Notes On The GISTEMP Ratio Of Land To Sea Surface Temperature Data
Guest post by BobTisdale
Over the past few days there has been some blogosphere buzz about the apparent ratio of Land Surface Temperature (LST) Data used in the GISTEMP combined land and sea surface temperature data with 1200km smoothing. I’ll provide a few comparison graphs to explain.
Figure 1 includes a time series graph of the Hadley Centre’s HADCRUT3 Combined Surface Temperature product, from January 1982 to April 2010. Also included is the weighted average of the two datasets that make up the HADCRUT3 data, with the weighting of 27% LST [CRUTEM3] and 73% Sea Surface Temperature (SST) [HADSST2]. Those weightings were required to match the linear trend of the weighted average to the linear trend of the HADCRUT3 combined product. The weighting makes sense, since the global oceans represent about 70% of the surface area of the globe.
http://i28.tinypic.com/2yoe59f.jpg
Figure 1
Figures 2 and 3 provide similar comparison graphs. Figure 2 shows the NCDC combined surface temperature product and the weighted average of its LST and SST components. To align the linear trends, the weighting required for the components of the NCDC product was also 27% LST data and 73% SST data. Figure 3 shows the GISTEMP product with 250km radius smoothing. For this GISTEMP product, the weighting required for the components was 28.5% LST data and 71.5% SST data. Again, the relationships of the SST and LST data make sense.
http://i26.tinypic.com/2qmi3yh.jpg
Figure 2
##########
http://i25.tinypic.com/1rx6v7.jpg
Figure 3
Here’s the curiosity. It appears in the GISTEMP Product with 1200km radius smoothing when we apply the same component weighting (28.5% LST data and 71.5% SST data) that we had used on the other GISTEMP combined product. The weighted average of the components of the GISTEMP combined product with 1200km radius smoothing has a significantly lower trend than the actual GISTEMP data. This can be seen in Figure 4.
http://i28.tinypic.com/9jot4x.jpg
Figure 4
In order to achieve the same linear trend as the GISTEMP combined product with 1200km radius smoothing, the components have to be weighted with 67% LST data and 33% SST data, almost reversing the ratio of the areas of global oceans and continental land masses.
http://i30.tinypic.com/p9g5d.jpg
Figure 5
Figure 6 is a map illustrating the GISTEMP LST data (trends) from 1982 to 2009. Note how the GISTEMP LST data extends out over the oceans. This is not the case for their combined product, because GISS masks the LST data over the oceans in its combined product. So in order to properly create a weighted average of GISTEMP land and sea surface temperature data with 1200km radius smoothing, the land surface data where it extends out over the oceans would first need to be masked.
http://i26.tinypic.com/4ieop2.jpg
Figure 6
A NOTE ON THE DIVERGENCE BETWEEN GISS AND THE OTHER DATASETS
Much of the divergence between GISTEMP and the Hadley Centre and NCDC combined surface temperature products is likely caused by the fact that GISS deletes SST data in the Southern and Arctic Oceans and replaces it with LST data, which has a significantly higher linear trend than the SST data it replaces. This was discussed in the post GISS Deletes Arctic And Southern Ocean Sea Surface Temperature Data.
ANOTHER CURIOSITY
It can also appear that GISS extends LST data out over the oceans in areas other than those with seasonal sea ice. In fact, I made this mistake in a comment at Lucia’s The Blackboard this morning. Refer to my Comment#49191 at the bottom of her post NOAA: Hottest June in Record. This illusion can be seen in the following .gif animation of GISTEMP trend maps for the period of 1982 to 2009. The April trend is presented in Figure 7. Note how, in the highlighted area of the North Atlantic, there are differences between the SST trend and the trend of the GISTEMP combined product with 1200km radius smoothing. The Faroe Islands are located between Scotland and Iceland, and GISS uses station data there, so that explains the differences in that area. But what of the area of the North Atlantic west of Ireland and south of Iceland, with the approximate coordinates of 50N-60N, 20W-15W? There aren’t any islands there with weather stations.
http://i29.tinypic.com/2i0vhif.jpg
Figure 7
In its GISTEMP LST products, GISS also includes surface station data identified as Ship followed by a letter; that is, “Ship J”, “Ship R”, etc. Refer to Figure 8. These can be found using the station locator feature on the GISTEMP Station Data webpage.
http://i32.tinypic.com/2i09k7r.jpg
Figure 8
Here’s a link to the webpage presented in Figure 8.
