by Judith Curry
Update: New comment from Xie
My mind has been blown by a new paper just published in Nature.
Just when I least expected it, after a busy day when I took a few minutes to respond to a query from a journalist about a new paper just published in Nature [link to abstract]:
Recent global warming hiatus tied to equatorial Pacific surface cooling
Yu Kosaka and Shang-Ping Xie
Abstract. Despite the continued increase in atmospheric greenhouse gas concentrations, the annual-mean global temperature has not risen in the twenty-first century, challenging the prevailing view that anthropogenic forcing causes climate warming. Various mechanisms have been proposed for this hiatus in global warming, but their relative importance has not been quantified, hampering observational estimates of climate sensitivity. Here we show that accounting for recent cooling in the eastern equatorial Pacific reconciles climate simulations and observations.We present a novel method of uncovering mechanisms for global temperature change by prescribing, in addition to radiative forcing, the observed history of sea surface temperature over the central to eastern tropical Pacific in a climate model. Although the surface temperature prescription is limited to only 8.2% of the global surface, our model reproduces the annual-mean global temperature remarkably well with correlation coefficient r 50.97 for 1970–2012 (which includes the current hiatus and a period of accelerated global warming). Moreover, our simulation captures major seasonal and regional characteristics of the hiatus, including the intensified Walker circulation, the winter cooling in northwestern North America and the prolonged drought in the southernUSA.Our results show that the current hiatus is part of natural climate variability, tied specifically to aLa-Nina-like decadal cooling. Although similar decadal hiatus events may occur in the future, the multi-decadal warming trend is very likely to continue with greenhouse gas increase.
The authors used the GFDL coupled global climate model. They conducted three simulations:
The historical (HIST) experiment is forced with observed atmospheric composition changes and the solar cycle. In Pacific Ocean–Global Atmosphere (POGA) experiments, SST anomalies in the equatorial eastern Pacific (8.2% of the Earth’s surface) follow the observed evolution (see Methods). In POGA-H, the radiative forcing is identical to HIST, and in the POGA control experiment (POGA-C) it is fixed at the 1990 value [natural internal variability only]. Outside the equatorial eastern Pacific, the atmosphere and ocean are fully coupled and free to evolve.
The results in terms of global-average surface temperature are shown in Fig 1 below:
In Fig 1 a, you can see how well the POGA H global average surface temperature matches the observations particularly since about 1965 (note central Pacific Ocean temperatures have increasing and significant uncertainty prior to 1980).
What is mind blowing is Figure 1b, which gives the POGA C simulations (natural internal variability only). The main ’fingerprint’ of AGW has been the detection of a separation between climate model runs with natural plus anthropogenic forcing, versus natural variability only. The detection of AGW has emerged sometime in the late 1970′s , early 1980′s.
Compare the temperature increase between 1975-1998 (main warming period in the latter part of the 20th century) for both POGA H and POGA C:
- POGA H: 0.68C (natural plus anthropogenic)
- POGA C: 0.4C (natural internal variability only)
I’m not sure how good my eyeball estimates are, and you can pick other start/end dates. But no matter what, I am coming up with natural internal variability associated accounting for significantly MORE than half of the observed warming.
Like I said, my mind is blown. I have long argued that the pause was associated with the climate shift in the Pacific Ocean circulation, characterized by the change to the cool phase of the PDO. I have further argued that if this is the case, then the warming since 1976 was heavily juiced by the warm phase of the PDO. I didn’t know how to quantify this, but I thought that it might account for at least half of the observed warming, and hence my questioning of the IPCC’s highly confident attribution of ‘most’ to AGW.
Although this was not a specific conclusion of the paper (the focused on the period 2002-2012), the conclusion jumps out from their Fig 1 (and my eyeball analysis).
