In this post, we’ll discuss a recent article and blog post about the recently published England et al. (2014). This post includes portions of past posts and a number of new discussions and illustrations.
We’ve already discussed (post here) the paper England et al. (2014) Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Since then, NBC News has an article by John Roach with the curious title Global Warming Pause? The Answer Is Blowin’ Into the Wind. And the team from RealClimate have agreed and disagreed with England et al. (2014) in their post Going with the wind.
I find it surprising that England et al. is getting so much attention. It’s simply another paper that shows quite plainly that the past and current generations of climate models are fatally flawed…because they cannot simulate coupled ocean atmosphere processes that cause global surface temperatures to warm and that stop that warming. Maybe the attention results from their use of “wind” as a metric. Everyone understands the word wind.
A FEW PRELIMINARY COMMENTS
We’ve illustrated and discussed in past posts how the current generation of global models cannot simulate how, when and where the surfaces of the oceans have warmed since 1880 and during the satellite era. See the posts:
- CMIP5 Model-Data Comparison: Satellite-Era Sea Surface Temperature Anomalies
- Models Fail: Land versus Sea Surface Warming Rates
- IPCC Still Delusional about Carbon Dioxide
We’ve also illustrated this recently, but as a reminder: The sea surface temperature anomalies of the tropical Pacific are a part of this discussion, because that’s where El Niño and La Niña events take place, and because that’s where the trade winds in question blow. The satellite-enhanced sea surface temperature data for the tropical Pacific show that the surface of the tropical Pacific has not warmed over the past 32+ years—the full term of the Reynolds OI.v2 sea surface temperature data. See Figure 1. On the other hand, climate models indicate that, if the surface temperatures of the tropical Pacific were warmed by manmade greenhouse gases, they should have warmed more than 0.6 deg C (or about 1.1 deg F).
So the problems with climate models are not limited to the past decade and a half.
OVERVIEW OF ENGLAND ET AL. (2014)
England et al. (2014) are basically claiming that stronger trade winds in recent years are driving CO2-based global warming into the depths of the Pacific Ocean, and that the stronger trade winds are associated with a shift in the frequency, magnitude and duration of El Niño and La Niña events. They use an abstract metric called the Interdecadal Pacific Oscillation (IPO) to define the periods when El Niño or La Niña events dominated.
As an expanded overview of England et al., during the period from the mid-1970s to the late 1990s, the trade winds were weaker because El Niño events dominated, so, according to their modeling efforts, more global warming was occurring at the surface. But since the late 1990s, the stronger trade winds associated with more-frequent La Niñas are causing the CO2-based global warming to be driven into the depths of the Pacific Ocean.
Figure 2 presents a commonly used index for the strength, frequency and duration of El Niño and La Niña events. It is a graph of the sea surface temperature anomalies of the NINO3.4 region. I’ve also highlighted NOAA’s official El Niño and La Niña events, based on their Oceanic NINO Index (but the data in the graph are not from the Oceanic NINO Index). And as we can see, there were a series of strong and long El Niño events from 1982 through 1998: the 1982/83, the 1986/87/88 and the 1997/98 El Niños. Although the series of El Niños in the first half of the 1990s are now considered independent events, Trenberth and Hoar proclaimed them as one long event in their 1996 paper The 1990-1995 El Niño-Southern Oscillation Event: Longest on record. The El Niño events since 1998 have not been as strong, and the frequency of La Niña events has increased.
Because trade winds are weak during El Niños and strong during La Niñas, the change in the frequencies of El Niño and La Niña events indicate the trade wind should have increased during that time…and they have. We illustrated and discussed this in the recent post El Niño and La Niña Basics: Introduction to the Pacific Trade Winds.
But that’s not where the problems exist with the findings of England et al. (2014).
THE BASIC PROBLEMS WITH ENGLAND ET AL. (2014)
England et al. (2014) have the same problems as the recent Trenberth papers. I discussed those in my Open Letter to Kevin Trenberth – NCAR. The following is a revised portion of that post. I’ve changed a few of the graphs to reflect the differences in the start date for the hiatus. Trenberth used 1999 in one of his recent papers, while England et al. used 2001.
Based on England et al (2014), the ocean heat content of the western tropical Pacific should be increasing during the hiatus period. As noted earlier, England et al. used 2001 as the start of the hiatus. Figure 3 presents the NODC ocean heat content for the western tropical Pacific (24S-24N, 120E-180), for the depths of 0-700 meters, for the period of January 2001 to December 2013. We can see that the western tropical Pacific to depths of 700 meters has, in fact, warmed.
Before we proceed, let’s confirm that the variability in the ocean heat content of the tropical Pacific takes place in the top 700 meters. The Tropical Atmosphere-Ocean (TAO) project buoys have sampled subsurface temperatures, etc., in the tropical Pacific since the early 1990s, so the NODC data should be a reasonably reliable there. Over the past decade, ARGO floats have supplemented the TAO buoys. And now for the data: the source Ocean Heat Content data in the tropical Pacific for the depths of 0-700 meters and 0-2000 meters (represented by the unadjusted UKMO EN3 data) during the TAO project and ARGO eras are exactly the same, see Figure 4, and that suggests that all of the variability in the tropical Pacific ocean heat content is taking place in the top 700 meters.
