2010-09-08 13:08:41Blog post on South American hockey stick
John Cook

john@skepticalscience...
121.222.93.62

One thing everyone is welcome to use the Authors Forum for is test driving new blog posts, regardless of whether they're used as rebuttals. It's a good way to get feedback and ideas before going live. I've just thrown together a blog post based on a new paper I've just had emailed to me and welcome any thoughts (struggling for a take-home point here). Let me know if you'd like me to email you the paper.


A South American hockey stick

A new paper has just been published employing a new technique for reconstructing past temperatures (Kellerhals 2010). It uses Ammonium concentration from an ice core in tropical South America (the eastern Bolivian Andes) as a proxy for temperature. This enables them to build a temperature record going back 1600 years in a region which has had little proxy data available until now. The results are interesting. They find a distinguishable Medieval Warm Period and Little Ice Age in the record. However, the last few decades show unprecedented warmth over the last 1600 years.


Figure 1: Reconstructed tropical South American temperature anomalies (normalized to the 1961–1990AD average) for the last 1600 years (red curve, smoothed with a 39‐year Gaussian filter). The shaded region envelops the ±2 standard deviation uncertainty as derived from the validation period. Poor core quality precluded any chemical analysis for the time interval between 1580 and 1640 AD.

Of course, this is a temperature record for a single location. There are proxy records that find some locations were actually warmer than now during the Medieval Warm Period. One can be misled by looking at these isolated records as evidence that the Medieval Warm Period was overall warmer than current conditions. However, when we combine all the various temperature records, we find modern temperatures are significantly warmer than medieval temperatures. This is demonstrated in Moburg's reconstruction of Northern Hemisphere temperature (which happens to bear a striking resemblance to the South American proxy record).


Figure 2: Northern Hemisphere Temperature Reconstruction by Moburg et al. (2005) shown in blue, Instrumental Temperatures from NASA shown in Red.

 


Note: all I can find by way of temperature reconstructions are Northern Hemisphere ones. Eg - this composite from Mann 2008:

 

Are there any global time series going back 1000 years?

2010-09-08 16:54:55in Figure 1
nealjking

nealjking@gmail...
84.151.34.204
Why is there a break in the record around 1600 AD ?
2010-09-09 00:34:54
Ned

ned.flounders@yahoo...
129.170.23.6

That looks good, John.  Are the instrumental temperatures at the right side of Fig 2 from the global land/ocean data set?  If so, just be aware that you're comparing a "northern hemisphere" reconstruction with "global" instrumental temperatures. That's OK, but be aware that people might ask about it.

There is definitely less proxy data available for the southern hemisphere.  

You probably already know about this, but you can find a list of reconstructions (global, regional, etc.) here:

http://www.ncdc.noaa.gov/paleo/recons.html

You might be better off using Moberg in place of Mann ones, even if Moberg is NH only, just to avoid the inevitable tantrums that any link to Mann et al will probably produce.

It's too bad that Ljungqvist 2009 just compiled lots of proxies and didn't offer an actual global reconstruction.

2010-09-09 07:44:17Comment
Robert Way

robert_way19@hotmail...
142.162.205.242
I believe on figure 2 we are looking at Northern Hemisphere temperatures. I'm not 100% sure and I can go look over the data again but i think I used northern hemisphere referenced to the 1961-1990 just like moberg. I'll confirm this later tonight.

Yeah that is the point of using Moberg in place of Mann as it just tries to supercede all the mann haters and tree-ring stuff. It really is too bad about Ljungvist 2009...
2010-09-09 14:33:29Comment
Robert Way

robert_way19@hotmail...
142.162.205.242
I checked the data and I can't find the actual temperature data I used. It seems highly likely that you are right about me using the land only index anomalized to 1961-1990 from the NOAA state of the climate data. That's what I think I did. Either way, here is the revised figure...
http://www.skepticalscience.com/pics/Mobergetal.2005.jpg

It is using hadley NH data because I couldn't find the temperature reference period for GISS northern hemisphere data... so i had to use hadley which was the same baseline 1961-1990 as the moberg reconstruction...



It had to be some sort of land/ocean or anomalized period splice based upon how little warming is in the other graph. Can't believe we didnt catch that sooner... sorry guys...
2010-09-09 16:38:06NH Temperature
John Cook

john@skepticalscience...
121.222.93.62
Robert, don't suppose you could email me the spreadsheet file of the latest graph? I'd like to use it in the blog post.
2010-09-09 17:38:12
nealjking

nealjking@gmail...
91.33.118.20

John,

I still have my question about Figure 1: Why does it show a gap in the plot around the year 1600 AD?

2010-09-09 23:04:06Poor core quality
John Cook

john@skepticalscience...
121.222.93.62

The reason given by the authors:

Poor core quality precluded any chemical analysis for the time interval between 1580 and 1640 AD.

I'll include this in the caption as I'm sure others will wonder.

2010-09-09 23:30:46OK
nealjking

nealjking@gmail...
91.33.118.20
Yes, I think so: ""Inquiring minds want to know..."
2010-09-10 01:33:05Comment
Robert Way

robert_way19@hotmail...
142.162.205.242
John,
I will email you the spreadsheet. The unfortunate thing is I have been a tad messy in my rush to put together the revised version so i'll clean it up and send it along.

Sorry if this initiated any problems. I apologize for my oversight. It seems in my hurry to find data with a base period of 1961-1990 I forgot to make sure I was talking about NH versus Global. When we talk NH it becomes even more unequivocal..
2010-09-10 10:58:26Thanks for Moberg spreadsheet
John Cook

john@skepticalscience...
121.222.93.62

Thanks heaps. Just made some minor formatting tweaks and extended it a little past 2000 as arguably the most important part of the time series was getting cut off - the hottest recent decade:

 

2010-09-10 13:47:35Comment
Robert Way

robert_way19@hotmail...
142.162.205.242
Certainly puts things into context doesn't it... And to think some people argued Moberg showed too much variability...
2010-09-19 17:38:04I recommend
nealjking

nealjking@gmail...
84.151.40.102
that everything be ready for "inspection", as the "Hockey Stick" "anti-Team" will be likely to come out in force, looking to beat the argument up over anything.