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

Tuesday, April 29, 2008

Is There A Cumulative ENSO Climate Forcing?


At present, I will consider this an oddity or curiosity until someone figures out how and why this works. Also, before you run off and try this with another ENSO index, be forewarned: the effect of the running total is very sensitive to the base year.


While looking for a trend in long-term NINO3.4 Data, I created a running total of the data, where the value of each subsequent year was added to the prior year total, as illustrated in the following table and graph.

Year ...... NINO3.4 (deg C) ...... NINO3.4 Running Total
1871 .......... -0.42267....................... -0.42267
1872 .......... -0.76575....................... -1.18842
1873 .......... -0.76925....................... -1.95767
1874 .......... -1.2345.......................... -3.19217

Figure 1

There was no mistaking the shape of the curve, but the scale was wrong. So I added only a percentage of each NINO3.4 reading to the previous total, finally settling on 6.8%.

The following graph is the “NINO3.4 Running Total with a 6.8% Coefficient” compared to the global temperature anomaly (HADCrut3GL).

Figure 2

Both smoothed with a 5-Year filter.

The divergence prior to 1909 is easily attributable to data errors in sampling and calculating global temperature and NINO3.4. In fact, in the following reference and data source, the authors state, “Because the number of observations in the tropical Pacific drops greatly prior to about 1950, the quality of the analyses is not as good as in recent years and structures of SST anomalies are partially imposed by the method of analysis, which used empirical orthogonal functions as a means of spatial interpolation. Data are very sparse in the Pacific prior to the opening of the Panama Canal in 1914, and many months do not contain any observations in the N3.4 or Niño 4 regions.”


The smoothed and normalized El Niño Reconstruction is from “Indices of El Niño Evolution” (2001), Kevin E. Trenberth, and David P. Stepaniak J. Climate, 14, 1697-1701.

Updated raw NINO3.4 data attached to the second link:


Why would a running total of ENSO data correlate to global temperature anomaly? It appears that the atmospheric and SST heat gain (or loss) from each El Niño (or La Niña) impacts global temperature directly and then mixes and decays while the Rossby waves travel throughout the oceans over the following years, continuing their effect on atmospheric temperatures but to a lesser degree. This would require a longer decay rate than normally calculated by GCMs.

In a controversial paper, “Decade-scale trans-Pacific propagation and warming effects of an El Niño anomaly”, (1994) Jacobs et al identified decade-long residual effects of the 1982-83 El Niño.

The abstract:

“El Niño events in the Pacific Ocean can have significant local effects lasting up to two years. For example the 1982-83 El Niño caused increases in the sea-surface height and temperature at the coasts of Ecuador and Peru1, with important consequences for fish populations2,3 and local rainfall4. But it has been believed that the long-range effects of El Nino events are restricted to changes transmitted through the atmosphere, for example causing precipitation anomalies over the Sahel5. Here we present evidence from modeling and observations that planetary-scale oceanic waves, generated by reflection of equatorial shallow-water waves from the American coasts during the 1982–H83 El Niño, have crossed the North Pacific and a decade later caused northward re-routing of the Kuroshio Extension—a strong current that normally advects large amounts of heat from the southern coast of Japan eastwards into the mid-latitude Pacific. This has led to significant increases in sea surface temperature at high latitudes in the northwestern Pacific, of the same amplitude and with the same spatial extent as those seen in the tropics during important El Niño events. These changes may have influenced weather patterns over the North American continent during the past decade, and demonstrate that the oceanic effects of El Niño events can be extremely long-lived.”

Note that in 1994 the Pacific Decadal Oscillation (PDO) had yet to be identified. Though the above description bears a similarity to the PDO, the source of the PDO is still controversial. Do higher frequencies of El Niños (La Niñas) produce a positive (negative) PDO, or does a positive (negative) PDO trigger a higher frequency of El Niños (La Niñas)? Regardless, El Niños/La Niñas appear to have a longer effect on climate than climate models predict--or a longer effect than modelers share in their reports.

Is the divergence of high-latitude Northern Hemisphere temperature in the following graph an example of the long-term effect of the 1997/98 El Niño? We’ve been told that GCM-predicted polar amplification results from greenhouse gas emissions, but the amplification from 2000 to present seems to stem from the 1997/98 El Niño, not from anthropogenic sources, which should be gradual.

Figure 4

The data for the above graph is available at:


Someone will note that the data I originally used was incomplete and ended in 2002. After finding the above, I scrambled for data and spliced on ONI data from 2000 to 2007 as a quick fix. You'll soon discover that my data mixing made little difference.
To determine the effect of the normalization of the NINO3.4 data, I ran through the “Recipe” included in the second linked reference, but excluded the smoothing and normalization steps.
First the Raw SST data:

Figure 5

The Anomaly data with 1950 to 1979 average base as reference:

Figure 6

Using a new coefficient, 0.09, I recreated the graph comparing NINO3.4 running total versus global temperature anomaly.