I have found little to no information on these GHCN “ship stations”. Are they presenting SST or Nighttime Marine Air Temperature? Dunno. They may have served a purpose when GISS first prepared their GISTEMP product due to the sparseness of SST data in the early SST datasets, but now, these “ship stations” only add an unknown bias to well-documented optimally interpolated SST data. (And if they don’t add a bias either way, then there’s really no reason to have them. All they do is add confusion.)
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Frank Lansner: You wrote to Steven Mosher, “In the present case for example we have a GISS LST+SST that is the same as the SST (HADISST) 1900-1920.”
What’s the source of your data? A quick trip to the KNMI Climate Explorer shows that to be incorrect. There are significant differences between Global GISTEMP LST + SST (1200km radius smoothing) and Global HADISST from 1900 to 1920:
http://i25.tinypic.com/2mi14bm.jpg
And here’s the difference (GISTEMP MINUS HADISST):
http://i25.tinypic.com/2rwrbpg.jpg
Steve Keohane says:
July 18, 2010 at 8:34 am (Edit)
Thanks Bob, that looks like a lot of work. The bottom line wrt temperature measurement seems to be that lacking enough data points, the numbers can be fiddled with to look like anything. ”
******************************
actually quite the OPPOSITE. Its very hard to make the trend in the data disappear.
As a simple experiment ( RonB and others) have taken a simple
straight average of all the stations. That is, average the stations with NO area weighting whatsoever.
http://rhinohide.wordpress.com/2010/07/10/trb-0-01-ghcn-simple-mean-average/#more-832
There isnt much you can do to the numbers to get the basic facts to disappear.
davidmhoffer says:
“Now I don’t know what the right ratio is, I think Frank Lansner suggested 30% land 70% ocean. For trending purposes, I just don’t see how you can change the weighting between land and ocean of the combined anomaly based on the number of places you have weather station data for. Suppose the number of land cells doubled, but their new ones had the exact same average anomaly as the original ones.”
I’m not sure you understand how spatial averaging works or the temporal aspect.
Lets start with the spatial.
Box A: 5 degrees square: it has 50 stations.
Box B. 5 degrees square. it has one station.
The first step is to average the stations in box A. that gives you 1 time series for that
Box. Box B has 1 station. That station represents that box.
Assume an area of 25 for each. So Box A gets 25/50 (50%) weight and box B gets
50%. Now if Box B is 50% land, you have to scale for that
Box A would get 25/37.5 weight and box B would get 12.5/37.5 weight.
Adding stations to A will not change it much. Adding stations to B can change B..
but the leverage is not that great.
New stations also have to have a temporal overlap ( in RSM or CAM)
The best approach ( from a stats perspective) is the LSM approach (Nick Stokes,
RomanM) but they are not substantially different than other approaches.. they do use more data, but MORE DATA does NOT shift the curves. That’s because the field is already oversampled in certain areas,despite what people think about the sampling requirements. Take stations away or add them, the warming (TREND) does not change appreciably.
davidmhoffer: You replied, “It isn’t the specific data that is added it is the manner in which it is handled in the analysis that matters.”
Are you suggesting that GISS should standardize on only those Land and Sea Surface Temperature grids that are available over the entire 1880 to 2010 period? If not, how should they introduce new data as it becomes available or drop it when it is no longer available?
For GISTemp, either the Ocean SST number is wrong, the Land Temp is wrong or the Global Temp is wrong.
It is just basic math.
I would normally say “just fix the one that is wrong” but that would likely involve an increase in all of them.
