UPDATE July 2, 2011: I’ve modified the wording of the second paragraph after Figure 4.
I’ve written numerous posts that describe the Pacific Decadal Oscillation (PDO), what the PDO represents, and, just as important, what it does not represent. For those new to the PDO and for those needing a refresher, refer to An Introduction To ENSO, AMO, and PDO — Part 3.
IceCap recently published a post about the PDO. One of the illustrations in that post confirms a point I have been making: that the PDO does not represent the Sea Surface Temperatures of the North Pacific. And I’ve added a few additional comments and clarifications about the IceCap post.
In my post An Introduction To ENSO, AMO, and PDO — Part 3, the discussions of the PDO patternwere about the spatial pattern. But the word pattern can also be time-related and, therefore, it could pertain to the PDO’s “behavior in time”. A recent discussion on a blog thread gave me the impression that the multiple meanings of the word pattern could be the cause of some of the confusion about the PDO.
Also, much of the post An Introduction To ENSO, AMO, and PDO — Part 3discussed and illustrated the fact that the PDO does not represent the Sea Surface Temperature Anomalies of the North Pacific. There is another factor that may lead to some of the continued misunderstandings about the PDO. The PDO data is standardized. This could greatly exaggerate the magnitude of its variations. Could the standardization also inflate its perceived importance?
I had also wanted to include the errors in the SkepticalScience post It’s Pacific Decadal Oscillation . The SkepticalScience post was obviously written by someone who never plotted the Sea Surface Temperature anomalies of the North Pacific, who misunderstands how the PDO is calculated and what it represents, and who misunderstands or elects to misrepresent climate oscillations. But this post is long, so I’ll present the Skeptical Science errors in a separate post.
THE ICECAP PDO POST
Figure 1 is a screen cap of the opening of Joe D’Aleo’s post about the PDO at IceCap. I’ve included the screen cap because if you were to click on the post title Is the PDO real or a skeptic inventionat IceCap the link does not bring you to the post; it links to a .pdf document titled “THE PDO.” Most of the IceCap post and the linked .pdf document appear to be the same, but their opening paragraphs and the titles are different.
That aside, most of the IceCap post is intended to confirm the existence of the PDO, which is something I don’t dispute. If you’ve watched any of the SST .gif animations or videos I’ve prepared, the positive and negative PDO spatial patterns are very visible. And they should be; the PDO spatial pattern is the most prevalent of the many that form in the North Pacific SST anomalies. The IceCap post also mentions using the Pacific Decadal Oscillation in weather forecasts. Since there are a number of papers that describe weather patterns associated with the PDO, I imagine the PDO data is a useful index for meteorologists.
But one of the points that I have illustrated and discussed a number of times is that the PDO does not represent the Sea Surface Temperature anomalies of the North Pacific, north of 20N. The IceCap post includes a group of PDO- and ENSO-related maps, Figure 2. And those maps actually illustrate that the PDO does not represent the Sea Surface Temperature (SST) anomalies of the North Pacific.
I’ve modified the illustration, Figure 3, by highlighting the area used to calculate the PDO data and by covering the rest of the maps with text. The PDO is not calculated from the SST data south of 20N, so all of that additional visual information exaggerates the surface area represented by the PDO.
My text in Figure 3 reads: In The Positive (Warm) Phase Map Of The PDO Shown Above Left, The Sea Surface Temperature Anomaly For The North Pacific North Of 20N Appears To Be Negative. That Is, The Negative SST Anomalies Cover A Greater Surface Area And They Are More Intense Than The Positive Anomalies. The Opposite Holds True For The Map On The Right-Hand Side. How Then Could A Positive (Warm) Phase Of The PDO Raise Global Surface Temperatures, And Vice Versa?
Someone is bound to note that the IceCap post does not mention that the PDO has an impact on Global surface temperatures. Agreed. Their post doesn’t. This part of my discussion is about the common belief that the sign of the PDO dictates whether global temperatures rise or fall. Since the sign of the PDO does not represent the Sea Surface Temperature of the North Pacific the belief is unfounded.
