I’ve moved to WordPress. This post can now be found at Met Office Prediction: “Climate could warm to record levels in 2010”But They Fail To Tell You Why
The Met Office begins their December 10, 2009 press release…
…with, “A combination of man-made global warming and a moderate warming of the tropical Pacific Ocean, a phenomenon known as El Niño, means it is very likely that 2010 will be a warmer year globally than 2009.” They continue, “The latest forecast from our climate scientists, shows the global temperature is forecast to be almost 0.6 °C above the 1961–90 long-term average. This means that it is more likely than not that 2010 will be the warmest year in the instrumental record, beating the previous record year which was 1998.”
Most people would assume the “man-made global warming” portion of the opening sentence refers to anthropogenic greenhouse gases. But in this instance I believe “man-made” means the bias in the instrument temperature record, the Hadley Centre’s HADCRUT3GL data, the Hadley Centre created when they spliced two SST datasets together in 1998.
A year ago today, in the post The Step Change in HADSST Data After the 1997/98 El Nino, I plotted the differences between HADSST2 SST anomaly data and the SST anomalies of the three NOAA/NCDC SST datasets: ERSST.v2, ERSST.v3b, and OI.v2 datasets. Refer to Figures 1 through 3 for updated graphs.
The HADSST2 data acquired this man-made shift in Global SST data when they changed source SST datasets from ICOADS to NCEP-GTS in 1998. A coincidence? The supplier writes:
“The total period of record is currently 1784-May 2007 (Release 2.4), such that the observations and products are drawn from two separate archives (Project Status). ICOADS is supplemented by NCEP Real-time data (1991-date; limited products, NOT FULLY CONSISTENT WITH ICOADS).” [Emphasis added.]
Recently I also compared the global SST anomalies of HADSST2 and the other Hadley Centre’s SST dataset, HADISST, again plotting the difference between the two. Refer to Figure 4.
If the Global HADCRUT data is in record territory in 2010, it is in large part the product of poor source dataset splicing of the HADSST2 portion of the HADCRUT data. The period bias in their SST data varies from 0.06 to 0.12 deg C, depending on the dataset one uses for comparison.
The other portion of the record is the product of the multiyear aftereffects of the 1997/98 El Nino that I’ve discussed in numerous posts. The most recent overview of this is included in Global Temperatures This Decade Will Be The Warmest On Record….
ERSST.v2, ERSST.v3b, OI.v2, HADISST, and HADSST2 data is available through the LNMI Climate Explorer:
As though the El Nino discontinuity provided a convenient covering point for jacking up the ambient background. "He who controls the present, controls the past. He who controls the past, controls the future."
I find it interesting that all of your graphs here start in 1987.
Could this in any way be related to the fact that there is another discontinuity just before that time in your chart from September 24?
It seems at one point you use this discontinuity to being up this step increase and the to push your agenda that global temperature response to El Nino is not a linear one, but then use that a cherry picked discontinuity in the data set and conveniently ignore (on purpose?) earlier discontinuities just to continue your criticism the Hadley Centre's data?
Are you doing a little bit of cherry picking of the data time periods to skew it to your talking points at the moment? Rather convenient I must say and one that shows what your true agenda really is.
Dennis: You asked, “Could this in any way be related to the fact that there is another discontinuity just before that time in your chart from September 24?” and you linked a graph from the Thompson et al “High Tech Wiggle Matching” post.
Actually, no. I started the graphs in this post in 1987 to provide approximately the same period (11 years) before and after the 1998 shift.
Also, had you bothered to read the earlier post on this same subject from a year ago that I linked above…
… you’d note that I started the data in 1978 for that one (November 1981 for the OI.v2 data). There are some earlier differences but step changes in 1998 are still plainly visible.
In the graph you linked…
…from the Wiggle Matching post…
…the “Tdyn/ENSO/Volcano residual global mean” data is global surface temperature (land + ocean) that has had the impacts (as poorly calculated by Thompson et al) of the following variables removed: “Dynamic Fit”, which they define as “variations in the advection of marine air masses over the high latitude continents during winter,” ENSO, and volcanic eruptions. But in this current post, there have been no adjustments for “Tdyn/ENSO/Volcano”. This post only deals with the difference between Global SST datasets (SST anomalies of the HADSST2 Minus the SST anomalies of the other datasets). So your later accusations are unfounded. Your accusations are based on your misunderstandings of the differences between this post and the wiggle-matching post.