The running total data also had to be shifted -0.1 deg C to account for the absence of previous normalization adjustments.
In upcoming posts, I’ll illustrate that the same process works with other ENSO indices and while using monthly ENSO and global temperature anomaly data.

Sunday, April 27, 2008

Mann et al "Weighted Dust Veil Index"

It appears the data in the Mann et al "Weighted Dust Veil Index" has been manipulated. Those who use this Northern Hemisphere volcanic DVI data must be aware of its deficiency. I downloaded the data and plugged it into a spreadsheet, then wondered why the data or results appeared unusual when graphed and used in a comparative study.

It's easy to read past what the file title really says. Read the title again. “Weighted Dust Veil Index”. It includes the word “Weighted”, which indicates the data has been biased, or slanted, or prejudiced. When I determined what Mann et al had done, I asked myself, “What part of weighted didn’t you understand, Bob?”

The Mann et al Weighted Dust Veil Index (DVI) data for the Northern Hemisphere is available here: http://www.ncdc.noaa.gov/paleo/ei/ei_data/volcanic.dat

A graph of the data:


The magnitudes of the Nicaraguan Coseguina eruption in 1835, the Indonesian Tambora eruption in 1815, and the 1766 eruption of the Mayon volcano in the Philippines mask the problem with the Mount Pinatubo data.

The 1991 through 1993 weighting is significant. It more than doubles the impact of the Mount Pinatubo eruption and, by doing so, diminishes all those before it.

First, an overview of the Lamb Dust Veil Index from NASA: “Lamb (1970) formulated the Dust Veil Index (DVI) in an attempt to quantify the impact on the Earth's energy balance of changes in atmospheric composition due to explosive volcanic eruptions. The DVI is a numerical index that quantifies the impact of a particular volcanic eruption's release of dust and aerosols over the years following the event. This package ... provides DVI's for the period 1500 - 1983 (DVI = 1,000 for Krakatoa in 1883), along with DVI estimates for the eruptions of Santorin in 1470 B.C., Vesuvius in 79 A.D., and Oraefajokull in 1362 A.D.”

For further information: http://gcmd.nasa.gov/records/GCMD_CDIAC_NDP13.html

For Lamb Dust Veil Index Data: ftp://cdiac.esd.ornl.gov/pub/ndp013/

Back to Mann et al DVI: The Mann et al DVI data is a supplement to the "Mann, M.E., et al., 2000, Global Temperature Patterns in Past Centuries: An Interactive Presentation, IGBP Pages/World Data Center for Paleoclimatology Data Contribution Series #2000-075. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA," henceforth Mann2000. It contains one volcanic reference: "Briffa, K. R., P. D. Jones, F. H. Schweingruber, and T. J. Osborn, 1998: Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years. Nature, 393, 350–354." There were no values provided for the Dust Veil Index in the Briffa reference.


The Mann 2000 paper also references Mann et al 1998, henceforth MBH98, which provides other possible sources of the data.


MBH98 discusses DVI as follows: "…and (3) the weighted historical 'dust veil index' (DVI) of explosive volcanism (see Fig. 31.1 in ref. 40) updated with recent data 41." The references are:

"40. Bradley, R. S. & Jones, P. D. in Climate Since A.D. 1500 (eds Bradley, R. S. & Jones, P. D.) 606–622 (Routledge, London, 1992)."


"41. Robock, A. & Free, M. P. Ice cores as an index of global volcanism from 1850 to the present. J. Geophys. Res. 100, 11549–11567 (1995)."


The Robock & Free "Ice Core" reference: In it, there is no assignment of a DVI value for Pinatubo or for the year 1991. It is a study of Ice Core data that uses the Lamb DVI as one of many sources of data for comparisons of volcanic indices. Robock and Free modified Lamb DVI for this study, to "eliminate the temperature influence…" in the original Lamb data. This revision is not reflected in the Mann et al DVI data, since Mann DVI uses unmodified Lamb DVI data as its source. Robock and Free also discuss in this report that the DVI was updated "to include estimates for the period from 1983 to 1995, accounting for the decay of the El Chichon cloud, the small Augustine eruption, and the 1991 eruptions of Mount Pinatubo and Hudson." But Robock and Free do not provide the DVI values for those years except in comparative graphs expressed as optical depth, not DVI.

El Chichon-1982: Mann et al use the Lamb DVI data for this year, which includes data for El Chichon and two additional eruptions in 1982: Pagan and Galunggung. No weighting was performed by Mann et al on this data.

Augustine-1986: The Mann et al DVI data does not reflect any DVI in 1986 for this Alaskan volcano or in the three years that follow. Mann et al weighted its DVI value by omission.