In past posts, I’ve presented that there is an inverse relationship between the PDO and the Sea Surface Temperature (SST) Anomalies of the North Pacific. This can be shown with a graph that compares the PDO data and the difference between the SST anomalies of the North Pacific north of 20N and global SST anomalies. For simplicity sake, we’ll use the term “North Pacific Residual” for the data that’s calculated as the North Pacific SST anomalies minus the Global SST anomalies. Refer to Figure 4. We’ll use the North Pacific Residual in the comparison graph because one of the steps taken to calculate the PDO is to subtract Global SST anomalies from the SST anomalies of each 5 degree by 5 degree grid of the North Pacific north of 20N. Both datasets in Figure 4 have been smoothed with 121-month running average filters. Other than the agreement between the multidecadal variations in the two curves, there are a couple of things to note about the graph. First, the PDO data has been inverted; that is, it’s been multiplied by a negative number. Second, the PDO data has also been scaled by a factor of 0.2. That was an arbitrarily chosen round number I used to bring the variations of the PDO down into line with the North Pacific Residual data. But let’s look at the scaling in another way: the multidecadal variations in the PDO data are five times higher than the actual variations in the North Pacific Residual data in that graph. One might conclude the PDO data exaggerates the actual multidecadal variations in North Pacific SST anomalies. More on that later.
Figure 4 is taken from the post An Inverse Relationship Between The PDO And North Pacific SST Anomaly Residuals.
CORRECTION: I have modified the wording of the next paragraph. Note the strikes through the 2 “a” and the replacements with “the”.
The SST anomaly data for the North Pacific is part of the Global Surface Temperature data, but the PDO data is not. This inverse relationship between the PDO and the SST anomaly data of the North Pacific directly contradicts the assumption that a positive (warm) PDO is responsible
a the rise in global temperatures or that a negative (cold) PDO is somehow responsible for a the drop in global temperatures.
The IceCap post attempts to resurrect the argument that there is no clear evidence that ENSO drives the PDO. IceCap bases their brief discussion on a quote from the 14-year old Mantua et al (1997) paper “A Pacific interdecadal climate oscillation with impacts on salmon production.” The IceCap post reads, “The authors made no claim as to which (PDO or ENSO) was the chicken and which the egg.” And IceCap includes a quote from Mantua et al (1997):
“The ENSO and PDO climate patterns are clearly related, both spatially and temporally, to the extent that the PDO may be viewed as ENSO-like interdecadal climate variability (Tanimoto et al. 1993; ZWB). While it may be tempting to interpret interdecadal climatic shifts as responses to individual (tropical) ENSO events, it seems equally conceivable that the state of the interdecadal PDO constrains the envelope of interannual ENSO variability.”
That paragraph from Mantua et al (1997) appears to contradict a paper they reference. Mantua et al (1997) calculate the PDO using a method that was presented in Zhang et al (1997) ENSO-like Interdecadal Variability: 1900–93. In Zhang et al (1997), the PDO was identified as “NP”, and they use Cold Tongue Index SST anomalies (CT) as the El Niño-Southern Oscillation (ENSO) proxy. Zhang et al (1997) note:
“Figure 7 shows the cross-correlation function between CT and each of the other time series in Fig. 5. The lag is barely perceptible for TP and G and it increases to about a season for G – TP and NP, confirming that on the interannual timescale the remote features in the patterns shown in Fig. 6 are occurring in response to the ENSO cycle rather than as an integral part of it…”
It would be difficult for the PDO to drive ENSO if ENSO leads the PDO by a season and if the PDO spatial pattern is a “response to the ENSO cycle rather than…an integral part of it.”
In the 14 years since Mantua et al (1997) was published, there have been a number of papers that confirm that ENSO drives the PDO. In ENSO-Forced Variability of the Pacific Decadal Oscillation, Newman et al (2004) also found that the PDO lags ENSO. They describe cell d of their Figure 1 as:
“ENSO also leads the PDO index by a few months throughout the year (Fig. 1d), most notably in winter and summer. Simultaneous correlation is lowest in November– March, consistent with Mantua et al. (1997). The lag of maximum correlation ranges from two months in summer (r ~ 0.7) to as much as five months by late winter (r ~ 0.6). During winter and spring, ENSO leads the PDO for well over a year, consistent with reemergence of prior ENSO-forced PDO anomalies. Summer PDO appears to lead ENSO the following winter, but this could be an artifact of the strong persistence of ENSO from summer to winter (r = 0.8), combined with ENSO forcing of the PDO in both summer and winter. Note also that for intervals less than 1yr the lag autocorrelation of the PDO is low when the lag autocorrelation of ENSO (not shown) is also low, through the so-called spring persistence barrier (Torrence and Webster 1998).”