Steve, do you think that if someone creates a “method” i cannot be .. well … rotten? A method cannot be developed with “so-so” intensions?
Frank:
“Steve, in CRU program code released in climate gate its a permanent method that temperature proxy graphs are cut of in 1960, and the same can be read in their mails. So they have “a method” to cut of data most often in 1960, in then the sceptics cannot complain?”
As I explained in my book on climategate there is nothing WRONG with truncating a proxy series when it fails to correlate. Given the lack of correlation the analyst has several choices:
1. Reject the underlying science that tree rings work to capture temperature.
2. Include the series and live with the uncertainty this creates.
3. Drop the WHOLE series
4. Truncate.
The issue is full disclosure of the method and the Sensitivity to the analyst choice.
each of the decisions above is a rational defensible choice. Each choice has an impact. In my experience I would aways back such choices up by doing the sensitivity analysis. So yes you can complain, but you ALSO have to acknowledge that some choices make no difference.
“In the present case for example we have a GISS LST+SST that is the same as the SST (HADISST) 1900-1920. The GISS LST is higher. So the weight of land data is zero. Obviously this is build into a method , a program etc. and then its suddenly ok??”
I wouldnt characterize the method as OK.
Every method has limitations. They are all approximations. All “wrong.” Giss has a method. Its weak points Are well know. CRU also has a method. Its weak points are known. Some of those weak points are addressed by Nick and Roman M. Some addressed by Zeke and me. The bottom line is the weaknesses do not make the warming disappear.
Here is another method to ESTIMATE the average temp of the world.
The highest temp ever recorded was 136F. The coldest -128F
Can you estimate the average temp of the globe today with those two facts?
Sure. the average would be 4F. Not a very good estimate. but given those two facts
its the best estimate you have. Thats always the question. whats the best estimate given the facts. every method of estimating has limitations. hence the word estimate.
Now lets look at the place that is consistently the warmest.. Dahli ethiopia about 94F and coldest antartctica.. about -70F.. that gives us an average of about 12F. better estimate it uses average temps and not extremes
Now Add england to that Average: CET is around 50F over time..
Now your average is about 25F
Now lets add a long american record orland CA.. about 62F and our average is
up above 30F.. keep adding stations and you will get to the numbers that Ron shows.
Thats one method. This method, however, is limited becuase we didnt ask how close the stations were to each other. that brings in the issue of area averaging
“Are you suggesting that GISS should standardize on only those Land and Sea Surface Temperature grids that are available over the entire 1880 to 2010 period? If not, how should they introduce new data as it becomes available or drop it when it is no longer available?”
one advantage of GISS method ( RSM) is you can add stations over time. With CAM thats less likely
Steven Mosher: And what do the acronyms CAM and RSM stand for?
Bob Tisdale says:
July 18, 2010 at 3:35 pm
davidmhoffer: You replied, “It isn’t the specific data that is added it is the manner in which it is handled in the analysis that matters.”
Are you suggesting that GISS should standardize on only those Land and Sea Surface Temperature grids that are available over the entire 1880 to 2010 period?>>
No, I’m not saying that. I’m saying that the current method can result in a slightly amplified trend based on the method used now. If you can increase the trend by adding more data points with the exact same anomaly as the ones you already have, then you have a problem. I think dispensing with ocean data reported from within a certain distance of a land station is a similar issue. As more land stations in proximity to ocean come on line, less ocean data is used. These things may be small, but we’re arguing about tenths of a degree per century and they do add up. Does it change the over all trend? No. We’re talking perhaps 0.6 degrees per century instead of 0.8 or in that neighbourhood. But if governments are going to take action at potentialy massive financial and social costs, do they (and we the public) not deserve that decisions be made on the most accurate analysis? Suppose, just to illustrate, that our globe had 50 ocean points and 50 land points. On day one all the ocean points have anomalies of 1 and there’s only one of the 50 land points that has data and it has an anomaly of 2. So someone averages out the 51 data points and gets an over all anomaly of 1.0196. The next morning the ocean anomalies are all 1 again, but there is data for all 50 land points and they are all 2, just like the single one from yesterday. So now we have 50 at 1 and 50 at 2. Despite the extra 49 land points being exactly the same as the first one, I’m supposed to be OK with the anomaly now being 1.5? I think not.
