. (2006) and Tait (2008).” which is virtual climate station (VCS) output, relate to the 7SS?
I suspect little relation going by Figure 5 VCS Tmax and Tmin. Although the Mean of the two may correspond perhaps, hard to tell unless the respective profiles are graphed one against the other. Figure 5 is only 1980 – 1999 when there was very little warming in the 7SS anyway and there is certainly no linear warming in either VCS TMax or TMin over that period. No prominence given to this by NIWA I note, inconvenient perhaps.
Tait et al. (2006) and (2008) should confirm that VCS originates from raw CliFlo site data rather than homogenized location series. Seems probable given the spatial coverage and quote.
I don’t know the date VCS commences. Both RCM1 and RCM2 were initialized on 1st December 1969 but with ERA 40 (RCM1) and HadlSST1.1 (RCM2), not with VCS. It would be the ERA 40 profile that should be compared to the 7SS (and 11SS). What then is the ERA 40 input? Raw or homogenized?
Again, the contentious 7SS era is prior to 1970 so this exercise is not very helpful in respect to the 20th Century trend. But good to work through because it appears to place the 7SS on the outer i.e. not relevant to New Zealand regional climate modeling when it comes to the crunch.
[RC – as I mentioned, this info is technical and I find it hard to absorb in one bite. That means others will struggle, too. Now that’s fine, it’s a technical topic, but you make the bite larger by increasing the number of lines. Would you kindly paste text that will flow in the normal way, please? I’ve put this slab through NoteTab to join the lines and show what I mean. You’re helping to create a quite unique research tool here, but I’d ask you to respect the prime purpose of this blog, which is to make the information accessible. Thanks for your help. – RT]
BEST states warming for New Zealand:
0.97 ± 0.24 °C / century since 1960
2.41 ± 0.60 °C / century since 1990
http://berkeleyearth.lbl.gov/regions/new-zealand
Their Mean profile 1980 – 1999 looks nothing like Ackerley et al’s Fig 5 VCS TMax/TMin above, neither of which exhibits anything like what BEST returns. I haven’t heard of NIWA protesting to BEST about this.
NIWA’s VCS and BEST are similar products, but the respective results are very different.
I’ve been anticipating this paper for a while. Bob Dedekind some time back said something like “I’ve been very busy” to explain his absence from this blog. That gave me the clue that something was up, along with the obvious need for this paper.
Now it’s in the literature. No doubt there will be those in certain quarters who will gag on the citation. Jim Salinger claimed NZCSC made “lots of mistakes” in their ‘Statistical Audit of the NIWA 7-Station Review’, not actually specifying any of course. Now he will have to address this paper formally. The NZ judiciary, J Venning in particular, chose just to ignore the Statistical Audit. Perhaps he might take the time to read this paper and reflect on how he could have conducted his hearing differently.
But how will NIWA respond if they can’t reveal their own methodology?
Good work C. R. de Freitas, M. O. Dedekind, and B. E. Brill. Now go out and buy yourselves hazmat suits – you’ll need them.
“But how will NIWA respond if they can’t reveal their own methodology?”
Precisely. Fascinating. Get popcorn.
Hazmat suits or >>>>
http://visual.merriam-webster.com/images/sports-games/winter-sports/ice-hockey/goalkeeper_1.jpg
Heh, heh! Good one.
>”Our national policy decisions have been guided by IPCC recommendations, which are formed from global average temperatures, not local temperatures.”
Yes except NZ local temperatures (7SS, 11SS, plus another 2S) contribute to the CRUTEM4 global average for this neck of the woods, after NIWA adjustment:
CRUTEM4 Temperature station data
http://www.cru.uea.ac.uk/cru/data/temperature/crutem4/station-data.htm
NZ source: Homogenized series , NIWA, New Zealand
http://www.niwa.co.nz/climate/information-and-resources/nz-temperature-record
Therefore the local trend contributes to the global trend, Southern Hemisphere trend, and regional trend. All of which the IPCC includes in consideration.
