“Mr Lynn (16:54:38) :
Now a paper from 1980 throws all the New Zealand data out:”

http://www.investigatemagazine.com/australia/latestissue.pdf

Includes scanned copy of a 1980s journal article.

http://www.investigatemagazine.com/hessell1980.pdf

Wow, Salinger, Trenberth, Lawrence, these guys go way back.

If the series is extended a hundred years to 2020, all with 11 degrees, the linear trend STILL increases by 0.2 degrees.

Proof of global warming for hundreds of years.

“RE: your 20:00:54 comment… Not sure what you mean. What data did u use? Annual averages or monthly? I tried it with annual and did see a bit of a trend.”

If you meant the part about a trend being no more than 0.2C or flat,

The first 14 years (Thorndon) adjusted down 0.6C averages 12.6C.
The last 14 years (Kelburn) adjusted down 0.2C averages 12.8C.
So there is an average increase from those periods of 0.2C.

The linear trends for Thorndon first 14 years and Kelburn last 14 years are both descending. This, if you trust “trends”, indicates that Thorndon was cooling and Kelburn is cooling, which puts any real increasing trend in doubt, which is why I said Wellington is flat.

Linear trends can be deceptive, and can be more so for a long period than a short. If all data is combined and plotted from 1913 to 2008, a linear trend = 0.6C.
But a little test shows the problem with linear trends such as this.

Take a 20 year sample of temps shown below, the linear trend increases 0.2C, even though temps end up cooler than the beginning:

year temp

2000 12
2001 12
2002 12
2003 12
2004 12
2005 8
2006 8
2007 8
2008 9
2009 9
2010 10
2011 10
2012 11
2013 11
2014 12
2015 12
2016 11
2017 11
2018 11
2019 11
2020 11

“RE: your 20:00:54 comment… Not sure what you mean. What data did u use? Annual averages or monthly? I tried it with annual and did see a bit of a trend.”

I used annual from monthly averages. What trend do you mean? For the ’62 to ’08 stations difference? If so, I wouldn’t expect a trend, but a pattern indicating weather shifts. Oh, I made a mistake, it looks like the difference is greater, not less, when there are dramatic ups and downs in temperature. Don’t have a clue as to what could be the reason behind that.

“That would not provide you with overlap. Lapse rates change with the weather.
You’d be calibrating Thorndon with 2008 weather. ”
Better than nothing, not as good as having had the overlap in the first place.”

Nothing is not better than nothing. Might as well read tea leaves. You can do that now, though, just as I did with the airport and Kelburn, and lower Thorndon by .9C, or use the Karori station and adjust Thorndon by 75.1C. At least a long term record of two stations like I did is something, as weather comes and goes but a long average has a better chance of hitting on a more accurate average.

“On the met models, I didn’t mean the GCM’s used in modern forecasting. I meant simple stuff to take into account local conditions, based on climatological data like wind patters, cloud patterns, etc.”

It’s far from being simple, and it changes. I wouldn’t want to be the one that spent years developing algorithms and programs just to find out after firing it up that it didn’t predict temperatures within a degree in a week.

I suspect your surmise at the end of the previous comment is closest to the truth.

The adjustment of exactly the standard lapse rate makes me the most suspicious.

………

“That would not provide you with overlap. Lapse rates change with the weather.

You’d be calibrating Thorndon with 2008 weather. ”

Better than nothing, not as good as having had the overlap in the first place.

On the met models, I didn’t mean the GCM’s used in modern forecasting. I meant simple stuff to take into account local conditions, based on climatological data like wind patters, cloud patterns, etc.

I suspect your surmise at the end of the previous comment is closest to the truth.

I

“I’m probably not going to have time to do much more fiddling.”

Yes you do. Come on, don’t quit now.

“One thing to do would be to put stations in historical locations (like Thorndon) no longer present and record a few years of calibration data. That would help anywhere you didn’t have enough overlap.”

Time travel hasn’t been developed to the degree necessary. (Sorry, couldn’t resist)
That would not provide you with overlap. Lapse rates change with the weather. You’d be calibrating Thorndon with 2008 weather.

“Another would be to try to measure average lapse rate for similar station differences in the same general area, and crank that in.”

Yes, but we don’t have the average lapse rate for Thorndon, nor any stations that overlap.

“nother would be to try to create a meteorological model of each station, and use it to “forecast” the past in some sense.”

Perhaps, but the old records are scarce enough, there is not much reason to expect that in the late 1800s and early 1900s detailed enough records would have been kept on daily basis of such variables you mention. And even now models have a hard time forecasting precise temperatures to tenths of a degree. It’s one reason we use thermometers.

“NIWA obviously didn’t try very many of these. If I were them, I’d use the adverse publicity to get funding to do more of this sort of thing, while at the same time making everything transparent.”

That isn’t obvious. The guy who started this survey has published work. NIWA claims to have a pamphlet in their library of the methodology used as well. I’ve read a little, about studying old sea charts and wind patterns, pictures, tea leaves and such.
But when it comes right down to it, a massive amount of data and analysis could have been attempted; how much was actually considered is a different matter, and whether there is information anywhere that could support a serious scientific study may be elusive or nonexistent. It could have been not much more than “ok, do the lapse rate, no stinkin uhi, I got to hurry off, busy with a bunch of different things, thats enough to get through review, here ya go” kind of thing. I have very little confidence that isn’t closer to the truth.

Kelburn data is available from 1962 through 2008 (complemented by Kelburn AWS for the last four years), and Wellington Aero (Airport) data exists from 1962 through 2008.

The average temperature difference between these locations from 1962 through 2008 is 0.6C. NIWA adjusted Thorndon the same amount as the “Airport” location, by .8C.

There is no trend in the stations differences between 1962 and 2008, although recently, from 1995 through 2008 the difference has been .9C. As an aside, from 1995 through 2008, Kelburn has a .2C decreasing linear trend, while Aero has a .2C increasing trend. Both trend downwards from 2000 through 2008.

Analyzing a comparison chart of the two station records, the long term average can be clearly seen to be affected the most by times of steep increases or decreases in temperature, which brings the stations temps closer together and affects the long term average differential.

A graph of the Thorndon temps is a good example of steep increases and decreases. If differences between stations are an indication of lapse rate, using the best estimate of a real, long term local average lapse rate would necessitate adjusting Thorndon down by .6C, not .8C.

It is entirely possible and IMO likely that the Wellington area has not experienced an overall increasing trend of more than a couple tenths of a degree in a century. Time of measurement errors likely prevail throughout earlier records, and as more attention was paid to climate change in the mid 70s, more effort and more records were likely done and kept which could easily cause an increasing divergence between the old and newer records. Local land use changes could have an effect, and could station UHI, for example with the two airport stations that combine in the NIWA record, the station that NIWA combines with “AERO” being an average of .2C higher than the other operating station which is not used by NIWA in their graph.
If all this is taken into consideration along with an assumption that Kelburn itself after 2000 read a couple tenths of a degree higher due to siting or UHI problems, Wellington trend is flat between 1913 – 2009.

I’ve done some and from what I can tell it does work.

The actual pieces are going to be relatively long, several years in most cases.

I’d have to look at the math a bit on that.

However, you only need to stitch them together if you need to know what the absolute temperature record has been in the past.

If you ask a different question, what’s been the trend in temperatures, I don’t think you need to stitch them together.

Since AGW is the claim, the world has warmed, knowing the trend us sufficient.

Nick