by Judith Curry
Lennart Bengtsson’s recent statement on climate research has elicited a response from Andy Lacis, that directly points to the fundamental debate in climate dynamics.
In his statement discussed in the previous post, Bengtsson stated:
Climate is nothing but the sum of all weather events during some representative period of time. The length of this period cannot be strictly specified, but ought to encompass at least 100 years. Nonetheless, for practical purposes meteorologists have used 30 years. For this reason alone it can be hard to determine whether the climate is changing or not, as data series that are both long enough and homogenous are often lacking. Because of chaos theory it is practically impossible to make climate forecasts, since weather cannot be predicted more than one or several weeks. For this reason, climate calculations are uncertain even if all model equations would be perfect.
In a comment on the Bengtsson post, Andy Lacis states:
But fundamentally, that statement is flat wrong. This is because climate is a boundary value problem in physics, while weather is an initial value problem. The physical nature of these two problems is quite different, so also is the numerical approach that has to be taken in order to model climate change, and to forecast the changing weather.
Andy Lacis’ perspective is more extensively discussed in his Climate Etc. post discussing CO2 as a Control Knob.
Context
Before digging into the arguments, this is a reminder that appeal to authority arguments won’t work here (well they never carry much weight at Climate Etc) – both Bengtsson and Lacis are leading senior scientists that are very highly regarded in the climate science community. They come from different perspectives: Lacis’ perspective is from radiative transfer, whereas Bengtsson’s perspective is from atmospheric/fluid dynamics. I do not regard either as an issue advocate.
This same debate is what motivated Tomas Milanovic’s recent post How Simple is Simple – a response to Isaac Held’s article Simplicity Amidst Complexity.
This same debate is at the heart of the controversy surrounding Michael Mann’s recent paper about the AMO, as discussed in a recent post by Nic Lewis.
I have alluded to this debate on several previous threads, notably Trends, change points, and hypotheses. Excerpt:
Consider the following three hypotheses that explain 20th century climate variability and change, with implied future projections:
I. IPCC AGW hypothesis: 20th century climate variability/change is explained by external forcing, with natural internal variability providing high frequency ‘noise’. In the latter half of the 20th century, this external forcing has been dominated by anthropogenic gases and aerosols. The implications for temperature change in the 21st century is 0.2C per decade until 2050. Challenges: convincing explanations of the warming 1910-1940, explaining the flat trend between mid 1940′s and mid 1970′s, explaining the flat trend for the past 15 years.
II. Multi-decadal oscillations plus trend hypothesis: 20th century climate variability/change is explained by the large multidecadal oscillations (e.g NAO, PDO, AMO) with a superimposed trend of external forcing (AGW warming). The implications for temperature change in the 21st century is relatively constant temperatures for the next several decades, or possible cooling associated with solar. Challenges: separating forced from unforced changes in the observed time series, lack of predictability of the multidecadal oscillations.
III: Climate shifts hypothesis: 20th century climate variability/change is explained by synchronized chaos arising from nonlinear oscillations of the coupled ocean/atmosphere system plus external forcing (e.g. Tsonis, Douglass). The most recent shift occurred 2001/2002, characterized by flattening temperatures and more frequent LaNina’s. The implications for the next several decades are that the current trend will continue until the next climate shift, at some unknown point in the future. External forcing (AGW, solar) will have more or less impact on trends depending on the regime, but how external forcing materializes in terms of surface temperature in the context of spatiotemporal chaos is not known. Note: hypothesis III is consistent with Sneyers’ arguments re change-point analysis. Challenges: figuring out the timing (and characteristics) of the next climate shift.
There are other hypotheses, but these three seem to cover most of the territory. The three hypotheses are not independent, but emphasize to varying degrees natural internal variability vs external forcing, and an interpretation of natural variability that is oscillatory versus phase locked shifts. Hypothesis I derives from the 1D energy balance, thermodynamic view of the climate system, whereas Hypothesis III derives from a nonlinear dynamical system characterized by spatiotemporal chaos. Hypothesis II derives from climate diagnostics and data analysis.
The stadium wave falls between II and III.