And the first sentence of the Conclusions of Newman et al (2004) reads (their italics):
“The PDO is dependent upon ENSO on all timescales.”
My post An Introduction To ENSO, AMO, and PDO — Part 3 included the same discussions of those two papers under the heading of DOES THE PDO DRIVE ENSO?
A more recent paper, Shakun and Shaman (2009) “Tropical origins of North and South Pacific decadal variability” also confirms that the PDO is an aftereffect of ENSO. In addition to the PDO, they use the acronym PDV for Pacific Decadal Variability.
The Shakun and Shaman (2009) Conclusions read:
“Deriving a Southern Hemisphere equivalent of the PDO index shows that the spatial signature of the PDO can be well explained by the leading mode of SST variability for the South Pacific. Thus, PDV appears to be a basin-wide phenomenon most likely driven from the tropics. Moreover, while it was already known PDV north of the equator could be adequately modeled as a reddened response to ENSO, our results indicate this is true to an even greater extent in the South Pacific.”
These papers confirm my statements from past posts that the PDO is an aftereffect of ENSO. And that brings us to the two tables from the IceCap post, which I have merged into one graphic, Figure 5. IceCap introduces the tables by stating:
“During the positive phase see the dominance of more frequent, stronger, longer La Ninas and the positive PDO mode, more frequent, stronger and longer El Ninos.”
There is an obvious error in that sentence. It should begin with “During the negativephase…”
Now I do realize that IceCap has not stated that the positive PDO is responsible for the more frequent, stronger and longer El Niño events and vice versa, but they implied it. And that contradicts the papers above. Since the PDO is an aftereffect of ENSO, a period when El Niño events dominated would cause the PDO to be positive, and vice versa for epochs when La Niña events dominate.
A SIMPLE DESCRIPTION OF HOW THE PDO SPATIAL PATTERN IS FORMED
To reinforce the discussion above, the following is a simple explanation of the processes that cause the PDO spatial pattern during ENSO events.
A positive PDO spatial pattern is characterized by SST anomalies in the eastern North Pacific that are higher than the SST anomalies in the central and western North Pacific. That positive PDO pattern is created by the response of the North Pacific SST anomalies to an El Niño event. Changes in coupled ocean-atmosphere processes, resulting from the El Niño, can cause an increase in the SST anomalies in the eastern North Pacific. Since the El Niño causes a reversal of trade winds in the western tropical Pacific, less warm water than normal is spun up into the western and central North Pacific (an area called the Kuroshio-Oyashio Extension or KOE), and SST anomalies of the western and central North Pacific drop. The initial drop in the western and central North Pacific is also driven by the changes in coupled ocean-atmosphere processes that are caused by the El Niño. The reverse holds true during a La Niña in the eastern North Pacific. For the western and central North Pacific during a La Niña, the leftover warm water from the El Niño also gets spun up into the KOE, adding to the warm waters being brought there by the increased strength of the trade winds, both of which raise SST anomalies there.
There are differences between the PDO and an ENSO proxy such as NINO3.4 SST anomalies from time to time. (NINO3.4 SST anomalies are a commonly used proxy for the frequency and magnitude of El Niño and La Niña events.) The reason for this is that other factors can impact how the North Pacific SST anomalies respond to ENSO events. These other factors include shifts in sea level pressure in the North Pacific and a phenomenon called The Reemergence Mechanismalong the Kuroshio-Oyashio extension (KOE). Aerosols from explosive volcanic eruptions should also account for some of the differences between the PDO and an ENSO proxy, though I have never seen this discussed in any papers.
CAN THE MULTIPLE MEANINGS OF THE WORD PATTERN ADD TO THE CONFUSION ABOUT THE PDO?
