Arguments
The role of the ocean in tempering global surface warming
Posted on 27 October 2014 by Guest Author
This is a re-post of an article by Richard Allan for NOAA
It is well known that the surface has warmed over the past few decades, primarily in response to rising concentrations of greenhouse gases. ENSO variability and other natural factors, have additionally contributed toward year-to-year fluctuations about this warming trend (dark red line in Figure 1). Strong El Niño events add a few tenths of a degree Celsius to the global average surface temperatures. However, there has recently been an observed slowing in the rate of surface warming (compare the red and orange trend lines in Figure 1) which may be related in part to a greater number of cold La Niña events in the 2000s compared to previous decades (see article by Climate.gov).
Figure 1: Observed changes in global annual average surface temperature relative to 1961-1990 from the HadCRUTv4 dataset which is updated to account for gaps in data coverage (version 2.0 Long Reconstruction). The temperature difference is compared with 1961-1990 average using data from Cowtan & Way (2014). The rate of warming from 1970-2013 (red trend line) is larger than the rate of warming between 1998-2013 (orange line).
On its own, though, ENSO is only part of the story, and it cannot fully explain how and why extra heat trapped by rising greenhouse gas concentrations is unable to raise surface temperatures; recent research indicates that, if anything, the Earth is gaining heat at an increasing rate. How can warming at Earth’s surface have slowed when energy accumulation is becoming larger? The role of our oceans is central in answering this.
Where is the increasing heat coming from?
As greenhouse gas concentrations continue to rise, infra-red radiative cooling by the surface and the atmosphere (1) to space becomes less effective. This sends the planet out of balance, with more energy arriving through absorbed sunlight than leaving through infra-red radiation. Or to put simply, more energy stays on our planet than leaves, which results in the Earth warming. The heating effect is modified by knock-on effects which amplify or reduce climate change through vicious cycles or "feedbacks."
For example, as the atmosphere warms, observations and basic physics agree that the atmospheric moisture content increases. This enhances the strength of the greenhouse effect, and therefore the overall heating effect, still further. Inexorable rises in greenhouse gas concentrations have driven a radiative imbalance leading to global warming at the surface, which has been amplified by the associated increases in atmospheric moisture. These processes are crucial in explaining rapid global climate change during my lifetime.
So where is the recent surface warming?
While ENSO influences Earth's overall heating rate and global surface temperatures from year to year (Figures 1-2), what explains the diminishing rate of surface warming over the longer period from the 1990s to the 2000s? A number of small volcanic eruptions (which make the planet reflect more sunlight) and a slightly weaker sun in the 2000s compared to the late 1990s are thought to have offset some of the heating effect of rising greenhouse gas concentrations. However, our recent analysis of satellite data, ocean measurements and detailed simulations (see Figure 2) indicates that Earth's heating rate has not diminished over the period 1985-2012. And if anything it has increased. Currently heat is accumulating at a rate approximately equivalent to every person worldwide using 20 tea kettles each to continuously boil the oceans. That's a big tea party.
Figure 2: Changes in Earth's yearly average heating rate in observations and simulations 1985-2013 (adapted from Allan et al., 2014). Shown here are an average of 9 simulations which applied observed surface temperatures and were adjusted to match the observed average heating rate over the 2005-2010 period. Earth loses heat following large volcanic eruptions such as Mt. Pinatubo in 1991 which injected aerosol particles into the stratosphere, reflecting more of the incoming sunlight back to space. Also important are changes in ocean circulation associated with ENSO: during the strongest El Niño events, such as 1997/8, warm ocean surface waters heat the atmosphere and more thermal infra-red radiation is emitted to space (there is a lag between the peak of the El Niño and the decrease in the heating rate of the planet) . Conversely, during La Niña events colder conditions in the east Pacific generally lead to a reduced rate at which energy is lost from the global oceans causing a greater overall heating rate of the planet (see above).
As we all know from boiling up a pan of vegetables for dinner, you have to put in quite a bit of energy to raise the temperature of water: approximately 4,000 Joules to heat a kilogram of water by 1 degree Celsius. And it turns out that Earth's capacity to take up heat is primarily determined by our vast oceans, all 1.4 billion billion tons of them (1.4×1021kg) (2). But only the upper layers of the ocean are well mixed over the course of a year or a decade. Exchanging energy between the surface layers and the deep sea (defined here as depths below 700 meters (2,300 feet)) can be a glacially slow processes for much of ocean.