Back to our discussion of the hiatus period: The NODC ocean heat content data also show the ocean heat content (0-700m) of the eastern tropical Pacific, a much larger region, has been cooling from 2001 to 2013. See Figure 5.
As a result, there has been an overall decrease (not increase) in the ocean heat content of the tropical Pacific since 2001, Figure 6, and a substantial decrease in the ocean heat content of the tropical Pacific as a whole since the peak around 2004.
Therefore, based on data, there appears to have been a rearrangement of heat within the tropical Pacific and not an addition of heat as suggested by England et al. (2014).
Also, in the recent post If Manmade Greenhouse Gases Are Responsible for the Warming of the Global Oceans… I presented the NODC’s vertical mean temperature anomaly data for the Indian, Pacific, North Atlantic and South Atlantic Oceans, for the depths of 0-2000 meters, during the ARGO era (starting in 2003). Figure 7 is an update of that illustration, including the recently released 2013 data. The flatness of the Pacific trend indicates there has not been a substantial increase in the subsurface temperatures of the entire Pacific Ocean to depths of 2000 meters over the past 11 years…same with the North Atlantic. It cannot be claimed that manmade greenhouse gases caused the warming in the South Atlantic and Indian Oceans, when they obviously have had no impact on the warming of the North Atlantic and Pacific Oceans to 2000 meters over the past 11 years.
The problems with the England et al. (2014) model-based assumptions are blatantly obvious. The ocean heat content of the tropical Pacific has cooled quite rapidly since 2001. And subsurface temperatures of the entire Pacific Ocean during the ARGO-era show little to no warming.
The body of the RealClimate post by Matthew England is a summary of the England et al. (2014) paper, and we outlined the failings of the paper above. Eric Steig wrote the introduction for the RealClimate blog post. For support, Eric linked a few papers:
- Foster and Rahmstorf (2009) – We discussed the failings with Foster and Rahmstorf (2009) in the post here, and with the follow-up Rahmstorf et al. (2012) in the post here. And as you’ll recall, even SkepticalScience threw Foster and Rahmstorf (2009) under the bus.
- Balmaseda et al. (2013) – Balmaseda et al. (2013) was a primary topic of discussion in the “Trenberth still searching for the missing heat” series of posts: here and here and here and here and here and here and here and here and here.
- Cowtan and Way (2013) – We discussed how the Cowtan and Way (2013) infilling of HADCRUT4 global land+ocean surface temperature data did nothing to explain the hiatus over 90% of the globe–but exaggerated the model failings at the poles–in the posts here and here.
THE NBC NEWS ARTICLE ABOUT ENGLAND ET AL. (2014)
John Roach begins his article with (my boldface):
For the past 13 years, global surface air temperatures have hardly budged higher despite continual pumping of planet-warming gasses into the atmosphere from the engines of modern life. Does this prove global warming is a giant hoax? No, according to a new study, which says the missing heat is being blown into the western Pacific Ocean by extraordinarily powerful and accelerating trade winds.
The reference to “global surface air temperatures” is curious. I suspect John Roach relied on England et al. (2014) for it. The abstract of England et al. (2014) begins (my boldface):
Despite ongoing increases in atmospheric greenhouse gases, the Earth’s global average surface air temperature has remained more or less steady since 2001.
And the first sentence of the paper reads (my boldface):
Observations of global average surface air temperature (SAT) show an unequivocal warming over the twentieth century1, however the overall trend has been interrupted by periods of weak warming or even cooling (Fig. 1).
Yet England et al. did not present “global surface air temperature” data in cell a of their Figure 1. See my Figure 8.
Figure 8 (cell a of Figure 1 from England et al.)
England et al. (2014) presented GISS Land-Ocean Temperature Index (LOTI) data in their Figure 1, which is a combination of land surface air temperature data and sea surface temperature data, with the vast majority being sea surface temperature data since the oceans cover 70% of the planet.
To confirm that, under the heading of Methods, England et al. write [my brackets]:
Observations and reanalysis data. SAT [surface air temperature] is taken from the Goddard Institute for Space Studies (GISS) climatology.
And no, I’m not being pedantic. As we’ve shown in numerous posts over the past year, climate models do a reasonable job of simulating land surface air temperatures over the past 30+ years, but in order to achieve that warming, the climate models have to double the observed warming rate of the surface of the oceans. See Figure 9.
And using marine air temperature data do not help the models, either—see Figure 10—though it has to be kept in mind that the ICOADS marine air temperature data are not corrected for the shipboard “heat island effect” that plagues that dataset.
John Roach began his discussion of climate models with a catchy heading:
The shortcomings of the climate models highlighted in this new paper feed into larger criticism that the models play down the importance of natural variability in the global climate system. “You want to have enough noise in your system” in order to get a realistic result, noted Xie.
That this shortfall is highlighted in the new research, he added, “is quite a nice result, but in a sense it is bad news for the climate research community because it does point to a potential problem for the climate models.”
It’s not a potential problem. It’s a major problem. One contributing factor to the problem is that climate scientists (example Shang-Ping Xie’s quote) view coupled ocean-atmosphere processes as “noise in your system”. ENSO is not noise; ENSO is a coupled ocean-atmosphere process that climate models still cannot simulate. Sea surface temperature data and ocean heat content data indicate that ENSO acts as a chaotic, sunlight-fueled, coupled ocean-atmosphere, recharge-discharge oscillator—with El Niño events acting as the discharge mode, and with La Niña events acting as the recharge and redistribution mode. If this topic is new to you, refer to illustrated essay “The Manmade Global Warming Challenge” [42MB pdf] for an introduction.
Climate models failings with respect to ENSO—their failures to properly simulate of El Niño- and La Niña-related processes—have been known for years. See Guilyardi et al. (2009) and Bellenger et al (2012). It is very difficult to find a portion—any portion—of El Niño and La Niña processes that the models simulate properly.
Then John Roach allowed Matthew England and others some more leeway:
A problem with the models, in turn, could erode trust in climate science, noted England. But “that would be akin to writing off the medical profession for finding out something new about an illness that they didn’t know about earlier,” he said.
The inability of the models to capture the observed wind trends and thus the hiatus is “just one small process in the global system that seems to need improvement,” he noted. The long-term global warming trend, he added, is independent from decade-to-decade variability in the Pacific Ocean.
Fyfe echoed the sentiment. Instead of undermining climate science, he said, “What you are seeing here in this discussion is the natural evolution of science and improving our understanding. The overall big picture that the planet is warming and that that warming is due to human influence stills stands with or without the hiatus.”
For those who understand climate model failings, the trust in climate science has been eroding for years. In fact, for many persons, it has eroded to the point that we have no confidence in climate models…none at all.
This sentence is a classic: “The long-term global warming trend…is independent from decade-to-decade variability in the Pacific Ocean.” And the claim “just one small process in the global system that seems to need improvement” is the understatement of the year. Combined they form the most bizarre assertions I’ve seen attributed to a climate scientist to date…
- …especially when the paper that England authored indicated the lack of global surface warming has been caused in part by the “decade-to-decade variability in the Pacific Ocean”. In other words, without that “decade-to-decade variability in the Pacific Ocean” there would not have been the hiatus. (His paper also failed to address the contribution to the long-term warming from the mid-1970s to the turn of the century caused by the domination of El Niño events during that period.)
- …especially when England et al. (2014) presented multidecadal changes in surface temperatures in response to multidecadal “variability in the Pacific Ocean”, not “decade-to-decade variability”. (See their Figure 1, which is my Figure 8.)
- …especially when one considers that the Pacific is the largest ocean on this planet, that it covers more of the surface of the planet than all of the land masses combined, and that its surface area dwarfs the area of the other ocean basins. See Figure 11.
- …especially when one considers than the monthly, annual and decadal variations global sea surface temperatures mimic the variations in the Pacific sea surface temperatures…because the Pacific is so massive and because the dominant coupled ocean-atmosphere processes that express themselves as El Niño and La Niña events take place in the Pacific. See Figure 12.
- …especially when one considers that the multidecadal variations in the sea surface temperatures of the global oceans mimic the variations in the Pacific…again because the Pacific is so massive and because the dominant coupled ocean-atmosphere processes that express themselves as El Niño and La Niña events take place in the Pacific. See Figure 13, which presents the two datasets detrended and smoothed with 121-month running-average filters.
- …especially when one considers that the forced component of the climate models (represented by the multi-model mean) cannot simulate the multidecadal variations in the sea surface temperatures of the Pacific Ocean (Figure 14) or the global oceans (Figure 15), again represented by detrended and smoothed data and model outputs.
# # # # # #
- …and, last but not least, especially when everyone understands that climate models were tuned to (and model projections extend from) a naturally occurring upswing in global sea surface temperatures, not the long-term trend. See Figure 16. (For the years used for model tuning, refer to Mauritsen, et al. (2012) Tuning the Climate of a Global Model [paywalled]. A preprint edition is here.)
Imagine how foolish the models would look if the modelers had tuned their models to the warming period from the early-1910s to the mid-1940s.
[Note: If you’re wondering why the climate models performed so poorly in Figures 14 and 15, refer to the post IPCC Still Delusional about Carbon Dioxide. The climate model simulations of sea surface temperatures do not capture the cooling that took place from 1880 to the early-1910s (see Figure 16 above) and, consequently, they do not capture the warming that took place from the early-1910s to the early-1940s.]
And once again, we find climate science being compared to medicine. But let’s put the climate model failings into perspective. The failures of the climate models to properly simulate coupled ocean-atmosphere processes are akin to doctors not being able to explain respiration and circulation. Climate models are in the dark ages compared to medicine.
The data and model outputs presented in this post are available from the KNMI Climate Explorer.