The Joint Institute for the Study of the Atmosphere and Ocean Joint Institute for the Study of the Atmosphere and OceanJoint Institute for the Study of the Atmosphere and Ocean (JISAO) Pacific Decadal Oscillation (PDO) webpage is a primary source for PDO information and data. Two maps are used by JISAO to illustrate the spatial patterns that can exist during the warm and cool phases of the PDO, Figure 6. I’ve highlighted the area of the North Pacific represented by the Pacific Decadal Oscillation. It is the area north of 20N, and only that area. The word “pattern” in the opening paragraph on that webpage refers to the “spatial climate fingerprint” of the North Pacific north of 20N, not the multidecadal variability of its data. As discussed previously in this post, when the PDO data is positive (warm phase), the SST anomalies in the eastern North Pacific are warmer than those in the western and central North Pacific, and when the PDO is negative (cool phase), the SST anomalies in the eastern North Pacific are cooler than the SST anomalies in the western and central north Pacific.
Farther down on the JISAO PDO webpage, the PDO is described as:
The “Pacific Decadal Oscillation” (PDO) is a long-lived El Niño-like pattern of Pacific climate variability. While the two climate oscillations have similar spatial climate fingerprints, they have very different behavior in time.
Again, the word pattern is being used to describe spatial characteristics of the SST anomalies.
As discussed in An Introduction To ENSO, AMO, and PDO — Part 3, the phrase “El Niño-like pattern” does NOT mean that the North Pacific (north of 20N) has a separate El Niño-like event.It refers to the fact that a typical El Niño event creates a spatial pattern in the North Pacific where it is warmer in the east than it is in the central and western portions, and a typical La Niña event will create the opposite pattern, cooler in the east than it is toward the center and west of the North Pacific.
Figure 7 is a time-series graph that compares the PDO and NINO3.4 SST anomalies (a commonly used proxy for the frequency and magnitude of El Niño and La Niña events). Both datasets have been smoothed with a 121-month filter. Keep in mind that the NINO3.4 SST anomalies represent exactly that, the SST anomalies of an area of the tropical Pacific called the NINO3.4 region, which is bordered by the coordinates of 5S-5N, 170W-120W, while the PDO does not represent the SST anomalies of the North Pacific. The PDO is a statistically manufactured dataset. As illustrated, the multidecadal variations in the PDO and the NINO3.4 SST anomalies are different. Both vary from positive to negative in the mid-1940s and rise from negative to positive in the late 1970s, but the NINO3.4 SST anomalies have an extra period of positive values in the 1960s. As discussed earlier in this post, the reason the PDO has a different “behavior in time” is because the PDO is also strongly impacted by other factors, including sea level pressure and volcanic eruptions.
In summary, on the main JISAO Pacific Decadal Oscillation (PDO) webpage, the word pattern always refers to the “spatial” characteristics of the North Pacific SST anomalies, not its behavior in time.
THE WORD PATTERN CAN ALSO BE TIME RELATED
The second illustration on the main JISAO Pacific Decadal Oscillation (PDO) webpage is a time-series graph of the PDO data. It was missing from the website as I prepared this post. But there are copies posted at other websites. Refer to Figure 8.
On their PDO Index Monthly Values webpage, JISAO uses the phrase “the pattern of variability” to describe the PDO’s “behavior in time” or the periodicity of the PDO data. Refer to the description of the dataset at the top of the page. It reads (my boldface):
Updated standardized values for the PDO index, derived as the leading PC of monthly SST anomalies in the North Pacific Ocean, poleward of 20N. The monthly mean global average SST anomalies are removed to separate this pattern of variability from any “global warming” signal that may be present in the data.
The uses of the word pattern are different, and their intents are different. Does the use of the word pattern in both instances add to the confusion about the PDO? I don’t have the answer. I’m asking the question. Clearly, the use of pattern in the JISAO description of “The ‘Pacific Decadal Oscillation’ (PDO) is a long-lived El Niño-like pattern” relates to its spatial characteristics. Likewise, the word pattern in the JISAO description of their maps, “Typical wintertime Sea Surface Temperature (colors), Sea Level Pressure (contours) and surface windstress (arrows) anomaly patternsduring warm and cool phases of PDO,” refers to the same thing, the spatial pattern.