Observations and simulations demonstrate that the upper ocean has warmed up much more than deeper layers (in fact, a recent study indicates that observed warming in the upper 700 meters, shown in Figure 3, has been underestimated in the Southern hemisphere due to the sparse distribution of measurements). Some parts of the deep abyss (down below 4 kilometres, or 2.5 miles) may even be cooling and an evaluation of current satellite data and ocean measurements cannot detect a significant warming and associated contribution to sea level rise from layers deeper than 2000 meters (1.2 miles); many parts of the deep oceans are apparently still blissfully unaware that anything untoward is occurring at the surface!
Figure 3: Top: Depth-averaged temperature trends (in degrees Celsius per decade) in the upper 700 metres of the ocean for 1971-2010 (colours). Bottom: Observed warming rate (in degrees Celsius per decade) in the upper 700 metres of the ocean averaged over lines of latitude for the period 1971-2010 (colours) with average ocean temperature also shown (black contour lines in degrees Celsius) [Source: IPCC (2013) Working Group 1 assessment report, Figure 3.1]. Variability in ocean circulation associated with ENSO generally affects the way heat is mixed in the upper few hundred metres of the ocean (a recent analysis indicates that 85% of the year to year variability in sea level change due to thermal expansion of the upper 2000m of the ocean is explained by the upper 700m depth).
What explains slowing rates of surface warming?
Recent research indicates that fluctuations within the ocean, operating over longer timescales and at greater depth than the circulation changes associated with ENSO, are stealing some of the heat from the surface layers, thereby depriving the atmosphere of the warmth from the ocean surface. Changes in the Pacific and Atlantic oceans have been implicated as prime suspects. Strengthening trade winds in the Pacific, blowing from east to west associated with a La Niña-like enhanced Walker circulation, have been linked to a well-studied decade to decade variability called the Pacific Decadal Oscillation. This appears to be the driving force in burying heat just a few hundred metres below the sea surface.
It has recently been suggested, however, that the unusual conditions in the Pacific have been driven by the Atlantic Ocean as part of a change in the global ocean circulation, causing accelerated heat uptake at deeper levels in the North Atlantic and Southern Oceans. These ocean basins are fully interconnected and recent unusual changes in the tropics also influence weather patterns in remote locations such as across Europe and the Arctic which are connected by “bridges” in the atmosphere. In such a complicated system it can, however, be all too easy to conclude that the tail is wagging the dog.
It has been long appreciated that surface temperatures will not rise smoothly in response to increasing greenhouse gas concentrations; other heating and cooling factors including changes in the sun, volcanic eruptions and particle "aerosol" pollution influence Earth's overall heating rate. And natural fluctuations within our vast oceans from one decade to the next determine what proportion of that heat warms the surface layers and what part is gobbled up by deeper levels. While a lot of the focus on ocean circulation revolves around ENSO, it is important to note there is much more that goes on within the vast oceans that can affect our climate.
A further intriguing question is can climate models realistically simulate ocean variability and heat uptake over a range of time-scales? Plus, it is yet to be established whether the unusual circulation patterns of the atmosphere and ocean are part of a natural fluctuation or whether greenhouse gas forcing has also played a role in generating unusual oceanic conditions. Understanding these processes is vital in gauging the likely rates of global and regional warming associated with human activities, important for how we plan for and adapt to future climatic change. These are important topics of research being tackled by hundreds of scientists worldwide as detailed in the latest comprehensive assessment by the Intergovernmental Panel on Climate Change and as documented by the growing pile of journal articles steadily accumulating... just like heat in the oceans.
FOOTNOTES
(1) In order to stay in energy balance, the earth must emit energy equal to that which is absorbed from the sun. The process in which the Earth’s surface and atmosphere emit energy back to space is referred to as infra-red radiative cooling as our planet is “losing” heat. Greenhouse gases act to reduce the amount of energy that is emitted back to space by Earth, leaving more heat within the atmosphere, land and, in particular, the oceans, resulting in an imbalance as more energy from the sun is absorbed by the Earth than is emitted back to space.
(2) If the heat currently accumulating was distributed evenly throughout the oceans, the temperatures of the entire ocean, including the sea surface, would rise by a paltry 0.017 degrees Celsius each decade. Observations show that the Earth is heating at 0.6 Watts per square metre and since the global surface area is 5.1×1014square metres, the buildup of energy is about 3×1014Joules per second which is 9.5×1022Joules per decade. Making a rough approximation, assuming the specific heat capacity of sea water is about 3,900 Joules per kg per degrees Celsius and the total mass of the oceans is 1.4×1021kg this would mean that it would take 5.5×1024Joules (5.5 trillion trillion Joules) to heat the entire ocean by 1 degrees Celsius (1.8 degrees F). Then we simply divide the heating rate (9.5×1022Joules per decade) by 5.5×1024Joules per degrees Celsius to get 0.017 degrees Celsius per decade (so it would take about 600 years to raise the temperature by 1 degrees Celsius). In reality most of this energy has been heating the upper few hundred metres (see Figure 3 above), explaining why temperatures have been rising at around ten times this rate since 1970.