Continuing to present data that contains a known mathematical construct or two that artificially enhance the trend is not the best possible data, the fact that this problem has been known for two years and it hasn’t been addressed disturbs me, and if nothing else it leaves wide open the question of what other oddities not yet discovered are hiding in that data simply from poor treatment from a mathematical or other fundamental perspective.
As I said before, I don’t know the right answer, just that this approach is wrong. As Steve Mosher suggested, there may be no “right” approach possible, just wrong ones and less wrong ones. Mosher suggested a possible approach as did Lansner.
Interestingly, I once graphed every 5th year of land anomalies using only those grid points with continuous data in each of those years starting in 1880. I think there were around 800 or so. The match to GISS Land was surprisingly good. I may still have that graph around somewhere.
I’m an idiot, of course I have it.
http://knowledgedrift.files.wordpress.com/2010/06/pic41.png
Every 5th year and downloaded from Global Maps so not to be relied on in any way, but I do plan to do it again using monthly data on the same grid points or something similar. I may even do it with 1200 km smoothing to nobody can say I’m not using the “official” data. 🙂
latitude says:
July 18, 2010 at 5:22 am
tarpon says:
July 17, 2010 at 11:06 pm
Has anybody found the warm CO2 greenhouse roofing placed over the equator by all the computer models? Isn’t the missing CO2 greenhouse roofing proof enough the whole theory is composed of falsehoods? In normal times it would disprove.
==========================================================
You would think, since that is a requirement for verifying the computer programs.
Without it, the computer programs are wrong
Last I heard, they said warming is there, because the computer programs say it is, they just can’t find it
Last I heard, since actual temperature measurements didn’t agree with the models, they had to find a proxy for actual temperature: Winds.
All the graphs look as if they climb till sometime in 1997 and then hold steady or dip from 1997 till now. One regression line doesn’t always cut it.
Is it just me or is there a pattern here of Mosher and Zeke defending some pretty ridiculous work by hansen?
Lucy Skywalker
Could you send me John Daly data? The person managing the site now is obviously overwhelmed.
davidmhoffer: You wrote as an example. “Suppose, just to illustrate, that our globe had 50 ocean points and 50 land points. On day one all the ocean points have anomalies of 1 and there’s only one of the 50 land points that has data and it has an anomaly of 2. So someone averages out the 51 data points and gets an over all anomaly of 1.0196. The next morning the ocean anomalies are all 1 again, but there is data for all 50 land points and they are all 2, just like the single one from yesterday. So now we have 50 at 1 and 50 at 2. Despite the extra 49 land points being exactly the same as the first one, I’m supposed to be OK with the anomaly now being 1.5? I think not.”
It’s wrong for me to speculate and hopefully Steven Mosher will be back to explain, but I believe you’re missing something. In your example, aren’t you missing the point that the ocean and land points are treated independently, meaning that in your first example with only one land point that the anomaly would 1.5, not 1.0196? Refer to Steps 3 and 4 of the GISS current analysis discussion:
http://data.giss.nasa.gov/gistemp/sources/gistemp.html
The coastal cells are another matter.
You wrote, “As more land stations in proximity to ocean come on line, less ocean data is used.”
I assume this concern is based on the GISS Step 5 from…
http://data.giss.nasa.gov/gistemp/sources/gistemp.html
…in which they write, “The same method as in step3 is used, except that for a particular grid box the anomaly or trend is computed twice, first based on surface data, then based on ocean data. Depending on the location of the grid box, one or the other is used with priority given to the surface data, if available.”
But the opposite would be true when fewer land stations are used and isn’t that what has happened over the past decade globally?
(An observation)
Ahhhh Yes!
Peer (and Equal, and Colleague, and Contemporary, and Cohort, and Friend, and Not) Review at it’s finest. Limiting ‘Old World Print Media’ Peer Review to 2 or 3 select individuals is such a long and shallow exercise. Here at the Forum we are all truly naked and wrestling for the truth in the finest traditions of the Great Greek Masters of Science AND Philosophy. What a World we live in.
This is a perfect example of why WUWT is so popular.
This should help solve any remaining confusion on the land ratio issue: http://rankexploits.com/musings/2010/the-gistemp-land-fraction/
Steven Mosher:
You write: “As I explained in my book on climategate there is nothing WRONG with truncating a proxy series when it fails to correlate. ”
I disagree!
I have recently collected the most comprehensive overview of NH temperatures, mostly land.
http://hidethedecline.eu/pages/posts/part1-the-perplexing-temperature-data-published-1974-84-and-recent-temperature-data-181.php
Most of the proxies that where cut in 1960, but IPCC and others used tree ring data from (land !!) and from NH.
The temperature decline in NH land after 1960 was STRONG. So any idea of cutting tree ring data off in 1960 because it decline is extremely poor judgement, and I am surpriced how you can defend this??
See rough illustration:
http://hidethedecline.eu/media/PERPLEX/fig4.jpg
How about we discussed this in a seperate writing?
I have earlier gone through some “arguments” explaining why trees shows decline “when they shouldn” and its a poor reading to say the least.
K.R. Frank
And while all the attention is being focused on surface temperature measurements, who is going to remind us of the decreasing upper atmospheric temperatures measured by satellites, and the increased COOLING of the thermosphere by CO2? NASA GISS? NOAA? Anyone?
Steven Mosher says:
July 18, 2010 at 4:06 pm
As I explained in my book on climategate there is nothing WRONG with truncating a proxy series when it fails to correlate. Given the lack of correlation the analyst has several choices:
1. Reject the underlying science that tree rings work to capture temperature.
2. Include the series and live with the uncertainty this creates.
3. Drop the WHOLE series
4. Truncate.
The correct conclusion is 1. in my opinion.
2 I could live with, if the errors take it into account
Climate scientists are cavalier about errors , and par excellence, propagation of errors.
To make a thermometer of some proxy, one has to prove very good correlation with temperatures measured with another proxy, i.e. classical thermometers.
i.e. I would need a paper with 1000 tree measurements per point ( three percent error) in time that showed good correlation up to the present temperatures. Then I could start thinking of what might have happened for this particular series to show a decline and whether truncation would be rigorous science.
But if I had 1000 tree measurements, what would I need this single one for?
As I said, climate science is not rigorous science in any sense, but there is a limit to how much video gaming can be done with the data and can be accepted/swallowed by the rest of the scientific community.
As with the particular navel gazing over the GISS temperature anomalies. So yes, the temperatures have been rising since the little ice age. It is as trivial as saying winters are colder than summers. But what do anomalies really have to say about the price of tea in China? i.e about energy inputted outputted from the planet?
The number of convolutions that lead to anomalies is too large to be able to use them as a gauge for real energy numbers. It is as if one is given a map without a scale, so one cannot tell if it is a country one is looking at or a county. It is a small r that is missing but very crucial for driving. It is even worse, because it it a distorted map without rigorous notion of the proportions .
Even if one used average temperatures, that the GCMs model so very badly, there would be an almost inpossible task of getting real energy numbers: ground skin surface temperatures, gray constants and radiation spectra are needed over the whole map, in addition to a large number of samples to satisfy the Nyquist criterion of statistics.
It is ground surface skin temperatures ( including SSTs) that are doing the bulk of radiation, and here one is navel gazing over the air temperatures at 2 meters as if they are controlling bulk radiation.
This is an interesting plot that shows the huge variations in sea surface temperature with the time of day http://www.ghrsst.org/images/rubbish.jpg from
http://www.ghrsst.org/SST-Definitions.html .
Remember , radiation goes like T^4 . The average day night temperature does not give the average radiated energy.
Simple it ain’t even for sea. Imagine what happens for land.
Thanks, Anthony.