“Therefore the local trend contributes to the global trend, Southern Hemisphere trend, and regional trend.”
It’s still hard to imagine that even overstating our trend by 300% would make much difference to the global trend. You disagree?
>”It’s still hard to imagine that even overstating our trend by 300% would make much difference to the global trend. You disagree?”
Not much difference, no. But in combination with other overstatements, yes. And in terms of the SH and regional trends, definitely yes. But which datasets?
Global sea level rise is skewed by one region north of Australia. New Zealand land is the only land for 12,500 km to the east and 4200 km to the west i.e. it typifies a surface area very much out of proportion to its land area in the land datasets. Until we see a NZ 7SS and 11SS) in CRUTEM4 at 0.3 C/century to compare with the current CRUTEM4 global, we don’t know what difference that would make.
Richard and friends, Sorry for being off topic (a bit) but I thought some of you might be interested in a little study I’ve done. It’s an empirical look at recent trends in the greenhouse effect, based on UAH and RSS temperature series, and NOAA’s outgoing LWIR records, followed by a determination of relative emissivity. Basically shows that the greenhouse effect is unaltered through the whole period, and therefore challenging the classical IPCC theory. Kind regards to all, Robin. http://www.kiwithinker.com/2014/10/an-empirical-look-at-recent-trends-in-the-greenhouse-effect/
Very good analysis Robin, thanks.
Yes very interested Robin, good work. Linked via The Hockey Schtick in the previous post here:
https://www.climateconversation.org.nz/2014/10/climate-research-needs-redirection/#comment-1201536
Fits very nicely with the vein of that post and thread. Something to refer back to where a lot of related stuff is all in one place. For example, I’ve annotated the Basic Energy Model (BEM) quote thus:
“This implies that GHG energy is returned to the atmosphere and space very quickly as latent heat of evaporation [see Kiwi Thinker/HS post upthread] ”
Better discussed there I think.
Publishing a paper makes much more sense than taking a case to the High Court.
The press release is incorrect though, 0.28 ± 0.29°C/century (95% CI?) does imply statistically significant warming. Remember that NZ is small and insulated by ocean, so is unlikely to have warmed as much as continents in higher latitudes. The estimate does look a bit low though, so the homogenisation and shelter correction methodologies could probably bear some scrutiny.
Shame NIWA didn’t let anyone apply scrutiny to their methodologies due to the fact they refused to let anyone know what methods they were using – if they ever had any. Pack of dishonest frauds the lot of them, along with those who supported them.
[Magoo, you’re an old friend here, so you get some latitude, but this ad hominem nonsense is beyond the pale. Please use more genteel (dare I suggest intelligent?) insults or I’ll be forced to clip your wings. – RT]
“Publishing a paper makes much more sense than taking a case to the High Court.”
Thank you, Simon—though you could probably notice they address different matters.
You make a good point about significance. I’ll check, or someone who knows will comment.
It’s most curious that you describe the likelihood of strong warming (actually, 50% more than the global average) as low, when NIWA scientists have for years confidently proclaimed strong past warming (0.91°C) while simultaneously predicting a future firmly ameliorated by the surrounding Pacific.
You finally remark that the trend should be increased, which contradicts your “unlikely” assessment. You should make up your mind.
>”Publishing a paper makes much more sense than taking a case to the High Court.”
It wouldn’t have been necessary if firstly NIWA, then the Judge, had just considered the merits of the evidence (e.g. the ‘Statistical Audit’). They didn’t, He didn’t. NiWA absconded, the Court failed its duty. So where does this paper place the Court now? And NIWA?
>”0.28 ± 0.29°C/century (95% CI?) does imply statistically significant warming”
CI -0.01 to +0.59 includes zero therefore implies statistically insignificant warming. If the CI was say +0.01 to +0.57 that would imply statistically significant warming.