Satellite data shows a similar step change following the 97/98 El Nino.
So is your conclusion of the "non-linearity of global temperatures and El Nino" more a result of the 1998 discontinuity due to the spicing of the data records or an actual data jump (step function)?
It just seems very curious to me that the 0.12C-0.13C temperature jump in the 1998 time period are of similar magnitude in your comparative temperature plots (these current ones and the earlier ones you produced back in September and last December).
Could it be your non-linearity conclusions are the result of the same biased spliced data problem and not actual data correlations?
If the spliced data problems are removed or corrected for, then doesn't that affect your non-linearity issue?
I know you will probably tell me they are two different data sets that just happen to show the same sort of temperature jump at exactly the same time.
You seem to want to have it both ways here depending on what points you are trying to make.
Dennis: You continue to confuse the points of the two entirely difference exercises. I am now convinced that you aware of the differences and this is just your way of attempting to confuse those who read your comments here.
You asked, "So is your conclusion of the 'non-linearity of global temperatures and El Nino' more a result of the 1998 discontinuity due to the spicing of the data records or an actual data jump (step function)?"
El Nino-induced step changes in the East Indian and West Pacific SST anomaly data appear in ERSST.v2, ERSST.v3b, OI.v2, HADISST, and HADSST2 datasets. And as you noted, there are also El Nino-induced step changes in the TLT anomalies of the Mid-To-High Latitudes of the Northern Hemisphere.
The HADSST2 data as noted above contains these El Nino-induced step changes in the East Indian and West Pacific SST anomalies. In addition, the HADSST2 data (Globally, not only the East Indian and West Pacific data) also has an upward shift in 1998 that does not occur in the global HADISST, ERSST.v2, ERSST.v3b, or the OI.v2 data. This additional shift is caused by the splicing of two datasets in 1998. I've illustrated it in this post by subtracting the other datasets individually from the HADSST2 data.
Based on your own comments in your September 24, 2009 blog, you state the following regarding the Thompson data: "Thompson et al (2009) identifies the “Global mean” data as HadCRUT3, which is the Hadley Centre’s combined land surface temperature and SST data."
You'll have to forgive me and my supposed ignorance of your various SST datasets, but based on your description of the Thompson data, isn't that the same HADSST data that goes into this chart from your September 24 blog?:
Isn't that the same HADSST data that you are now pointing and using here regarding the possible spliced data issue that your are conjecturing the Hadley Centre may or may not use to make higher temperature conclusion next year?
So again, just so I can understand this all better, is your conclusion of the non-linear step increase based on an artifact of the same HADSST data splice or isn't it?
Dennis: You asked, “You'll have to forgive me and my supposed ignorance of your various SST datasets, but based on your description of the Thompson data, isn't that the same HADSST data that goes into this chart from your September 24 blog?:
Their “Global Mean” data is described as HADCRUT3, and yes, as HADCRUT3 data, it contains HADSST2 data, BUT as you will note in the header of the graph you linked, that data is the Thompson et al “Tdyn/ENSO/Volcano residual global mean” data, which means it has been modified. It has had datasets subtracted from the HADCRUT3 data, including an SST dataset, their modified Cold Tongue Index data. So it is no longer HADCRUT3 (CTRTEM3 + HADSST2) data.
You asked, “Isn't that the same HADSST data that you are now pointing and using here regarding the possible spliced data issue that your are conjecturing the Hadley Centre may or may not use to make higher temperature conclusion next year?”
No. It has been modified by Thompson et al with the subtraction of their “ENSO Fit” and “Volcano Fit” data. (The “Tdyn Fit” only applies to the CRUTEM portion of the data.)
You asked, “So again, just so I can understand this all better, is your conclusion of the non-linear step increase based on an artifact of the same HADSST data splice or isn't it?”
The step increases in this post, as calculated by subtracting the other global SST anomaly datasets from the global HADSST2 anomaly data, are a function of the data splice. It is likely that a portion of the step in the graph you’ve linked from the “Wiggle Matching” post is from this splice problem, too, but there are many differences between the graphs in this post and the “Wiggle Matching” graph you are so fond of linking. If you’d like to go through the same process that Thompson et al went through but limit the data to SSTs and perform that process for all of the SST datasets listed in this post, you may be able to figure out what portion of the step in the “Wiggle Matching” graph is the result of the data splice. You’d also have to use period average SST instead of trends in your modified “Wiggle Matching” graphs so that you’re comparing apples to apples with the graphs in this post.
Post a Comment