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

Sunday, January 25, 2009

Reproducing Global Temperature Anomalies With Natural Forcings

I’ve moved to WordPress.  This post can now be found at Reproducing Global Temperature Anomalies With Natural Forcings
It Only Takes NINO3.4 SST Anomaly, Sunspot Number, and Volcanic Aerosols Data and A Different Mindset


In this post, only natural forcings are used to simulate the Global Temperature Anomaly curve. The correlation is closer than any other attempt of this kind that I’ve seen to date, with or without anthropogenic forcings and with or without General Circulation Models (GCMs). Consider two things: First, most GCMs that government entities employ to make projections about future global climate do NOT model El Nino-Southern Oscillation (ENSO) events, even though ENSO is a dominant natural climate phenomenon. Second, those few GCMs that attempt to model ENSO events do NOT model their processes or the climate responses before, during, or after the events with any accuracy, though they are improving.

In the following, I provide a recipe so that anyone familiar with a spreadsheet and who’s capable of downloading data from the links could reproduce the results with little effort.

It’s also important to note that I am not attempting to disprove the hypothesis of Anthropogenic Global Warming (though I am skeptical of the proejections). However, I am very clearly illustrating that more effort needs to be expended to isolate natural and anthropogenic causes for the warming over the past 130+ years, especially the long-term effects of ENSO. This should be done before attempting to project the anthropogenic causes of warming into the future by 100 years or more. GCM simulations of ENSO must be improved to accomplish that.


In a series of posts titled “Can El Nino Events Explain All of the Global Warming Since 1976?”, I illustrated a number of the processes that take place during and after an El Nino event, and also showed how the response of the East Indian and West Pacific Oceans to El Nino events created step changes in global Sea Surface Temperature (SST). Please review those posts (links follow) if you’re not familiar with them.
Can El Nino Events Explain All of the Global Warming Since 1976? – Part 1
Can El Nino Events Explain All of the Global Warming Since 1976? – Part 2
Supplement To “Can El Nino Events Explain All Of The Warming Since 1976?”
Supplement 2 To “Can El Nino Events Explain All Of The Warming Since 1976?”

Anthony Watts also posted them at Watts Up With That:

The questions, answers, comments and responses at Watts Up With That that remained on-topic provide additional insight.


One of the primary points discussed in the “Can El Nino Events Explain All of the Global Warming Since 1976?” series was how El Nino events redistribute warm water from beneath the surface of the Pacific Warm Pool to the surface of the East equatorial Pacific, in the NINO regions. Those warm waters are then driven to the East Indian and West Pacific Oceans by Pacific equatorial currents and by trade winds. In the East Indian and West Pacific Oceans, when not suppressed by volcanic eruptions, those redistributed warm waters cause a noticeable rise (step change) in SST anomalies, which then blend over time with the balance of the global oceans. I used a running total of NINO3.4 SST anomalies in Figure 21 of Can El Nino Events Explain All of the Global Warming Since 1976? – Part 2 to illustrate the phenomenon over the term of the data. As noted in the discussion of Figure 21 of that post, if a coefficient is applied to the NINO3.4 SST anomalies before the running total is calculated, the resulting curve will mimic the global temperature anomalies. Refer to Figure 1, below. This step in reproducing Global Temperature Anomalies creates a curve that simulates the slopes of the rises during the early and late 20th Century, and the slope of the decline in global temperature from the mid-1940s to the late-1970s.

The underlying curve of the Reproduction data was driven by ENSO events alone.

Figure 1


Note 1: The calculation of a running total is explained and illustrated well in Figure A of the following link:

Note 2: The NINO3.4 SST anomalies used in this post are from the Trenberth and Stepaniak NINO3.4 SST anomaly and TNI datasets found here:
http://www.cgd.ucar.edu/cas/papers/jclim2001b/ENflavorsr.htmlhttp://www.cgd.ucar.edu/cas/catalog/climind/TNI_N34/index.html#Sec5Specifically, it’s the NINO3.4 SST anomaly data found by following the links to:

Note 3: The Trenberth and Stepaniak NINO3.4 SST anomaly dataset is a standardized and filtered version of the HADSST NINO 3.4 SST (not anomaly) data here:
At the bottom of the last link is a notation that identifies HADSST as the source. The data runs from January 1871 to December 2007, which is the period of the graphs in this post.

Note 4: As illustrated later in this post, the same effect will occur if the Trenberth and Stepaniak NINO3.4 SST anomaly data is replaced with the HADSST version of NINO3.4 SST anomaly data. The anomalies, though, must be created with the period of 1950 to 1979 as the base years.

Note 5: The coefficient used as a scaling factor in Figure 1 is 0.0052. In the paper “Evolution of El Nino–Southern Oscillation and global atmospheric surface temperatures” (2000), Trenberth et al state on page 4, “The regression coefficient based on the detrended relationship is 0.094 deg C per N3.4 and is deemed more appropriate. The N3.4 contribution is given in Figure 3. It shows that for the 1997–1998 El Nino, where N3.4 peaked at ~2.5 deg C, the global mean temperature was elevated as much as 0.24 deg C (Figure 2)[Their Figure 2], although, averaged over the year centered on March 1998, the value drops to ~0.17deg C.” Dividing 0.17 deg C by 2.5 deg C leaves a yearly scaling factor of ~0.068 or a monthly scaling factor of ~0.00567 (0.068/12). I used 0.0052.