>”NZ is small and insulated by ocean, so is unlikely to have warmed as much as continents in higher latitudes”
The difference is WHEN the warming occurs. We don’t have to go to the NH for that. NIWA’s 7SS and 11SS contribute to CRUTEM4 so we can compare the 7SS to CRUTEM4 SH:
http://www.woodfortrees.org/graph/crutem4vsh/from:1909
Not much happened until 1970. But for NZ, according to NIWA’s graph, 0.96 ± 0.29°C/century warming started well before 1970: That is odd.
https://www.niwa.co.nz/sites/niwa.co.nz/files/styles/large/public/sites/default/files/images/0007/108925/nztemp7_anom_annual_100ytrendline_0.jpg?itok=_VyC66C-
The AU contribution to CRUTEM4 (HQ series) just looks like CRUTEM4 SH, not NIWA’s 7SS:
http://jonova.s3.amazonaws.com/guest/barnham/temp-mean.aus.0112.13002.png
In other words, NIWA’s 7SS is odd in SH terms prior to say 1970, let alone NH. And upthread you can see that it isn’t used for regional climate modeling purposes. NIWA prefer their VCS and ERA 40. The VCS doesn’t exhibit a linear trend (no warming) after at least 1980 and to 2000, which again is odd in SH terms. Neither does the 7SS in recent times so that’s corroboration rather than oddity, The contentious 7SS era is the very odd period prior to 1970.
Bizarre results in BOM’s ACORN (successor to HQ) getting press in AU, along with their “disappearing” of 118-year-old temperature records e.g.
http://www.dailymail.co.uk/news/article-2813394/Heat-Weather-Bureau-MP-accuses-wiping-118-year-old-temperature-records-justify-claims-climate-change.html
But ‘nuther story.
>”NIWA’s 7SS and 11SS contribute to CRUTEM4….”
‘Eleven-station’ series temperature data
https://www.niwa.co.nz/our-science/climate/information-and-resources/nz-temp-record/temperature-trends-from-raw-data
Similar story either side of 1970. If you look at Ruakura for example, it’s effectively flat after 1970. But apparently (in bold) “The warming trend over the 77 year period of this series is close to 1°C”.
>”NZ is small and insulated by ocean, so is unlikely to have warmed as much as continents in higher latitudes”
Here’s the insulation:
HadSST3 SH
http://www.woodfortrees.org/graph/hadsst3sh/from:1909/plot/hadsst3sh/from:1909/trend
I depends what you mean by “statistically significant”.
Regardless of the technical details, the error margins are of similar magnitude to the warming, which hardly makes the warming “significant” according to a layperson’s definition
Anyway, congratulations to the authors for getting this work into the peer-reviewed literature.
“Significant” has a specific meaning in statistics. Assuming the analysis is correct, there is only a 2.5% probability of there having been no warming over the past century. It is therefore correct to say that there has “almost certainly” been some warming over the past century.
The paper appears to be paywalled, but I wonder how many periods were considered when determining adjustments. The lower the number of periods, the less likely that an adjustment will be made, which would likely result in a smaller increase in average temperature. In the court case, NIWA stated that CSET used an insufficient number of periods, which resulted in a flawed analysis. I suspect the authors did not correct this or even examine the effect of period length on temperature.
I was alluding to the discussions that Doug Keenan had around time series and the concept of significance with respect to these. I don’t pretend to know very much about this topic so tend to keep my trap shut. However, the term “statistical significance” doesn’t in general mean what most people think it means
“The press release is incorrect though, 0.28 ± 0.29°C/century (95% CI?) does imply statistically significant warming. ”
I’m not sure if you’ve made a typo or not. It should read “0.28 ± 0.29°C/century (95% CI?) does imply statistically insignificant warming. ”
If the 95% CI includes zero, then it is accurate to state that the warming trend (whatever it is) is statistically insignificant at the 95% level.