This disagreement is further clarified by two recent comments from Climate Etc. regulars Robert Ellison (Generalissimo Skippy) and Fred Moolten, excerpts:
Robert Ellison: ‘Sensitive dependence and structural instability are humbling twin properties for chaotic dynamical systems, indicating limits about which kinds of questions are theoretically answerable. They echo other famous limitations on scientist’s expectations, namely the undecidability of some propositions within axiomatic mathematical systems (Gödel’s theorem) and the uncomputability of some algorithms due to excessive size of the calculation.’ James McWilliams
Climate and weather model share the same underlying mathematical dynamic. So models are undoubtedly chaotic and there are many feasible and divergent solutions within the bounds of feasible inputs.
‘The Earth’s climate system is highly nonlinear: inputs and outputs are not proportional, change is often episodic and abrupt, rather than slow and gradual, and multiple equilibria are the norm.’ http://www.globalcarbonproject.org/global/pdf/pep/Rial2004.NonlinearitiesCC.pdf
‘Technically, an abrupt climate change occurs when the climate system is forced to cross some threshold, triggering a transition to a new state at a rate determined by the climate system itself and faster than the cause. Chaotic processes in the climate system may allow the cause of such an abrupt climate change to be undetectably small.’ http://www.nap.edu/openbook.php?record_id=10136&page=14
Climate is what emerges from these abrupt transitions – and there are likely to be four or more this century – counting the 1998/2001 transition to a cooler planet. e.g. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-12-00626.1
Fred Moolten. I noticed exchanges of comments related to the roles of volcanic and solar forcing during pre-industrial times, including the LIA. Here’s one link concluding that solar forcing was the more important cooling influence, because unlike volcanic forcing, surface and stratospheric effects operated in the same direction – Solar and Volcanic Forcing. My purpose, though, is not to compare their respective roles but to point out that in combination, these forcings induced profound and long-lasting climate responses. The relevance to this blog topic is that neither volcanic nor solar variation during that era was periodic – i.e., there was no evidence to support phase-locking with a putative chaotic oscillation, as could be claimed for seasonal or diurnal variation. The observational record therefore appears consistent with evidence for simplicity (quasi-linearity) in response to strong forcings that dominates over the unpredictable variability inherent in chaotic behavior of the climate that may have operated at that time. Again, I believe Tomas is correct in emphasizing the limits to predictability, but Held is correct in reminding us that those limits don’t preclude our ability, with fairly high confidence, to make fairly accurate long term predictions under Holocene climate conditions.
Timescales
It seems that this debate largely boils down to the issue of time scales, and whether climate change can be regarded as truly just a boundary value problem (Hypothesis I). In some sort of long timescale equilibrium sense, external forcing from the sun and atmospheric composition determines the planets overall temperature; we know from comparative planetology that we can explain the broad differences in the climates of Earth, Mars and Venus by simple energy balance considerations. These considerations, however, don’t prima facie tell us anything about the amount or rate of warming to be expected by adding more CO2 to the Earth’s atmosphere.
On a previous control knob thread, Jim2 made the following statement: “CO2 is more like a pilot light than a control knob. It is a non-condensible gas that keeps a small flame alive to ignite the prima donna: Water.” I think that Jim2 has made an insightful statement.
So, how long is ‘long’? The 16+ year hiatus in global warming provides evidence that a 30 ppm increase of atmospheric CO2 (since 1998 – 25% of the post industrial anthropogenic contribution) has not acted to significantly increase global surface temperatures on a timescale of 16 years (the period 1998-present). This implies that changes in atmospheric CO2 of this magnitude is not a control knob on surface temperatures on timescales shorter than 16 years (whether or not the deep ocean is somehow heating because of increased CO2). This does not in any way mean that CO2 does not act to warm the surface on long time scales; it highlights the importance of initializing the ocean in order to make credible predictions of climate on decadal timescales.
Multidecadal ocean oscillations (e.g. PDO, AMO, stadium wave) on nominal timescales of 50-80 years seem to be important, if not dominant, climatic features on multidecadal timescales, even in the presence of CO2 forcing of the magnitude we have seen for the past 100 years.