Climate models are notoriously poor at simulating natural internal variability. My recent post Climate model simulations of the AMO provides some insights here. The bottom line is that if I were to pick a single climate model with which to conduct these experiments, I would choose the GFDL model. The interesting thing from a scientific perspective is that specifying the surface temperature in this region seems to anchor the coupled atmosphere/ocean circulations in a way that not only gives a better simulation of global average surface temperature, but also provides better simulations of the variability of key regional circulation features.
How might this anchor work? Well, this suggests a number of additional climate model experiments. But a recent paper by Marcia Wyatt entitled A secularly varying hemispheric climate signal propagation previously detected in instrumental and proxy data not detected in CMIP3 data base. I don’t want to go into this in any detail here since Marcia has a new paper that is almost published, which will be the topic of its own thread. But my idea here is that the network of connections is anchored in a few regions (notably equatorial Pacific, Norwegian Seas, Eurasian Arctic Shelf Seas), and numerically constraining what is going on in these regions can anchor the hemispheric circulations.
Checking in with Uncertain T. Monster
After my initial exuberance of reading this paper, I forced myself to step back and ask what might be wrong with this story? Here is what I came up with:
If you accept the following two premises:
- climate models are useful for untangling natural from anthropogenic climate variability/change
- the missing heat is being sequestered in the deep ocean for the past decade or so
then an inescapable corollary seems to be:
- the same natural internal variability (primarily PDO) that is responsible for the pause is a major and likely dominant cause (at least at the 50% level) of the warming in the last quarter of the 20th century.
Does this explanation rule out contributions to the pause from stratospheric aerosols, solar cooling, etc.? No, but I am not seeing the potential from these forcing mechanisms to dominate over the PDO given the ‘fingerprint’ evidence.
Are climate models useful for untangling anthropogenic climate change and natural internal variability? I have argued on a previous thread that the answer is no, i.e. climate models are not fit for this purpose. However, the POGA simulations, which seem to be successful at simulating natural internal variability, might be a better framework for interpreting AGW detection and attribution.
Apart from the actual results, the significance of this paper is the novel experimental design. Using climate models to understand how the climate system works is arguably the best use of climate models. Here is to hoping that this paper will stimulate some more interesting climate model experiments.
Update: new comment from Xie
Marcel Krok has obtained a comment from Xie: Xie reacts on Curry. Key comments:
Compare the temperature increase between 1975-1998 natural internal variability associated accounting for significantly MORE than half of the observed warming.
1975 was a La Nina year, and 1998 followed the strongest El Nino in the instrumental record. My estimate indicates that the El Nino-La Nina difference accounts for 0.2-0.3 C difference of her 0.4 C in POGA-C. So for multi-decadal trend, PDO accounts for only 0.1-0.2 C for the longer period of 1950-2010. El Nino and La Nina are part of the short climate cycle of ENSO, averaged out over several decades. Our paper noted that the warm phase of the tropical Pacific Decadal Oscillation contributed to the fast warming during the 1970s-1990s.
JC: I thought that it might account for at least half of the observed warming, and hence my questioning of the IPCC’s highly confident attribution of ‘most’ to AGW.
I have a different take on this. The IPCC conclusion applies to centennial warming from 1880. Much of the 0.8 C warming since 1900 is indeed due to anthropogenic forcing, because natural variability like PDO and AMO has been averaged out over this long period of time.
Our results concern the effect of tropical Pacific SST on global mean temperature over the past 15 years. It is large enough to offset the anthropogenic warming for this period, but the effect weakens as the period for trend calculation gets longer simply because it is oscillatory and being averaged out.
JC comments: I’m pleased that Xie has responded to what I have written. El Nino and La Nino don’t seem to me to be easily separable from the PDO. Picking ENSO neutral years at the beginning and end of the 30 yr period circa the 1970′s to 2000 still shows a strong increase during the period.
Also, I am interested that Xie seems to refer to the forthcoming attribution statement for the AR5, which apparently refers to the period since 1880. Which is different from the AR4, which referred to the latter half of the 20th century.