But that isn’t the main focus of the paper, the main finding is that, when a proper RS93 anaylsis is performed, and taking UHI/sheltering into account (Aguilar) the resultant trend is one third that of the previously published 7SS, which used an out-dated method that didn’t take ANY statistical significance checks into account, nor corrected for acknowledged sheltering and/or UHI effects.
Simon,
“I wonder how many periods were considered when determining adjustments.”
As I say in the post: ‘They claim to apply the method set out in RS93 exactly as described, “without adjusting it in any way,”’ The authors explain further:
“In dealing with a site change known a priori by way of comparison with “neighbouring” stations, RS93 faces the same measurement issues as did S81. Both papers consider statistical differences between before-and-after temperature averages at the target station and those at comparison stations.
In RS93, the measurement techniques for those differences contrast with those applied in S81 in the following respects:
1. RS93 uses monthly data whilst S81 used annual data.
2. RS93 takes short before-and-after periods (e.g. 24 months), whilst S81 used long periods (e.g. 10 years or more).
3. RS93 weights the averages of data from comparison stations based on relative correlation coefficients, whist S81 used unweighted averages.
4. RS93 excludes proposed adjustments that are not statistically significant at the 95 % level, whilst S81 did not measure confidence intervals for adjustments.”
NOTE: RS93 refers to [Rhoades, D. A., & Salinger,M. J. (1993). Adjustment of temperature and rainfall records for site changes. Int J Climatol, 13, 899–913.]; S81 refers to “the published 1980 paper [Salinger, M. J. (1980). The New Zealand temperature series. Climate Monitor, 9(4), 112–118.] in conjunction with the unpublished thesis elaborations.”
Your suggestion that “the authors did not correct this or even examine the effect of period length on temperature” is erroneous. As the above description methodology reveals, they include many more periods than NIWA does. In addition, they keep the overall period brief to avoid contamination by long-term changes, just as RS93 advises.
“Remember that NZ is small and insulated by ocean, so is unlikely to have warmed as much as continents in higher latitudes.”
So you admit that NZ’s warming should be lower than the 0.91°C/century of the previous 7SS, since that was about 50% higher than the average global trend?
Statistical significance cuts both ways. When (not if) global cooling really sets in sometime between now and around 2020, the upper CI bound will have to drop below zero to be statistically significant. See McKitrick (2014):
McKitrick, R. (2014) HAC-Robust Measurement of the Duration of a Trendless Subsample in a Global Climate Time Series. Open Journal of Statistics, 4, 527-535. doi: 10.4236/ojs.2014.47050.
http://www.rossmckitrick.com/
[Pre-emptively, anticipating Simon] The IPCC’s contention, citing Jones, Lockwood, and Stott (2012), that the solar recession wont offset GHG warming is a perpetual motion machine overturning the laws of physics. A reduced energy input to a system cannot produce more energy than at the original input level.
Andy,
“hardly makes the warming “significant” according to a layperson’s definition ”
That’s what I was thinking. We’re talking about less than half a degree over 100 years, when we can’t detect a whole degree of change on the bare skin. We may be somewhat accustomed to the scale of these changes, but to most people they’re quite profoundly underwhelming.
>”less than half a degree over 100 years”
Less than the accuracy of mercury thermometer readings (depending on the instrument – see below). This is coming up in the stoush across the Tasman where modern AWS “record” high temperatures are displacing 19th Century records to 0.5°C (those that haven’t been wiped from the record).
A mercury thermometer (LIG) damps the reading (and e.g. hysteresis) whereas a thermister or resistance detector (RTD) in a modern AWS logs to ± 0.1°C accuracy very quickly and continuously (every second):
http://www.delta-t.co.uk/product-display.asp?id=WS-GP2%20Product&div=Meteorology%20and%20Solar
BOM’s thermometers:
http://www.bom.gov.au/climate/cdo/about/airtemp-measure.shtml
‘The Metrology of Thermometers’
“Since we had this recent paper from Pat Frank [hotlink] that deals with the inherent uncertainty of temperature measurement, establishing a new minimum uncertainty value of ±0.46 C for the instrumental surface temperature record, I thought it valuable to review the uncertainty associated with the act of temperature measurement itself.”
“There are some enlightening things to learn about the simple act of reading a liquid in glass (LIG) thermometer that I didn’t know as well as some long term issues (like the hardening of the glass) that have values about as large as the climate change signal for the last 100 years ~0.7°C – Anthony”
“If the scale is marked in 1c steps (which is very common), then you probably cannot extrapolate [sic? interpolate?] between the scale markers. This means that this particular thermometer’s resolution is1c, which is normally stated as plus or minus 0.5c (+/- 0.5c)”
“Mercury forms a pronounced meniscus in a thermometer that can exceed 1c and many observers incorrectly observe the temperature as the base of the meniscus rather than it’s peak: ( this picture shows an alcohol meniscus, a mercury meniscus bulges upward rather than down)”
“25 years ago, very accurate mercury thermometers used in labs (0.01c resolution) had a calibration chart/graph with them to convert observed temperature on the thermometer scale to actual temperature.”
“Electronic temperature sensors have been used more and more in the last 20 years for measuring environmental temperature. These also have their own resolution and accuracy problems. Electronic sensors suffer from drift and hysteresis and must be calibrated annually to be accurate, yet most weather station temp sensors are NEVER calibrated after they have been installed.”
“Finally we get to the infamous conversion of Degrees Fahrenheit to Degrees Centigrade.”
http://wattsupwiththat.com/2011/01/22/the-metrology-of-thermometers/
So we have a division in the NZ 7SS (and many others) – LIG/RTD. The contentious era is under the LIG regime, producing by far the bulk of “less than half a degree over 100 years” where the accuracy is least (possibly). But I think we should be mindful of the respective issues, LIG vs RTD, as in AU sense too.
Gareth has written an unflattering piece on this paper. At least someone noticed.
Good old Gareth, gotta love him, the poor guy. He tries so hard, but has so little to back him up.
Simon:
No, read RS93 again.
What this means is that as you take longer periods, you end up including gradual effects such as sheltering growth or UHI, and also you increase the risk of including other, unknown site changes at any of the reference stations.
RS93 states quite clearly that using short periods of 1 or 2 years pre-and post will provide enough statistical significance to determine a result, provided monthly values are used, not annual.
[RS93] – “1 or 2 years before and after”
[Gareth] – “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.”
2 + 2 = 4 Gareth (or your ghost writer).
Classic! But, to set a good example, I won’t descend to calling him names, not even clot or dimwit. Aren’t I good?
We should emphasise, on this point, that using 4 years with monthly data is 12 times better than Mullan with his annual data. Twelve times!!
“RS93 uses four year periods for comparison…”
Yes, I suspect poor Gareth is confused about this. Does he mean four years in total, or four years before and after?
If four years in total, then he is correct, but that’s what we did, so he’s just confirmed we followed RS93.
If he means four years before and after, then he is wrong that RS93 uses that, or even suggests it. To prove your point, Mr Renowden, please provide us with the page number from RS93 where it says that.
Even the worked example in RS93 uses two years before and after.
Section 2.4, pg 906:
Poor Gareth.
>”We should emphasise, on this point, that using 4 years with monthly data is 12 times better than Mullan with his annual data. Twelve times!!”
Not in Mullen (2012):
2. Data and Method
(a) Station temperature data
The data used in this study are monthly temperatures, ……
The key issue to my mind (among other things) is the RS93 criteria – k Adjustment δ Contains zero? Valid adjustment? As for example Masterton 1920 downthread that NZCSC adhere to but NIWA and Mullen do not. The criteria is made plain in the ‘Statistical Audit’ Appendix, I don’t know how it could be made any more clear.
Gareth might grasp that one day.
[Gareth] – “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), ”
Search Publications: Applying the Rhoades and Salinger Method to New Zealand’s “Seven Stations” Temperature series
http://www.metsoc.org.nz/publications/journals
No Results
OK, found it under 2012
You’re going great guns, RC. Please continue; I want to exsanguinate Renowden’s malevolent post and this research is a good help.