So climate is not simply a boundary value problem on timescales out to at least 16 years, and almost certainly not out to 30 years, which is the period that is typically used to define ‘climate’. A climate model estimate of equilibrium sensitivity requires a very long model integration; fully equilibrating ocean temperatures requires integrations of thousands of model years. At what point does the climate transition from an initial value problem to a boundary value problem (for a secular, slow change in the boundary forcing)? At timescales less than 30 years, it seems that natural variability dominates over CO2 forcing (which has interesting implications for the attribution of warming in the last half of the 20th century). At timescales of 100-1000 years I suspect that external forcing is increasingly dominant, but there is internal variability on these timescales also (which is less well understood than the 50-80 year variability.) So I am not sure you ever get away from the initial value aspect of climate on the timescales of interest (decades out to a millennia).
Bottom line: what both Bengtsson and Lacis say are not incorrect; the key disagreement seems to be the timescale issue.
Policy relevance
The climate dynamics debate between hypotheses I, II, and III (and the variants) is intellectually interesting and exciting. The IPCC’s hypothesis (I) is being challenged in light of the hiatus in surface temperature since 1998, and scientists are paying increasing attention to natural variability. Scientists that are exploring natural climate variability do not say that there is no effect on climate from the anthropogenic increase in CO2 and pollution aerosol; rather they are exploring the increasingly likely possibility that natural variability is an important if not dominant factor in climate variability on multi-decadal timescales.
Even if we knew exactly the equilibrium climate sensitivity to CO2 doubling, the issue of ‘when’ the warming would be realized in terms of surface temperature is then another major uncertainty. If the policy relevant period is the 21st century (and increasingly only out to 2050), do climate scientists really have anything useful to say about the evolution of 21st century climate? Will it be dominated by anthropogenic forcing (greenhouse gases and pollution aerosol (I)? Or will 21st century climate be dominated by multi-decadal ocean oscillations, solar and volcanic forcing? In my opinion, the climate variability/change of the 21st century is definitely an initial value problem, with the attractor potentially changing in a significant way with external forcing.
Regardless of the cause of climate variability, we don’t really yet have a good handle on regional vulnerability to climate variability (both hot and cold, extreme events). Even if it was somehow convincingly demonstrated that a certain temperature threshold was ‘dangerous’ given our current vulnerabilities, we don’t know when we might actually encounter that particular threshold. If the threshold is at least a half century in the future, we have no idea what future regional vulnerabilities will be or what technologies might be available. If we are 100% convinced that a warming of say 3C will occur say by 2300 from anthropogenic CO2, does a near term (i.e. now) policy response make economic and political sense? These are clearly not issues for science to resolve, but the scientific debate, particularly regarding the time scale, is relevant to the policy debate. Attempts to argue that ‘dangerous climate change’ is already here depends on dubious links to anthropogenic climate change of extreme weather and rapid sea level rise in some locations (tied primarily to local geological processes and land use practices).
It would be great for science, and even for policy, if climate scientists would stop focusing solely on CO2-forced climate change, and look at natural internal variability and how this interacts with external forcing – both slow (e.g. solar or CO2) and fast (e.g. volcanoes). The big issue of scientific interest (not to mention policy relevance) is abrupt climate change. I don’t really regard the climate shifts of 1976 and 2001 to be ‘abrupt climate change’, although by some definitions these qualify. These climate shifts are certainly of great interest and significance in and of themselves, and there are some hypotheses on the table regarding how to predict the next shift (e.g. stadium wave).
JC conclusions
The science of climate change on decadal to century timescales most definitely is not settled, in spite of the IPCC’s highly confident proclamations. There are so many interesting and unsolved issues in climate dynamics. At this point, climate science seems relatively irrelevant for energy policies – the goals of carbon mitigation are in place, and whether anything meaningful can be achieved in the near term is doubtful. However, climate scientists are (in the words of Pointman) in a hurry towards some finishing line only they could see, and acted accordingly. I suspect that the IPCC becoming less and less relevant to the UNFCCC agenda.
I’m hoping that at some point soon, climate scientists will get fed up with trying to play politics with their science and get back to researching and debating these fundamentally interesting and unsolved issues in the science of climate dynamics, rather than attacking their colleagues for suggesting that there are other ways of thinking about climate change.