Note 6: The January 1871 data is also offset by -0.34 deg C to align it with the HADCRUT data.

Note 7: The accuracy of NINO3.4 SST anomaly data decreases before the opening of the Panama Canal in 1914. There was less ship traffic in the NINO3.4 area before that year. This could explain the divergence in the early part of the data.

Note 8: The Global Temperature Anomaly data (the blue curves in the following graphs) is HADCRUT3GL data. It is available through the Hadley Centre website:

This was as far as I originally intended to carry this post. I was going to use it as a supplement to the “Can El Nino Events Explain All of the Global Warming Since 1976?” series, but then the smoothness of the Global Temperature Anomaly Reproduction caught my eye. That curve would make a good basis from which to illustrate the minimal effects of the variations in global temperature due to changes in solar irradiance--those caused by the solar cycles. It could also highlight the short-term impacts of volcanic eruptions.


Figure 2 is a graph of monthly sunspot numbers that were scaled to illustrate the presently accepted impact on global temperature anomalies of variations in solar irradiance caused by solar cycles. Sunspot data was used as a proxy for Total Solar Irradiance because there are no TSI datasets that provide monthly data available--at least none that I’ve found. The data was shifted by the average sunspot number for the period of January 1871 to December 2007, creating “anomalies” from the average. The scaling factor used is 0.0006.

Note 9: Monthly Sunspot data from January 1749 to October 2008 is available through NASA’s Marshall Space Flight Center.http://solarscience.msfc.nasa.gov/greenwch/spot_num.txt
Figure 2

Adding that scaled Sunspot Number Anomaly data to the Global Temperature Anomaly Reproduction data shows the minimal effects of the solar-cycle-caused variations in solar irradiance on global temperature. Refer to Figure 3.
Figure 3

The differences between Figure 1 and Figure 3 are so small that it’s difficult to see them without a blink comparator or gif animation, Figure 4. Note how, due to the timing of the solar cycles, the La Nina trough at 1975/76 was lowered and El Nino peak in 1997/98 was increased, adding to the rise in global temperature over that period. With that said, it must also be noted that all La Ninas do not occur during solar minimums and all El Ninos do not occur during solar maximums.
Figure 4


The Sato Index of Stratospheric Mean Optical Thickness is available through GISS. In Figure 5, the Sato Index data has been inverted and scaled by a factor of -2.4 to show the impacts on Global Temperature anomalies of explosive volcanic eruptions. The scaling factor was selected to approximate the middle of the range of 0.2 to 0.5 deg C (the range of temperature drops associated with the Mount Pinatubo eruption of 1991).
Figure 5

Adding the scaled and inverted Sato Index data to the Global Temperature Anomaly Reproduction data, the effects of volcanic aerosols are illustrated. Refer to Figures 6.
Figure 6

Figure 7 is a gif animation that highlights the differences between the Global Temperature Anomaly Reproduction data that’s been adjusted for Solar Forcings and the data that’s also been adjusted for volcanic aerosols. Note how global temperature anomalies were depressed by volcanic aerosols from the early-1880s to the mid-1910s and from 1960 to the late-1990s. Excluding the two minor eruptions, the period between the mid-1910s and 1960 was quiet.

Figure 7


Figure 1 illustrated the slow, smooth variations in global temperature anomalies caused by the underlying small step changes in global temperature anomalies resulting from ENSO events. But ENSO events are also reflected in the global temperature record as positive “spikes” for El Nino events and negative “spikes” for La Ninas. Typical lag periods between the peak of an El Nino and the corresponding peak in global temperature anomaly are approximately 3 to 6 months. To simulate those effects (termed ENSO Noise in this post), the NINO3.4 SST anomalies are shifted 3 months and scaled by a factor of 0.09. Refer to Figure 8.
Figure 8

Adding the ENSO Noise to the Global Temperature Anomaly Reproduction data helps to amplify the peaks and valleys of the smoother data that was created with the running total. Refer to the gif animation, Figure 9. Note how it exaggerates the ENSO events and alters the impacts of the 1982 El Chichon and 1991 Mount Pinatubo volcanic eruptions.
Figure 9

This can be better illustrated in the short-term (January 1978 to December 2007) graph of the HADCRUT Global Temperature Anomaly data and that Global Temperature Anomaly Reproduction data that have been adjusted for Solar, Volcanic Aerosols, and ENSO Noise, Figure 10. There are some minor differences.
Figure 10


Figure 11 is a comparative graph of the HADCRUT Global Temperature Anomaly data and the Global Temperature Anomaly Reproduction data from January 1940 to December 1979. There are periods when the datasets diverge, but considering the recently found discontinuity in SST data in mid-1940s, the agreement is reasonable.
Figure 11