Mullen (2012):
Figure 2: R&S temperature adjustments as a function of k for the May 1920 site change in the
Masterton reference series, using four comparison sites (Albert Park, Thorndon, Christchurch
Gardens and Taihape). The three lines show the effect of different ways of handling the missing
value at May 1920: leave May 1920 missing (red diamonds and line), estimate the missing value
(black), apply a naïve estimate for the missing value (green). The vertical red dotted lines denote
the 95% confidence intervals about the estimates (red diamonds). The horizontal dashed blue line is
the NIWA 2010 adjustment (of -0.21°C) based on comparing annual temperatures 1912-1919 with
1921-1927
‘Statistical Audit’ (see Supplementary Information for Masterton details):
Masterton
Table 3: Comparison between NIWA and R&S results
Site Label Site Name From To NIWA Adj R&S Adj NIWA sum R&S sum
Site 4
Waingawa (2473)
Feb 1912
Apr 1920
-0.21
0.00
-0.55
0.00
https://www.climateconversation.org.nz/docs/Statistical%20Audit%20of%20the%20NIWA%207-Station%20Review%20Aug%202011.pdf
Don’t know yet what the new paper says about Site 4 but yes Gareth, this is what it is all about.
>”see Supplementary Information for Masterton details”
Site Change in 1920 page 48 pdf:
‘Supplementary Information: Statistical Audit of the NIWA 7-Station Review’
https://www.climateconversation.org.nz/docs/Statistical%20Audit%20of%20the%20NIWA%207-Station%20Review%20Aug%202011%20SI.pdf
For the case of the 1920 adjustment, the results are:
k Adjustment δ Contains zero? Valid adjustment?
1
-0.11 ± 0.39 °C
Yes
No
2
-0.24 ± 0.25 °C
Yes
No
So the adjustment is not made.
Mullan (2012) is far from a refutation of RS93. In order to show that RS93’s two-year method is incorrrect, Mullan would have to prove statistically, and therefore mathematically, that k=2 is insufficient. This he has failed to do – all he has done is provide examples where the results change with longer time periods.
But this reinforces the valid point made in RS93 that gradual effects in other stations introduce inaccuracies with longer time periods.
So all he’s done is strengthen the case for shorter periods.
Should point out that BOM would not make the Masterton 1920 adjustment either for ACORN (-0.21°C NIWA vs 0.00 RS93) – they don’t adjust for less then 0.3.
RC,
“Not in Mullen (2012):
2. Data and Method
(a) Station temperature data
The data used in this study are monthly
temperatures, …”
My apologies. Yet, somewhere NIWA uses annual not monthly data in defiance of RS93. Where is it?
>”somewhere NIWA uses annual not monthly data in defiance of RS93. Where is it?”
‘Report on the Review of NIWA’s ‘Seven-Station’ Temperature Series’
Page 17 pdf- “The focus in this document is on annual mean temperature1”
“1 Mean temperature is defined as the average of the daily maximum and minimum temperature. Further
research will determine adjustments to monthly maximum and minimum temperatures separately, and
apply statistical methods (e.g., RHtests, Wang et al., 2007) to identify other change-points in the data.”
http://www.niwa.co.nz/sites/niwa.co.nz/files/import/attachments/Report-on-the-Review-of-NIWAas-Seven-Station-Temperature-Series_v3.pdf
[Gareth in comments] – http://hot-topic.co.nz/nz-cranks-finally-publish-an-nz-temperature-series-but-their-papers-stuffed-with-errors/#comment-45023
>”RS93…..states that periods longer than 2 years are optimal,,,,”
[Bob] – “If he means four years before and after, then he is wrong that RS93 uses that, or even suggests it. To prove your point, Mr Renowden, please provide us with the page number from RS93 where it says that.”
https://www.climateconversation.org.nz/2014/10/paper-adds-interesting-perspective-on-nz-temperature-trend/#comment-1211945
>”….as Mullen (2012) demonstrates) …..”
