On Hartmann and Wendler 2005 “The Significance of the 1976 Pacific Climate Shift in the Climatology of Alaska.”

StateWide_Change_1949-2012_F[1]Note: this is a companion article for the story: The Guardian’s Suzanne Goldenberg takes a fossil fueled trip to a remote Alaskan village to tell us recent global warming caused it to sink – but that’s not the cause  

I this essay, you’ll see why what some people perceive as “global warming” in Alaska is just part of a cycle. – Anthony


On Hartmann and Wendler 2005 “The Significance of the 1976 Pacific Climate Shift in the Climatology of Alaska.”

This post presents a number of problems the 2005 Hartmann and Wendler paper “The Significance of the 1976 Pacific Climate Shift in the Climatology of Alaska.” These include the misuse of the Pacific Decadal Oscillation (PDO) index and the erroneous conclusions derived from that misuse.

On page 14 of the 16-page paper, Hartmann and Wendler (2005) note:

This current debate presents the fact that more work is needed to further define North Pacific climate indices and implications.

Much of the obvious confusion could be eliminated if researchers stopped using abstract forms of sea surface temperature data like the PDO and began to use actual sea surface temperature anomalies. In other words, they should use the right tool for the job.


The abstract of Hartmann and Wendler (2005) begins:

The 1976 Pacific climate shift is examined, and its manifestations and significance in Alaskan climatology during the last half-century are demonstrated. The Pacific Decadal Oscillation index shifted in 1976 from dominantly negative values for the 25-yr time period 1951–75 to dominantly positive values for the period 1977–2001.

And the abstract concludes with following two sentences:

When analyzing the total time period from 1951 to 2001, warming is observed; however, the 25-yr period trend analyses before 1976 (1951–75) and thereafter (1977–2001) both display cooling, with a few exceptions. In this paper, emphasis is placed on the importance of taking into account the sudden changes that result from abrupt climatic shifts, persistent regimes, and the possibility of cyclic oscillations, such as the PDO, in the analysis of long-term climate change in Alaska.

The authors should be commended for considering the impacts of natural variability on Alaskan climate. However, they fail to recognize that the Pacific Decadal Oscillation (PDO) does not represent the sea surface temperature anomalies of the North Pacific north of 20N (the area on which the PDO data is based), or the Pacific Ocean as a whole, or the coastal waters of Alaska. This is clearly visible in Figure 1, which compares the Pacific Decadal Oscillation Index data (red) to the sea surface temperature anomalies of the North Pacific north of 20N (brown), the Pacific Ocean as a whole (light blue) and of the coastal waters of Alaska (dark blue).


Figure 1

The decadal variability of the PDO does not agree with the variations in the other three datasets. This can be shown by detrending and standardizing the sea surface temperature anomalies of the North Pacific north of 20N, the Pacific Ocean and of the coastal waters of Alaska and smoothing them (and the PDO index) with 121-month running-average filters. See Figure 2.


Figure 2

The Pacific Decadal Oscillation index is a statistically created dataset that does not represent the sea surface temperatures of the region of the North Pacific from which it is derived, nor does it represent the sea surface temperature anomalies of the Alaskan coastal waters, which should have a direct impact on the land surface air temperatures of Alaska. It therefore appears as though the authors used the wrong dataset as their reference for the sea surface temperatures.


There’s another way to illustrate the confusion caused by the use of the PDO index as a reference for North Pacific sea surface temperature anomalies, and that is in the context of climate shifts.

The primary focus of the paper was the impact of the 1976 Pacific Climate Shift on Alaskan climate—hence the title of the paper. The authors used temperature data starting in 1951 so we’ll use the same start year, and we’ll end the sea surface temperature data at present times.

Contrary to assumptions made using the PDO index, the sea surface temperature anomalies of the North Pacific north of 20N (which is the area from which the PDO index is derived) do not include a climate shift in 1976. The climate shift for the extratropical North Pacific occurred in 1988-89. See Figure 3. That’s about the same time as the climate shift in North Pacific ocean heat content.


Figure 3

On the other hand, the sea surface temperature anomalies of the Alaskan coastal waters did experience a climate shift in 1976, Figure 4. Sea surface temperature anomalies cooled from 1951 to 1975, and cooled again from 1977 to present. The two cooling trends before and after the 1976 climate shift suggest that the 1976 shift is responsible for the long-term trend of 0.055 deg C per decade for the Alaskan coastal waters.


Figure 4

Hartmann and Wendler (2005) could have used the actual sea surface temperature anomalies for the Alaskan Coastal Waters in their paper. It would have kept to their overall objective of citing the 1976 Pacific Climate Shift as a cause for the long-term warming and would have avoided the confusion associated with the use of the PDO index, which is an abstract form of sea surface temperature data.


To help clarify a few things, let’s borrow the next two illustrations from an earlier post, Blog Memo to John Hockenberry Regarding PBS Report “Climate of Doubt”. As noted in Figure 3, the 1976 climate shift is not a North Pacific phenomenon. It’s a shift in East Pacific sea surface temperature anomalies, Figure 5. It effectively shifted the sea surface temperatures of the East Pacific (90S-90N, 180-80W) up 0.17 deg C. In other words, that’s a natural rise in sea surface temperatures of almost 0.2 deg C for 33% of the surface area of the global oceans.


Figure 5

The Great Pacific Climate Shift also refers to the change in the basic state of the ocean processes taking place in the tropical Pacific. See Figure 6. After 1976, El Niño events dominated, but for the period from the early-1940s to 1976, El Niños and La Niñas were more evenly matched, with La Niñas just a little bit stronger.


Figure 6

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
Rud Istvan
May 14, 2013 6:32 pm

Wonderful post, and wonderful comment: use the right tool for the job at hand.