— Tom Di Liberto, lead reviewer
Looking at figure 1, the orange line looks like cherry picking, by using a short time frame starting with a hot year, 1998. It seems like if you started from 1999 or 2000, the orange line would probably match the red line slope much closer. The red line seems to fit the total data set very nicely, with no significant roll-off obvious in the recent years. But I have seen other charts where the warming rate "pause" seems much more apparent. At least from the figure 1 alone, it seems the orange line is not necessary, and even misleading, like something you would typically show as how "skeptics" view global warming.
rkrolph, that's the point. The orange line is not merely "like something you would typically show as how 'skeptics' view global warming"... that's exactly what it is. The standard 'skeptic' meme of 'no warming for the past 17 years' is constructed by cherry-picking the anomalously hot year of 1998 as the starting point (and looking only at surface temperatures). As you note, starting from 1999 or 2000 (or 1997) yields results more in line with the long term trend. You have successfully demonstrated why this 'skeptic' talking point is wrong from a statistical standpoint.
However, there is also a larger picture to consider. While the 'change in warming rate' for surface temperatures is not statistically significant, we can still identify likely underlying causes for fluctuations like this. That's what this article is really about... how fluxes in the much larger heat reservoir of the oceans can impact surface temperatures on short time scales.
When discussing a chart like the one on the top of the article, I generally do as Tamino did on his blog - count the number of temperature points after 1998 above the full length (red) trendline and compare that to the number of points under it. Statistically there is no decline in warming. I'd rather say there evidence points to the contrary. 2014 will very likely be very close to the red trendline.
@johncl meant this post, for uninitiated.
And this is a different approach to convey the same thing, but addressing 'hasn't warmed since 2002'. Note that both trends are almost aligned. In essence, the pause is an artifact of fast warming during 1992-2006 period.
rkrolph @1, part of the reason for the small differences in slope between the two trend lines is the choice of data set. Richard Allen uses Cowtan and Way (2014) which uses one of two methods to correct for biases in the HadCRUT4 data set that arise due to (among other reasons) limited coverage of Arctic regions where much of recent warming has occurred. "Skeptics" tend to use HadCRUT4 because those very biases accentuate the apparent reduction in the global warming trend. They also like using satellite data sets which, because they show an enhanced responce to ENSO variations, also show an enhanced apparent reduction in the trend (which is yet more evidence that much of the reduction in trend is simply a response to changing ENSO states). Using other data sets which do not have the "no arctic" bias such as GISS or BEST would show similar effects to using Cowtan and Way.
BojanD @4, the " fast warming during 1992-2006 period" is primarilly an artifact of the effects of the Mount Pinatubo erruption. Absent that eruption, there would still be a reduction in recent global warming trends. That, however, is an artifact of the very large El Nino even in 1998, and the very large La Nina events in 2008 and 20011/12. In fact, there is no evidence of the purported "pause" if you only use data sets extending to 2007, showing clearly that the slow down is an artifact of a cherry picked start year along with the two large La Nina's after 2007.
Once again it is proven there is no such a thing as a Climate Change Denier, only Climate Change Liars .... They don't deny, they lie
There is a HUGE difference here folks .....
An interesting observation is that there is a 0.07 degree gap between the red line and the orange line in Figure 1 in 1998. If we were to accept this story that the warming slowed down after 1998, then that gap has to be explained by some physical mechanism. This is much worse for the "global warming stopped in 1998" canard because now a denier has to explain about a 0.2 degree instantaneous warming.
Tom Curtis @6, exactly, you can go in circles forever if you concentrate on short term trends.
Tony @8, interesting point and I actually saw a denier response to this. They feign that 0.2 increase is nothing special since you sometimes see this kind of volatility on per-month basis. They think in trends when that suits them but quickly turn to noise when it doesn't.
The place where SSTs are increasing the fastest are surely in the Arctic Ocean, particularly around the perimeter--places like the East Siberian Arctic Shelf (ESAS). Consequences of such extreme (relatively) heating in these particular places are potentially quite, well, consequential.
"Arctic Methane Spikes Continue — 2666 Parts Per Billion on October 26th"