I’ve moved to WordPress: http://bobtisdale.wordpress.com/

Saturday, May 31, 2008

Will Corrections to Global Temperature Make it Easier to Duplicate with Natural Influences?

Ever since Scafetta and West, in their recent paper “Is Climate Sensitive to Solar Variability?”, March 2008 “Physics Today”, provided the graph of Phenomenological Solar Signal (PSS) from 1950 to 2007, I’ve wanted to see what effect adding other natural climate influences (volcanic aerosols, ENSO, AMO, PDO) would have on the curve. Refer to Figure 1.

How closely would the result match the global temperature anomaly curve and trend? I needed to get the time to duplicate the PSS curve, which I’ve now completed. Figure 2 is the reproduced Phenomenological Solar Signal (PSS) curve from 1950 to 2007.

Figure 2

Scafetta and West paper:


Estimates of the effects on climate of explosive volcanic eruptions vary greatly. For the Mount Pinatubo eruption of 1992, these temperature estimates range from 0.2 to 0.5 deg C. I’ve elected to use 0.2, which is documented in “Short-term climatic impact of the 1991 volcanic eruption of Mt. Pinatubo and effects on atmospheric tracers”, Pitari and Mancini, Natural Hazards and Earth System Sciences (2002) 2: 91–108.

I’ve also used annual SATO Index of Mean Optical Thickness data, with a scaling factor of 1.65176, to adjust the curve based on the values for the Mount Pinatubo eruption. (0.2 deg C/0.121083 Mean Optical Thickness) Figure 3 shows the effect of volcanic aerosols on PSS.

Figure 3


Trenberth et al (2000) in “The Evolution of ENSO and Global Atmospheric Temperatures” identified two impacts of ENSO on global temperatures: the direct year-to-year effect and the impact on the linear trend.

They state: “It shows that for the 1997-98 El Niño, where N34 peaked at about 2.5 C, the global mean temperature was elevated as much 0.24 C (Fig. 2) although, averaged over the year centered on March 1998, the value drops to about 0.17 C.” And: “For 1950-98, ENSO linearly accounts for 0.06 C of global warming.”

I used NINO3.4 SST data available from NCDC, which I then converted to anomaly data based on the average temperature from 1950 to 1979. The annual NINO3.4 anomaly value centered on December 1997 (assumes a 3-month lag between NINO3.4 and global temperature) is 1.823 deg C. The annual temperature change of 0.17 deg C would provide a scaling factor for ENSO of 0.093. (0.17 deg C/1.823 deg C). I also shifted the NINO3.4 data 1 year, in an attempt to account for the time delay between ENSO and global temperature. The linear trend of 0.06 deg C was divided equally over the 49 year span. I then assumed it remained flat from 2000 to 2007.

With the NINO3.4 data added to the curve, Figure 4, the modified PSS curve begins to take shape, reflecting many of the annual variations in global temperature anomaly.

Figure 4


From Knight et al (2005) "A Signature of Persistent Natural Thermohaline Circulation Cycles in Observed Climate", GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L20708, doi:10.1029/2005GL024233, 2005: "The regression of simulated global and Northern Hemisphere mean decadal temperatures with the THC are 0.05 +/- 0.02 and 0.09 +/- 0.02_C Sv_1 respectively, implying potential peak-to-peak variability of 0.1 and 0.2_C." Peak to peak changes on the AMO are on the order of 0.45 deg C. With a global temperature change of 0.1 deg C and a 0.45 deg C swing in the AMO, the scaling factor would equal 0.22. (0.1 deg C/0.45 deg C)

Since the AMO oscillation splits the period being investigated, there is a time when it adds to global temperature and another when it subtracts. Refer to Figure 5.

Figure 5

From the mid-70s, when the AMO reached its minimum, to 2005, it added to global temperature. Between 1950 and 1975, its contribution was negative. And from 2005 to present, the AMO has been adding to the decline in global temperature, though minimally on an annual basis.


Figure 6 illustrates where the adjustments to the PSS curve stand now when compared to global temperature anomaly, as represented by HADCrut3GL. It’s not a bad fit. The obvious problem is in the trend line. The global temperature curve trend is significantly higher than the adjusted PSS curve. That appears to be caused for the most part by the variance prior to 1965.

Figure 6

In a “A Large Discontinuity in the Mid-Twentieth Century in Observed Global-Mean Surface Temperature” by Thompson et al, Nature 453, 646-649 (29 May 2008) doi:10.1038/ nature06982; Received 28 January 2008; Accepted 4 April 2008, the authors describe an instrument bias in SST that lowered global temperature as much as 0.3 deg C around 1945.

A preliminary “fix” by the Climate Research Unit of the University of East Anglia (CRU) is shown in Figure 7.