Mullen. inadvertently, doesn’t – just the opposite. As Bob points out here:
https://www.climateconversation.org.nz/2014/10/paper-adds-interesting-perspective-on-nz-temperature-trend/#comment-1211945
>”….a great deal of avoidance of the elephant in the room: the max/min issue.”
Must have the same affliction as the shrinking goats:
‘Climate change shrinks goats’
http://wattsupwiththat.com/2014/10/21/eye-roller-climate-change-shrinks-goats/
Mullen (2012):
“Both monthly minimum and maximum temperatures are
used on occasion, although most emphasis is
placed on the monthly mean temperature
(derived from the average of the daily
minimum and daily maximum).”
Yet another stone in the pyramid (mountain) of evidence that global warming is scientific pap. Why this is such a large block was stated at the very beginning of the thread, The New Zealand temperature record is one of the very few in the Southern Hemisphere. Thus it has an inordinate effect on the so called global temperature. As an aside, I wonder what records exist in South Africa and Argentina. We know about Australia. A significant record in either of the above (or both) nations would do much to broaden our knowledge of the Southern Hemisphere climate history.
As a second aside, I wonder how many of the global warmists realize the pyramids were tombs.
I left out Chile in possible Southern Hemisphere temperature records, and no, I’m not going to go back and rewrite the whole comment. It’s late in Marietta, GA, USA and I’ve had more then a few Sauvignon
Blancs.
I like Georgia. I was there earlier this year, during that really cold spell you had.
I also like Sauvignon Blancs, have one for me. 🙂
We finally have a rebuttal on what was ultimately a flawed paper: https://link.springer.com/article/10.1007%2Fs10666-018-9606-6
Abstract
de Freitas et al. (2015) (henceforth dFDB) report a trend of 0.28 °C per century over the period 1909–2009 for New Zealand land surface temperatures, from their reanalysis of a composite of seven long-term records. This is much lower than the warming trend of about 0.9 °C per century reported previously by other researchers and much smaller than trends estimated from independent sea surface temperature data from the surrounding region. We show these differences result primarily from the way inhomogeneities in temperature time series at individual stations due to site or instrument changes are identified and adjusted for in the dFDB paper. The adjustments reported in that paper are based on a method designed by one of us (Salinger), but use only a short (1–2-year) overlap period with comparison stations and consider only inhomogeneities in monthly mean (rather than monthly maximum and minimum) temperatures. This leads to underestimates of the statistical significance of individual temperature discontinuities and hence rejection of many valid adjustments. Since there was a systematic tendency for the seven-station sites to be relocated to colder locations as the early half of the twentieth century progressed, this rejection of valid adjustments produces an artificially low rate of warming. We therefore disagree with the trend calculations in the dFDB paper and consider there is no reason to reject the previous estimates of around 0.9 °C warming per century.
“We finally have a rebuttal…….”
Aaahahahahahaha….who’s the WE, Simon ? The royal WE ?….No….the right royal little band of lying troughers with a despicable but pathetic besmirchment of honest science from an honourable man.
These lying, corrupt clowns are just stubbing their toes, kicking Dr de Freitas’s gravestone.
Yes, a right wee, but nothing royal or Royal about it, Mack. Did you, by any chance, read a book by one John Fonte titled ‘Sovereignty or Submission’ ? This guy is worth reading, he can put into words better than anyone I know what We all know about scheming scamming manipulating perverted lying cheats.
No, I don’t read books any more, Jacob. The fiction is….well, fiction. The non-fiction is nearly always embellished with a bit of fiction…..and, I’ve come to realise, there’s non-fictional stuff that’s totally delusional fiction. ie incorrect, opinionated, and usually paranoid, nonsense. Alarmist climate books generally fall into that category.
Mack, so be it. It’s a free world, whatever people claim.