May 14, 2013 6:37 pm

Feel bad for alot of Alaskans living with the change. Alaska is already a tough place to live

Ian Wilson
May 14, 2013 7:43 pm

I read almost all your posts and I am always impressed with the quality and depth of your analysis of the long term changes in ocean and atmospheric conditions. However on this one point, I must point out that you are not being totally honest with your audience.
The PDO index was never supposed to be a direct indicator of the sea-surface temperature anomaly conditions in the Nth Pacific Ocean. Granted, it may have been incorrectly used by some/many to incorrectly imply that it was an indicator of the average sea-surface temperature anomaly of the Nth Pacific Ocean. In this case, you efforts in pointing out this factual error are to be applauded. However, you are not correct in implying (either directly or indirectly) that the PDO index was intended to be understood as an ocean-wide sea-surface temperature anomaly index.
The PDO index was designed to selectively highlight a PATTERN in sea-surface temperature anomalies.
When the PDO index is positive, the sea-surface temperature anomalies in the central and eastern equatorial Pacific are higher-than-normal, and.those of the central and western Nth Pacific ocean are lower-than-normal. During the positive PDO phase, the sea-surface temperature anomalies along the coast of Nth America (including Alaska) are the opposite of that of the central and western Nth Pacific (i.e north of about 20 degrees latitude).
When the PDO index is negative, the seas-surface temperature anomaly patterns are inverted for each of the regions discussed above.
Hence, when you take calculate a single sea surface temperature for the Nth Pacific ocean you are just averaging two distinct regions of the sea-surface temperature anomaly pattern that are almost always opposed in temperature. You then conclude that the PDO index is not measuring a ocean-wide temperature. Now, technically what you are saying is correct. However,
in this particular case, calculating an ocean wide average temperature anomaly hides the very thing that the PDO index was designed to measure. This you claim that the PDO index is a poor indicator of the ocean-wide sea-surface temperature anomaly is disingenuous at the very least.
The PDO index is a good statistical measure of change in the PATTERN of the sea-surface temperature anomalies for the whole Pacific ocean (yes, it extends to the South Pacific as well).
The problem lies with the wrong way people interpret and use the PDO index, not with the index itself. It does the job for which it was designed perfectly!

May 14, 2013 8:11 pm

Ian – I thought Bob did a good job of making exactly that point.

May 14, 2013 9:44 pm

The PDO is so fundamental, it is so likely the basis of the 60 year period observed in many disparate statistical analyses, that it must have a fundamental mechanism. What have we to choose from? The sun? Nope. ENSO? Nope. I submit it is the thermohaline circulation. In the current phase (I refuse to get into this positive/negative value judgment) the thermohaline circulation reinforces the Humboldt Current and drives cold subsurface water into the uterus of Ninos, aborting them. In the opposite phase the thermohaline circulation drives across the Pacific Warm Pool and across the equator to reinforce the Kuroshio Current bringing rain to China and warm water to Alaska.
Is this the reason atmospheric warming has stopped? Probably. Is this the reason the Oceans (so far on average) are retaining heat? Probably. Does this mean catastrophic atmospheric warming will resume in 45 years? That depends. Carbon dioxide does not warm the oceans. High energy radiation does. The oceans warm the atmosphere, when they can.
Bear in mind the age of the bottom water the THC is offering. What meaning has a 60 year cycle when there is a millennium cold of water in the pipeline and the water emerging today off the coast of Peru sank to the abyss hundreds of years ago?

May 14, 2013 11:38 pm

Interested to hear your views as to why Alaskan land temperatures follow PDO, whereas sea surface temperatures don’t!
Great article.

May 15, 2013 12:13 am

I remember that year of change. I faced death 3 times in 1976, where I didn’t see where I would live through the next 60 seconds, on June 6th then every 2 months. I remember feeling that something was different with the weather. It was a turning point. That was when the flood cycle switched from 9 years to almost 12 years in the Pacific Northwest.

May 15, 2013 3:12 am

Good stuff. There’s something especially interesting in Figure 1. The Alaskan waters seem to be following the shorter (7-year or so) wiggles of PDO, but NOT the longer 70-year wiggles. In other words, those waters are a high-pass filter. There aren’t a lot of HP filters in Nature, especially when big bodies of water are involved.

May 15, 2013 3:47 am

Faintly ridiculous to ascribe a cycle or osscillation to a statistically defined metric for which there are less than three full cycles known.
Like the ENSO ‘cycle’ it is just as likely to be an inherently unpredictable fluctuation that MAY have a charateristic timescale and magnitude, but these can vary from ‘cycle’ to ‘cycle’ by fifty percent.

Myron Mesecke
May 15, 2013 6:56 am

Figure 1. Brown, red, blue and dark blue. For people like me with a bit of color blindness this just looks like a mess. People need to understand that not everyone has perfect color perception and use four very different colors on graphs like this. Better choices are blue, red, green, orange, black, magenta, etc. While taking GIS classes I learned how important it is to choose a color scheme for attributes that enhance visibility rather than makes it difficult to see.

May 15, 2013 8:15 am

Bob, in Fig 3 consider the trends of the low temperature excursions. Down to 1976, up to 2004, and now down again. Could the cold excursions be more important, or more significant than the average?

lemiere jacques
May 15, 2013 9:23 am

worse than we though, in good old days, we had a changing climate, and that was quite bad dam ice age! , than we had climate change,and it was worse than before; than we re having climate change change, worse than worse.

James at 48
May 15, 2013 10:03 am

Peak warmth gave us Peak Civilization. Now, the long slide downward.

%d bloggers like this:
Verified by MonsterInsights