NIWA temperature series problems — Part 1
Richard Treadgold | May 24, 2010 | 5:14 pm
A survey begins
When we published our paper Are we feeling warmer yet? last November, criticising the Seven-station Series (SS), NIWA quickly produced what they call the “Eleven-station Series” (ES) to counter it. They went to the trouble of asking Dr Jim Salinger, recently dismissed and author of the original national temperature series, to help them create it, which makes us realise that nobody at NIWA understands how he produced the original figures.
According to the NZ Herald, Jim specially selected the stations for the ES.
Some people would be impressed by that news, but others find their antennae stiffening at the idea of “specially selecting” data for any reason. It might be justified, but then again, it might not. The situation calls for careful study. So what has happened? We can say a few things about it.
Scientists in the Coalition have had a look at the new series and found problems with it. Of course, they started with the page on NIWA’s web site that contains it.
Stations have moved
First of all, NIWA misled the country from the beginning, by claiming that these stations had not moved. NIWA announced the brand-new series on December 3 last year, in a statement to Scoop containing this unambiguous headline:
(b) measurements from climate stations which have never been shifted.
There’s no doubt what that means. But it isn’t true. I wonder how many people, supportive of NIWA, have read no more about this new series than that first press release and even now repeat the false statement that these stations have never moved?
But even on NIWA’s web site, they have recanted this statement and now say they chose eleven stations where “there have been no significant site moves for many decades.” As scientists seem to have it in their DNA to avoid anything resembling a straight answer, lest it either be turned against them, become “settled science” or, horrors, become a prediction, it is impossible to believe this statement was written accidentally. It’s significant they should have recanted it.
Now it becomes a matter of what “significant” means. For now, we’ll leave you to decide for yourself. There is more study to be done on this.
Coalition scientists have identified substantial site changes in no fewer than 6 out of the 11 stations. We’re still working through the detail and hope to offer a full analysis in time. What follows is a bit technical, but skim as desired.
Some of the sites that have shifted, contrary to NIWA’s assertion, are:
Campbell Is (6174)
A site change in 1956-57, as previous site had poor exposure and was shaded by surrounding hills for 4-5 hours per day. Shifted to lower site closer to sea (perhaps affected by SSTs), that is less shaded (1-3 hours per day). Evidence of a 10-month overlap of records, although using different instruments, as the autographic instruments were shifted on 1 April 1957 – Fouhy et al. (1992).
Gisborne Aero/AWS (2810)
Seven site changes were reported in the period Apr 1937 – Feb 1993 – Fouhy et al. (1992).
Invercargill Aero I68433 (5814)
Site change in 1950, Mayor complained about reliability of records in 1981 (reading lower than nearby stations), damaged by flooding in 1984 – Fouhy et al. (1992).
Queenstown I58061 (5446)
This has not moved and was not used by McGann et al. (1992) as a climate reference station due to lack of homogenisation (page 1). Sited in a heavily vegetated park, which affects readings, but thought to be representative – Fouhy et al. (1992). Incidentally, in later papers co-written by Salinger, this site shows the largest AGW effect in NZ [CORRECTION 29 May: was 'UHI effect'], followed by Hastings (while Napier shows none).
Queenstown
One of the Coalition’s scientists visited two of the sites, key early stations in the ES. Here are his photos and comments. First, here’s Queenstown:
Queenstown has not moved, but look at the nearby rocks, vegetation, car park and roads.
Those trees across the carpark, which tend to raise recorded temperatures more and more as time passes (because they reduce the wind speed over the weather station), undoubtedly give shelter from at least one quarter of the compass, perhaps more. The rocks, roads and paved parking areas are notorious for heating up, which is why a climate-class weather station should be sited well away from them. The cars themselves actually emit heated gases, in case you weren’t aware of that!
It isn’t visible in this picture, but Lake Whakatipu — a sizeable body of water — is less than 100 m from the site, contravening the specifications.
In a final indignity, we should point out that the Stevenson Screen needs cleaning and repairs.
Ruapehu (Chateau)
Here are two views of the Ruapehu site. The previously-used Stevenson Screen:
and the current AWS (Automated Weather Station):
Image (1) shows how close the Stevenson Screen is to the gravel road and rubbish skips. The original site for the start of the series was situated beyond the ridge in the background, closer to the Chateau and at a higher altitude. We can expect warmer temperatures to be measured at the current site, and must suggest that adjustments would surely be required by this move.
Considering that NIWA stated none of the eleven stations had been moved, one wonders what else is false about this series. NIWA claim that none of these stations have had adjustments applied to them, and the quality issues raised here question the wisdom of that.
Image (2) shows the current AWS. What influences might the rubbish skips, the gravel car park, the building housing the water treatment machinery, and the oxidation pond (behind the photographer) have on the recorded temperatures?
It’s hard to imagine much cooling.
The point of the ES produced so hurriedly by NIWA was to “prove” that New Zealand has been warming, without using the official national graph that we criticised so heavily.
Considering the faults we’ve found with just a few of the eleven stations, most of which would be expected to produce some spurious warming, it’s hardly surprising their “specially-selected” “long-term” stations confirmed a warming trend.
NIWA have made much of the fact that these stations were unadjusted, but so what? It’s cherry-picking. Apparently even the professionals do it.
I’ll add more to this series as I gather more material. Your comments and suggestions would be very welcome, as always. Thanks.
A-ma-zing ! Just when I thought nothing could shock me about NIWA’s shonky practice on temperature series, you expose this scandal. They must think we are all fools. But hang on. We are. It’s every kiwi picking up the tab for the ETS because of such blatant misrepresentafion of the facts. There’s going to be an accounting for this – I would not want to be in NIWA when the tough questions have to be answered and responsibility for such shonkiness attributed.
Yes, it’s surprising that they cut corners and fudge things when we’re bound to find out.
Now, Rodney, many people would look on you as being in a somewhat fortunate position. You’re part of the government of the country, you have the ear of the Prime Minister and you move in privileged circles, privy to information denied the rest of us. One might think that, if you had anything you wanted done, you are in a good position to have it done.
Yet you are not, for your efforts to delay or defeat the Emissions Trading Scheme are unsuccessful. John K seems adamant the ETS will be introduced, apparently convinced, relying no doubt on opinion polls, that we want one. It seems that you are just like the rest of us in this, that you cannot pull magic strings to get things done.
Is that true? Or do you have a secret scheme (please?) to delay the ETS without having to change public opinion? Or must we simply continue shouting at John Key that we don’t actually want the ETS? And fail if we cannot raise the required number of shouters?
Could you tell him to review this web site?
So six stations shifted – out of the eleven stations that were specially selected because they never shifted. If immobility was not the criterion, then why were these particular stations chosen for Jim Salinger’s rebuttal graph?
The accompanying spreadsheet shows that most of these stations weren’t even in the series when it began in 1931. It is, in fact, a four-station series involving only Mt Ruapehu, Palmerston North, Hamilton and Queenstown. Why were these four selected? They can hardly be considered as a fair representation of New Zealand as a whole.