In Figure 12, the comparison of HADCRUT Global Temperature Anomaly data and the Reproduction of Global Temperature Anomaly data has been shifted forward in time to January 1900 to December 1944. From 1903 to 1944, the Reproduction mimics the HADCRUT Global Temperature anomaly date remarkably well, especially when the accuracy of the data during that period is considered.
Figure 12

The comparison from January 1871 to December 1909 is shown in Figure 13. Note how they diverge in 1903/1904.
Figure 13

In Figure 14, the Global Temperature Anomaly Reproduction data have been shifted up 0.3 deg C. The Reproduction data now mimics the HADCRUT data from 1871 to 1903. It appears there may have been a step change in the NINO3.4 anomaly or the HADCRUT Global Temperature Anomaly datasets around 1903/04.
Figure 14


Figures 15 and 16 compare the Reproduction of Global Temperature Anomalies (Solar plus Volcanic Aerosols plus ENSO Noise) to GISS and NCDC Global Temperature Anomaly data. A different scaling factor (0.0045) was needed for the running totals and the ranges were shifted to accommodate the differences in base years.
Figure 15
Figure 16


In Figure 17, the Trenberth and Stepaniak NINO3.4 SST anomaly data was replaced with raw HADSST NINO3.4 SST anomaly data (Base years = 1950 to 1979). The Trenberth and Stepaniak data has been smoothed and standardized, but the HADSST data has not. Therefore, the scaling factor for the running total was changed (to 0.0073) and the starting point was shifted to account for these differences. This graph was provided to show that the Reproduction of Global Temperature Anomalies was not solely a function of the Trenberth and Stepaniak NINO3.4 SST anomaly data or the smoothing and standardizing they used.
Figure 17


Note: This was originally a two-part post. I discovered an error in the spreadsheet that was used to create the graphs for the second part. When corrected, the additional adjustment did little to enhance the reproduction so I’ve deleted Part 2 and the supplement to it.

There will be those who will note that I provided no scientific basis for the use of a running total to simulate the long-term effects of ENSO. I have, however, in this post and the posts I’ve linked, illustrated the processes and their effects and explained why the running total reflects the underlying step changes caused by ENSO.

There will be those who will say that the reproducing the Global Temperature Anomaly curve with a running total of the NINO3.4 data is to be expected. I disagree. If that were so, why are there no studies noting the effect?


Erl said...

This is a very interesting construction Bob. Like you I consider that the effects of ENSO on global temperature must be accounted for before one can begin to account for other factors that might or might not be involved. However, I believe that the ENSO effects play out via a change in cloud cover (changing the amount of sunlight reaching the surface) in the regions that feed water into the tropics around the globe, in particular in the southern hemisphere. I am in a position to demonstrate that and will post that data soon. It can be seen that temperature increases occur in the surface waters of the South East Pacific, south of Nino 1+2 prior to and synonymous with the change in Nino 1+2 temperature. This increase in temperature in the surface waters south of Nino 1+2 is preceded by a fall in sea level atmospheric pressure and a rise in the temperature of the upper troposphere.

In August each year the upper troposphere at 30-40°S reaches a strong thermal maximum at a time when surface waters are at their thermal minimum. There is another 200hPa temperature peak in February.

This August upper air maximum temperature is associated with a season peak in ozone content. From August to March 200hPa temperatures stay high with ozone gradually decreasing but short wave solar radiation increasing over the period. If an El Nino event begins, this is when it manifests.

Each year in southern winter the southern stratosphere cools and the cool zone gradually descends towards the troposphere between mid and late year. Ozone descends with it. If solar influences react with that ozone in Southern Spring and summer, warming occurs down into the upper tropsophere. That warming, when sufficient, drives El Nino events.

Carl said...

Great post. Have you considered attending the Heartland's Int'l Conference on Climate Change? Too many scientists and amateurs are too focused on the sun and cosmic rays. Maybe you could get some more attention directed at ENSO by attending and passing out papers. I'll be attending and I'll deffinately try to draw attention to oceanic responses to ENSO.

Also, have you been able to use this site lately? http://nomads.ncdc.noaa.gov/cgi-bin/ncdc-ui/define-collection.pl?model_sys=sst&model_name=ersst&grid_name=999
I specify the dates I need, but now it will only give me data back to 1955.

It would be interesting if you adjusted the entire global temperature record for the immediate effects of ENSO events and then compared that dataset with the modeled response pre-ENSO noise. That would be the real test to see if your model is reproducing observed temperature trends.

Bob Tisdale said...

Carl: I have no plans to the Heartland Conference, not unless someone invites me down to New York for a cup of coffee. There are a few people I'd like to meet, though.

NOMADS is undergoing updates to include 2009 data and the long-term "All" selection for ERSST.v2 data is currently not available. I've exchanged emails with the person working on it. No word on when it will be corrected.

You can always use the KNMI Climate Explorer website. The data starts in 1880, but at least it's data.

There are a few more things I want to try before I go back and do other checks. With that in mind, I'm going to get back to preparing Part 2 of the post so I can get to the "new" things.



Tips are now being accepted.

Blog Archive

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.