Is the pre-1965 variance in temperature curves illustrated in Figure 6 the result of that “instrument bias”, a.k.a. error, by the keeper of SSTs? Eye-balling Figures 6 & 7, the differences in both sets of curves appear to be close to the same magnitude. Or would another adjustment of the PSS curve be required? This leads us to…


The PDO is the only major natural climate phenomenon that hasn’t had its effect on global temperature specifically identified. I’ve been looking on and off for years. If a significant portion of the global ocean surface rises and falls naturally; by thermohaline circulation (THC), by meridional overturning circulation (MOC), by delayed response to ENSO signals, by some other unidentified source, or by a combination of all of the above; temperatures on land masses will change in response. The AMO causes Northern Hemisphere temperature changes, and ENSO causes temperature changes in both hemispheres. These responses have been identified. Why not the global or hemispheric response to the PDO?

The other mystery associated with the PDO is its actual magnitude. The data available from JISAO is standardized. By definition, a standardized value is the distance of one data point from the mean divided by the standard deviation of the distribution. Got that. For some climate data sets, the effect of standardization can multiply the data by a factor of four. For others such as ENSO, standardization might amplify the data by only 5 or 10%. Where does PDO data fall? Without downloading the entire global SST data set and extracting the Pacific Ocean data north of 20N, there’s no way for me to tell.

The annual PDO is illustrated in Figure 8. Based on the smoothed data, it would lower Northern Hemisphere and global temperatures between the mid-1940s and the late 1970s; magically the same period tropospheric aerosols are used by GCMs to lower global temperature.

Figure 8

So let’s add PDO data to the PSS curve and adjust the scaling factor until we get the slope of the PSS trend to match the trend of global temperature anomaly. The coefficient based solely on that requirement turns out to be 0.098, slightly higher than the coefficient for ENSO and less than half of the one used for the AMO. I’ve also shifted the PDO data one year, like ENSO. Figure 9 illustrates the change to the PSS curve and Figure 10 provides the comparison to global temperature anomaly. http://i26.tinypic.com/24mhwt2.jpg
Figure 9

Figure 10

Any future revisions to SSTs and global temperatures will also cause changes in the AMO, ENSO, and PDO, but the relationship between them will remain the same. It seems as though any attempt to raise the global temperature curve between 1945 and the mid-60s will simply require a smaller PDO scaling factor to make the curves fit.

UPDATE – JUNE 1, 2008


The following are duplicates of Figure 10, which compares Global Temperature Anomaly (HADCrut3GL) to the adjusted Phenomenological Solar Signal (PSS), but in these I replaced the Hadley Centre data with data from GISS (Figure 11) and NCDC (Figure 12). Due to the differences in SST data sets and in calculation methods, the Scaling Factor for the PDO needed to differ to match the trends. GISS required a much higher scaling factor (0.155), where NCDC require one significantly less (0.065).
Figure 11
Figure 12


I do understand that using undocumented PDO scaling to achieve the trend match corrupts the data. But is creating a non-existent tropospheric aerosol adjustment to tune GCMs any different?

I also understand the scaling factor used to calculate the ENSO contribution to annual global temperature may also contribute to changes in the PDO. This is possibly why adding PDO to the curve amplifies the ENSO signal. I have no means of removing the ENSO signal from PDO, to provide an independent contribution for it.

Hopefully, when all data sets are revised to correct the errors in mid-century SST data, the scaling of the PDO will not be required at present levels.


It occurred to me there might be interest in seeing all the adjustments that were made to the PSS curve illustrated on one graph. Refer to Figure 13. I’ve also included before and after linear trend lines.

Figure 13

UPDATE 2 – June 2, 2008

This morning at Prometheus, Roger Pielke Jr.’s Science Policy blog, in one of his battles with RealClimate, he posted a graph of projected changes to the global temperature anomaly data. Refer to Figure 14.

Figure 14

From the Pielke graph, I revised the global temperature anomaly data set and compared it to the adjusted PSS data. It’s important to note that in this comparison the PDO was not required to make up for any additional decline in the mid-20th century. It has been “zeroed.” Yet even without the PDO, the slope of the PSS trend exceeds the trend of the revised global temperature anomaly curve. Refer to Figure 15.

Figure 15

No comments:


Tips are now being accepted.

Comment Policy, SST Posts, and Notes

Comments that are political in nature or that have nothing to do with the post will be deleted.
The Smith and Reynolds SST Posts DOES NOT LIST ALL SST POSTS. I stopped using ERSST.v2 data for SST when NOAA deleted it from NOMADS early in 2009.

Please use the search feature in the upper left-hand corner of the page for posts on specific subjects.
NOTE: I’ve discovered that some of the links to older posts provide blank pages. While it’s possible to access that post by scrolling through the history, that’s time consuming. There’s a quick fix for the problem, so if you run into an absent post, please advise me. Thanks.
If you use the graphs, please cite or link to the address of the blog post or this website.