Mount Pinatubo-1991: In the "Ice Core" study, Robock and Free do cite another of Robock's studies from 1995 (Robock and Mao, "The Volcanic Signal in Surface Temperature Observations" in the Journal of Climate),


In it a DVI/Emax value of 1000 is assigned to Mount Pinatubo (1991), located at 15N. This value is confirmed by the Global DVI value of 1000 in a 2002 report by Pitari and Mancini ("Short-term climatic impact of the 1991 volcanic eruption of Mt. Pinatubo and effects on atmospheric tracers" published in Natural Hazards and Earth System Sciences (2002) 2: 91–108).


Note: Since Mount Pinatubo is a low-latitude volcano (within 20S and 20N), the EMax would be 1, making the Mount Pinatubo values in the Robock and the Pitari studies equal. The Mount Pinatubo weighting is discussed further later.

Hudson-1991: Cerro Hudson is a Chilean volcano, located at 46S. The Robock "Ice Core" study included the Southern Hemisphere where the Cerro Hudson data would used, but due to its location and relatively small size, it would have had little to no influence on Northern Hemisphere data.

The Second MBH98 DVI Reference: The Bradley and Jones discussion of Figure 31.1 in "Climate Science Since A.D. 1500" (page 608) states, "Cumulative DVI for the northern hemisphere, assuming the dust from an individual eruption is apportioned over four years with 40% of each DVI assigned to year 1, 30% to year 2, 20% to year 3, and 10% to year 4… …(DVI values from Lamb 1970, 1977, 1983)."

For example, the Lamb Northern Hemisphere DVI value for the 1500 A.D. Java eruption is 500. The DVI readings in the Mann et al data are 200 at 1500 A.D., 150 at 1501 A.D., 100 at 1502 A.D., and 50 at 1503 A.D. Prior to 1985 (1982 plus three years of decay), all Mann et al DVI values use the Lamb Northern Hemisphere DVI data and this relationship. There were no 1991 to 1993 DVI values (Mount Pinatubo) in the Lamb DVI data because the Lamb study ended in 1983. It was here that Mann et al weighted the Lamb DVI data by tagging on Mount Pinatubo data, which, as discussed later, is inflated, hence the use of the word "Weighted" in the title of the data.


Mann et al added one volcano to an existing data base of over 400 volcanic eruptions to create their Weighted Volcanic Dust Veil Index. By inflating the Pinatubo eruption they chopped the effects of all those that preceded it.

There is no explanation accompanying the Mann et al DVI data as to how the update was carried out. However, if Mann et al followed the data apportionment process in 1991 through 1993, this would dictate that the Northern Hemisphere DVI value for the Mount Pinatubo eruption would have to have been 1250 (500/0.4). This NH value is greater than the Global DVI/EMax value of 1000 for Mount Pinatubo listed in Table 1 of the second Robock report and the Global DVI value of 1000 in the report by Pitari and Mancini. Based on Figures 1 and 2 of the Robock "Ice Core" study, the impact of Mount Pinatubo was evenly distributed between NH and SH. This would equate to a Global DVI of 2500 if the MBH apportionment was not weighted.

Mann et al appear to have divided the Global value of 1000 by 2 to calculate the Northern Hemisphere DVI value of 500 at 1991. Unfortunately, it also appears they failed to follow the apportionment for a cumulative DVI value of 500, where the first through fourth year values should be 40% of 500 (200), 30% (150), 20% (100), 10% (50). They used 100% when applying it to the first year, then downscaled from there.

The cumulative DVI data for the Mount Pinatubo eruption in the Mann 2000 supplemental table reads:

Year DVI
1991 500.0
1992 375.0
1993 250.0
1994 125.0

The following graph illustrates Northern Hemisphere DVI data and polynomial trend from 1850 to 1995 using the Mann data.


Based on a Global DVI of 1000 and a hemispheric DVI of 500, the values calculated by the cumulative method (40%, 30%, 20%, 10%) should be:

Year DVI
1991 200.0
1992 150.0
1993 100.0
1994 50.0

The following graph illustrates Northern Hemisphere DVI data, with polynomial trend, that has not been weighted, from 1850 to 1995, using the proper method of calculating the cumulative DVI.


Note the improved relationship between the two eruptions of equal DVI, 1883 Krakatau (Plus St Augustine) and 1991 Mount Pinatubo. Also note the disappearance of the "Hockey Stick" effect in the polynomial trend line.

It could be argued that a cumulative DVI value of 200 in 1991 is not reflected by the Robock graphs, where the optical depth values for Mount Pinatubo is significantly higher than El Chichon and slightly higher than Krakatau in 1883. Refer Figure 9 of the Robock "Ice cores" study.


It appears the Robock and Free adjustments did more to lower the El Chichon value than it did to raise Mount Pinatubo's. If Mann et al intended to weight the Mount Pinatubo eruption so that its relationship with El Chichon was represented better, then the El Chichon values should have been lowered, so that the correlation between Mount Pinatubo and other eruptions of equal magnitude was preserved.


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.