Note: I mentioned above that there are no El Niño or La Niña events in the North Pacific north of 20N. I have shown, however, that there are secondary releases of heat in the North Pacific from the warm waters left over from an El Niño. These secondary releases of heat from the North Pacific occur along the Kuroshio-Oyashio Extension (KOE) and they occur during La Niña events that follow major El Niño events. Refer to The ENSO-Related Variations In Kuroshio-Oyashio Extension (KOE) SST Anomalies And Their Impact On Northern Hemisphere Temperatures.
DOES THE PDO DATA EXAGGERATE ITS RELATIVE SIGNIFICANCE?
Figure 9 compares the Pacific Decadal Oscillation (PDO) data and NINO3.4 SST anomalies. The scales are similar and that might lead one who is unaware of the differences between the two datasets to believe the two “signals” are similar in magnitude. That’s wrong for a number of reasons. First, the NINO3.4 SST anomalies represent the SST anomalies of an area in the equatorial Pacific, but the PDO data does not represent the SST anomalies of the North Pacific. The PDO data is a statistically manufactured dataset that represents an abstract form of the SST data there. Second, the NINO3.4 SST anomalies are presented in Deg C. The PDO data is not. The PDO data has been standardized.
Unfortunately, as far as I know, there is no PDO data available online that has not been standardized. So to illustrate the PDO data before standardization, one would have to duplicate the process JISAO uses to create it. Two of the three SST datasets JISAO uses are obsolete and the differences between those older datasets and the current spatially complete datasets are significant, so the results could be very different. And we really do not need to go through all of that trouble to show that the PDO exaggerates the variability of the North Pacific SST anomalies. We can show that other ways.
The inverse relationship between the PDO and the North Pacific Residual was illustrated in Figure 4. And as you’ll recall, the North Pacific Residual was calculated by subtracting Global SST anomalies from North Pacific SST anomalies, north of 20N. Both datasets in Figure 4 were smoothed with a 121-month running-average filter and the PDO data was scaled and inverted (multiplied by -0.2) to bring its variations into line with the North Pacific Residual data. For the next illustration, Figure 10, let’s leave the PDO data in its raw monthly form, and only invert (multiply by -1) the North Pacific Residual data. That is, we won’t scale either dataset. As shown in Figure 10, the actual variations in the North Pacific Residual are miniscule compared to those of the standardized PDO data.
The standard deviation of the North Pacific Residual data is 0.177, so to standardize that dataset we divide it by that value, or multiply it by its reciprocal of 5.65. Refer to Figure 11. Note how well the inverted and standardized North Pacific Residual data (using a different SST dataset, HADISST) captures the underlying multidecadal variations of the PDO data.
And we can detrend the North Pacific SST anomaly data to also show how the PDO exaggerates actual North Pacific Sea Surface Temperature variability. Again in this example, both datasets are in their “raw” form, and the detrended North Pacific SST anomalies are inverted (multiplied by -1.0). The PDO data as shown in Figure 12 greatly exaggerates the actual variations in the detrended North Pacific SST anomalies.
To standardize the detrended North Pacific SST anomalies, we’ll divide the data by its standard deviation (0.182), or multiply it by its reciprocal of 5.5. Once again, as shown in Figure 13, much of the multidecadal variability of the PDO can be captured by inverting and scaling the adjusted North Pacific SST anomalies.
Whether we present the North Pacific SST anomalies detrended or as a residual, the PDO data exaggerates the actual variations in North Pacific SST anomalies by a factor of at least 5.5. This may also lead to some the misunderstandings of the effects of the PDO on global temperatures.
The PDO is a useful index. Based on the IceCap post, it is used by meteorologists for weather predictions. The early papers about the PDO discussed its impact on salmon production, so it is also useful in those endeavors. But the PDO cannot be used to explain epochs of global warming or cooling because the PDO does not represent a process through which the North Pacific could raise or lower global temperatures.
This post also illustrated how the PDO data is inversely related to North Pacific SST anomalies and how the PDO data greatly exaggerates the actual variations in the Sea Surface Temperatures of the North Pacific.
In a follow-up post, we’ll discuss the error-filled SkepticalScience post It’s Pacific Decadal Oscillation.
The PDO data and the HADISST anomalies presented in this post are available through the KNMI Climate Explorer: