A closer look at turbulent oceans and greenhouse heating

Adventure = Disaster that didn’t quite happen.

A few such events when sailing Indian Ocean in 1980 and 1984

(i.e. storm front with waterspout, Andrew C. Revkin with Lon Bubeck on sailboat Rashika, near Maldives).

From NYTimes by Andrew C Revkin

Earth’s climate is shaped by the interplay of two complicated and turbulent systems — the atmosphere and oceans.
(The photo above is from the two years I spent at that interface as crew on ocean-roaming sailboats.) The oceans hold the majority of heat in the system, are full of sloshy cycles on time scales from years to decades and, despite an increase in monitoring using sophisticated diving buoys, remain only spottily tracked.

It’s no wonder, then, that assessing the mix of forces shaping short-term wiggles in global and regional atmospheric temperature (years to decades) remains a daunting exercise.
That’s why it’s worth stepping back after weeks of news about studies of the role of oceans in retarding, and sometimes accelerating, global warming to reflect a bit on the difference between edge-pushing analysis and firm scientific conclusions.

What’s firmly established is that the climate is warming, that the buildup of human-generated heat-trapping greenhouse gases is contributing substantially to the warming and that while the buildup of gases is steady, the rise in temperatures is not.

There’s been a burst of worthy research aimed at figuring out what causes the stutter-steps in the process — including the current hiatus/pause/plateau that has generated so much discussion.
The oceans are high on the long list of contributors, given their capacity to absorb heat.
The recent studies have pointed variously to process in the Pacific and Atlantic and Southern oceans (the latter being the extraordinary band of seas in the Southern Hemisphere where winds circulate around the globe unimpeded by continents).

There’s important work to be done on this question but — as the oceanographer Carl Wunsch notes at the end of this post — the paucity of data on ocean heat makes it tough to get beyond “maybe” answers.

Peter Spotts of the Christian Science Monitor wrote a nice piece on the battle of the ocean basins. Here’s his description of the Atlantic mechanism:

[I]n the Atlantic, the heat is carried north as part of a powerful current system known as the Atlantic thermohaline circulation. The north-flowing Gulf Stream is the most visible manifestation of this circulation.

By the time it reaches the far North Atlantic, the dense, salty water has cooled and sinks. It plunges toward the seafloor and heads south at depth, retaining some of the heat it accumulated on the surface.

In a news article in the journal Science, which published the latest paper on the Atlantic’s role in decades-long global temperature fluctuations, Eli Kintisch described the Pacific argument this way: 

[I]n the 17 August Nature Climate Change study, a team led by [Kevin] Trenberth suggests that natural variability in the Pacific explains more than half of the hiatus. Based on data and climate simulations, they argue that a pattern known as the Pacific Decadal Oscillation, which shifts every 20 to 30 years, is driving the increased upwelling as well as other climate trends, including the rapid warming of the Arctic and recent cold winters in Europe.

The newest paper, in the current issue of Science, “Varying planetary heat sink led to global-warming slowdown and acceleration,” argues that the Atlantic not only has shaped the current plateau, but also was responsible for half of the sharp global warming at the end of the 20th century.
The paper, by Xianyao Chen of the Ocean University of China and Ka-Kit Tung of the University of Washington, has a remarkably trenchant abstract:

A vacillating global heat sink at intermediate ocean depths is associated with different climate regimes of surface warming under anthropogenic forcing: The latter part of the 20th century saw rapid global warming as more heat stayed near the surface. In the 21st century, surface warming slowed as more heat moved into deeper oceans. In situ and reanalyzed data are used to trace the pathways of ocean heat uptake. In addition to the shallow La Niña–like patterns in the Pacific that were the previous focus, we found that the slowdown is mainly caused by heat transported to deeper layers in the Atlantic and the Southern oceans, initiated by a recurrent salinity anomaly in the subpolar North Atlantic. Cooling periods associated with the latter deeper heat-sequestration mechanism historically lasted 20 to 35 years.

In an e-mail exchange, Ka-Kit Tung noted how this work can help reveal the steady warming in the background that is attributable to human activities:

The underlying anthropogenic warming trend, even with the zero rate of warming during the current hiatus, is 0.08 C per decade.* [That's 0.08 degrees Celsius, or 0.144 degrees Fahrenheit.] However, the flip side of this is that the anthropogenically forced trend is also 0.08 C per decade during the last two decades of the twentieth century when we backed out the positive contribution from the cycle….

This aspect of the work was largely missed in press coverage.
I asked a range of climate and ocean scientists to weigh in on the paper.
Many focused on details of the Atlantic-Pacific debate.
A few took a broader view that’s worth sharing:

Joshua K. Willis of NASA’s Jet Propulsion Laboratory said this:

In regards to your question, if you mean how robust is the “slowdown” in global surface warming, the answer is it just probably just barely statistically significant. If you are wondering whether is it meaningful in terms of the public discourse about climate change, I would say the answer is no. The basic story of human caused global warming and its coming impacts is still the same: humans are causing it and the future will bring higher sea levels and warmer temperatures, the only questions are: how much and how fast?

As far as the cause of the slowdown, I think there is still some debate, not just about the cause but about the details of what’s going on. For example, there have been several studies including this one to suggest that some deeper layer of the oceans are warming faster now than they were 10 or 15 years ago. This suggestion of an accelerated warming in a deep layer of the ocean has been suggested mostly on the basis of results from reanalyses of different types (that is, numerical simulations of the ocean and atmosphere that are forced to fit observations in some manner). But it is not clear to me, actually, that an accelerated warming of some sub-surface layer of the ocean (at least in the globally-averaged sense) is robustly supported by the data itself.

Until we clear up whether there has been some kind of accelerated warming at depth in the real ocean, I think these results serve as interesting hypotheses about why the rate of surface warming has slowed-down, but we still lack a definitive answer on this topic.

Here’s Andrew Dessler of Texas A&M University:

There are a few interesting things to note here.

First, the hiatus is example of how science works. When it was first observed a few years ago, there were lots of theories — including things like stratospheric water vapor, solar cycles, stratospheric aerosol forcing. After some intense work by of the community, there is general agreement that the main driver is ocean variability. That’s actually quite impressive progress and shows how legitimate uncertainty is handled by the scientific community.

Second, I think it’s important to put the hiatus in context. This is not an existential threat to the mainstream theory of climate. We are not going to find out that, lo and behold, carbon dioxide is not a greenhouse gas and is not causing warming. Rather, I expect that the hiatus will help us understand how ocean variability interacts with the long-term warming that humans are causing. In a few years, as we get to understand this more, skeptics will move on (just like they dropped arguments about the hockey stick and about the surface station record) to their next reason not to believe climate science.

As far as this particular paper goes, I think the findings that the heat is going into the Atlantic and Southern Ocean’s is probably pretty robust. However, I will defer to people like Josh Willis who know the data better than I do.

What’s most exciting to me is that this is really a fascinating conundrum. People like Kevin Trenberth and Kosaka and Xie have published quite convincingly that the action seems to be in the Pacific. So the challenge is to try to resolve that evidence with the ocean heat data that shows that the energy is going into other ocean basins. Ultimately, the challenge come up with the parsimonious theory that fits all of the data.

I do think that ocean variability may have played a role in the lack of warming in the middle of the 20th century, as well as the rapid warming of the 1980s and 1990s. But the argument that the hiatus will last for another decade or two is very weak and I would not put much faith in that. If the cycle has a period of 60-70 years, that means we have one or two cycles of observations. And I don’t think you can much about a cycle with just 1-2 cycles: e.g., what the actual period of the variability is, how regular it is, etc. You really need dozen of cycles to determine what the actual underlying variability looks like. In fact, I don’t think we even know if it IS a cycle.

And this brings up what to me is the real question: how much of the hiatus is pure internal variability and how much is a forced response (from loading the atmosphere with carbon). This paper seems to implicitly take the position that it’s purely internal variability, which I’m not sure is true and might lead to a very different interpretation of the data and estimate of the future.

Thus, their estimate of 1-2 more decades before rapid warming resumes might be right; but, if so, I’d consider them lucky rather than smart.

John Michael Wallace, a professor emeritus of atmospheric sciences at the University of Washington, offered these thoughts:

Back in 2001 I served as a member of the committee that drafted the National Research Council report, “Climate Change Science: An Analysis of Some Key Questions.” The prevailing view at that time, to which I subscribed, was that the signal of human-induced global warming first clearly emerged from the background noise of natural variability starting in the 1970s and that the observed rate of increase from 1975 onward could be expected to continue into the 21st century. The Fourth Assessment Report of the IPCC, released in 2007, offered a similar perspective, both in the text and in the figures in its Summary for Policymakers.

By that time, I was beginning to have misgivings about this interpretation. It seemed to me that the hiatus in the warming, which by then was approaching ten years in length, should not be dismissed as a statistical fluke. It was as legitimate a part of the record as the rapid rises in global-mean temperature in the 1980s and 1990s.

In 2009 Zhaohua Wu contacted me about a paper that he, Norden Huang, and other colleagues were in the process of writing in which they attributed the stair-step behavior in the rate of global warming, including the current hiatus, to Atlantic multidecadal variability. I was initially a bit skeptical, but in time I began to appreciate the merits of their arguments and I became personally involved in the project. The paper (Wu et al.) encountered some tough sledding in the review process, but we persisted and the article finally appeared in Climate Dynamics three years ago. [See Judith Curry's helpful discussion.]

The new paper by Tung and Chen goes much farther than we did in making the case that Atlantic multidecadalvariability needs to be considered in the attribution of climate change. I’m glad to see that it is attracting attention in the scientific community, along with recent papers of Kosaka et al. and Meehl et al. emphasizing the role of ENSO-like variability. I hope this will lead to a broader discussion about the contribution of natural variability to local climate trends and to the statistics of extreme events.

Carl Wunsch, a visiting professor at Harvard and professor emeritus of oceanography at the Massachusetts Institute of Technology, offered a valuable cautionary comment on the range of papers finding oceanic drivers of short-term climate variations.
He began by noting the challenge just in determining average conditions:

Part of the problem is that anyone can take a few measurements, average them, and declare it to be the global or regional value. It’s completely legitimate, but only if you calculate the expected uncertainty and do it in a sensible manner.

The system is noisy. Even if there were no anthropogenic forcing, one expects to see fluctuations including upward and downward trends, plateaus, spikes, etc. It’s the nature of turbulent, nonlinear systems. I’m attaching a record of the height of the Nile — 700-1300 CE. Visually it’s just what one expects. But imagine some priest in the interval from 900-1000, telling the king that the the Nile was obviously going to vanish…

 

Variations in the height of the Nile River over the centuries.

Credit Carl Wunsch

Or pick your own interval. Or look at the central England temperature record or any other long geophysical one. If the science is done right, the calculated uncertainty takes account of this background variation. But none of these papers, Tung, or Trenberth, does that. Overlain on top of this natural behavior is the small, and often shaky, observing systems, both atmosphere and ocean where the shifting places and times and technologies must also produce a change even if none actually occurred. The “hiatus” is likely real, but so what? The fuss is mainly about normal behavior of the climate system.

The central problem of climate science is to ask what you do and say when your data are, by almost any standard, inadequate? If I spend three years analyzing my data, and the only defensible inference is that “the data are inadequate to answer the question,” how do you publish? How do you get your grant renewed? A common answer is to distort the calculation of the uncertainty, or ignore it all together, and proclaim an exciting story that the New York Times will pick up.

A lot of this is somewhat like what goes on in the medical business: Small, poorly controlled studies are used to proclaim the efficacy of some new drug or treatment. How many such stories have been withdrawn years later when enough adequate data became available?

Addendum, 6:30 p.m. | Ka-Kit Tung responded to Wunsch and Dessler in an e-mail.

Here’s his reply to Carl Wunsch’s reaction:

Carl Wunsch’s concern over the sparsity of the ocean data, as expressed in his recent papers, is mostly related to the part of the ocean below 2000 m (the abyssal ocean). He pointed out the signal in the abyssal oceans were mostly at least 500 years old. The signals that we are interested for the current hiatus of the past 15 years came down from above and have not reached the part of the ocean below 2000 m. We used only data above 1500 m and our case was made in Figure 2 of the paper using recent data with better coverage.

And Andrew Dessler’s reaction:

We did not predict in our Science paper that the current hiatus will last another decade or two. The statement that it will last another “15 years” was found in the press release by Science magazine. We were not given a chance to approve it; it probably was not their practice. In the paper itself, we discussed the fact that “historically” lasted 20-35 years. In our university’s press release, we emphasized that it is difficult to predict how long it will last given the changing climate conditions.

Dessler mentioned that there is only 1-2 cycles of this 60-year variability in the short climate record. We discussed this issue in our paper: The global instrumental record since 1850 contains only 2 and half cycles of this 65-year cycle. Tung and Zhou (2013, PNAS) extended it a few hundred years using Central England temperature data. We are currently reexamining Greenland ice-core data that extends the cycle back another thousand years. In addition, free-running models have produced this multidecadal cycles in their control runs (i.e. without anthropogenic forcing), although the latest batch of models have problems getting the period right.

Links :

• The Guardian : Unpacking unpaused global warming – climate models got it right
• Phys : Ocean circulation explains why the Arctic affected by global warming more than the Antarctic
• Washington University : Cause of global warming hiatus found deep in the Atlantic Ocean