It’s a nonlinear world
There are many complicating factors with dynamical systems containing multiple feedback loops. Most feedback is only manifest after a time delay. An example of this in nature is heavy winter snowfall resulting in higher than normal freshwater runoff in the spring. Different factors may respond on markedly different time scales with the full impact of feedback interactions remaining unclear for years, decades or even centuries.
Another example of delay is the absorption of heat by H2O creating water vapor. Later, this latent heat may be released at a far distant location when the water vapor condenses to form rain. This mechanism is only one of a myriad of ways that energy is transported around the planet. These types of delayed reaction can cause yearly, decadal or longer cycles. Such factors are ubiquitous in nature and require the use of delay differential equations to model them. These delayed responses of natural feedbacks seem to indicate that climate change takes a while to propagate through the Earth system. But we do know, from paleoclimate data, that there have been relatively sudden climate shifts in the past. Past bouts of rapid change is another argument that climate change alarmists use to prove the urgency of their cause.
Climate feedback loops. After Pittock.
One of the favorite scare tactics of global warming promoters is to warn that there are “tipping points” lurking in nature that can suddenly throw the climate into a tizzy. To some extent they are correct, there are historical examples of nonlinear responses from the climate system that have caused reactions far more dramatic than the level of forcing that triggered them. In fact, the mechanisms responsible for the sudden transition from glacial to interglacial conditions could well be labeled a tipping point.
In mathematics, a nonlinear system is one whose output is not proportional to its input. A recent paper reported that dropping sea levels during a glacial period can cause a nonlinear response by exposing the Bearing Strait land bridge between Siberia and Alaska. When this happens the flow of salty water from the northern Pacific into the Arctic basin is cut off, triggering other rapid changes in ocean circulation patterns. The response is a general warming that allows sea levels to rise, restoring the northward flow from the Pacific. A side effect may be to create large lakes within the continental ice sheets containing huge amounts of fresh melt-water.
As I reported in “The Long Road Ahead,” scientists are now quite certain that for Earth’s climate to shift from glacial conditions, huge continental ice sheets must exist in the Norther Hemisphere. The Milankovitch cycles still operate during glacial periods, cycling the climate between cold and relatively warmer conditions. But, once a glacial period has started, orbital variations do not seem sufficient to bring the planet back to a warm, interglacial climate. However, things do warm up enough create instabilities in the ice sheets, leading to pulses of iceberg discharge and allowing the formation of vast ice dammed lakes filled with glacial melt water. These lakes represent a bigger lurking nonlinearity than the Bering Strait land bridge.
Glacial lake Missoula. Painting by Byron Pickering.
Rapid climate change is a result of nonlinearities in the Earth system’s natural processes, and many physical systems are inherently nonlinear in nature. In this case, the freshwater glacial lakes were waiting for just enough forcing warmth to weaken an ice dam and cause it to break. Scientist speculate that an outpouring of freshwater from such glacial lakes can disrupt the flow of ocean currents sufficiently to cause wild swings in climatic conditions.
What causes the warming in the first place is still a mater of raging debate. Some claim it is the melting of mid-latitude glaciers, which are very sensitive to warm summer weather but not cold winters. Others argue that extensive cover of sea ice in the Norther Pacific and Atlantic oceans create more zonal climatic circulation system than exists today. During termination changes in freshwater runoff destabilizes the zonal ocean current patterns, forcing warm water north resulting in a sudden thaw. This science is certainly not settled.
I have heard climate change boosters claim that in the past historical climate shifts have been out of proportion to the forcings that seem to have caused them. This, they say, justifies their assumption that the climate system’s feedback loops are a net amplifier of temperature. In the face of discoveries about the nonlinearities that control glacial-interglacial transitions the argument is not compelling.
Global warming proponents have yet to identify the mechanisms present in Earth’s current environment that will cause a rapid nonlinear climate warming. Today there are no continent spanning ice sheets, precariously holding back glacial mega-lakes. Sea levels are rising marginally, so no chance of exposing a new land bridge that would block ocean currents. We can only conclude that, barring some unforeseen and unidentified future disaster, there will be no rapid global warming.
Other factors
There are also questions regarding the data on which the AGW theory is based. Most climate data are proxy data. This means those data both inexact and open to multiple interpretations. For example, different groups of scientists, working from the same set of tree ring data, arrived at much different interpretations of Holocene climate history (see “Medieval Warm Period Rediscovered“). Furthermore, the farther back in time science looks, the more uncertain the data become.
A good example of the uncertainty surrounding paleoclimate data is the recent announcement that, after nearly 30 years of often heated debate, a team of researchers has finally produced a 50,000 year calibration curve for radiocarbon dating. The basic principles of radiocarbon dating are fairly simple and one would expect that consensus would be easy to achieve among scientists on how to perform such dating. But as with most climate related experimental data there were many data sets that had to be combined. These data sets diverged from each other by up to several thousand years after 26,000 years ago, and researchers could not agree on which ones were most accurate or how to combine them.
For the first 12,000 years the team used thousands of overlapping tree-ring segments from the Northern Hemisphere. For dates older than the available tree-ring record, the researchers had to turn to several other, less-precise data sets, including fossil diatoms and corals. How important can this work be? Surely the differences that the scientists were arguing over were small and any changes from using the new work would merely reinforce previous, if less accurate work. As it turns out, not really.
One case in point, the raw radiocarbon dates for the spectacular paintings of horses, lions, bison, and other animals at Chauvet Cave in southern France, the oldest known cave art, indicate the paintings were made 32,000 years ago. This was right after a major cold spell hit Europe. Using the new calibration curve dates the earliest paintings at Chauvet 36,500 years old, a period of relative warmth. According to Clive Gamble, an archaeologist at the University of London, getting those dates right is critical to understanding questions such as whether humans began painting caves when the climate was colder or warmer. You can imagine how accurate dating could change the interpretation of other climate related data.
Paintings from the Chauvet-Pont-d’Arc Cave in southern France.
Although the new curve is a major improvement in radiocarbon calibration, it is “definitely not the last word,” says team leader Paula Reimer, a geochronologist at Queen’s University Belfast in Northern Ireland. Her team is already planning an update for 2011. Remember, this is for a fairly simple historical measurement. Most historical climate data has known uncertainties with the same order of magnitude as the predictions offered up by the IPCC and other climate change alarmists. The only “good” climate data we have has been collected since the maturation of orbiting remote sensing satellites within the past 20 years or so. You simply cannot accurately predict Earth’s climate a century in the future based on a couple of decades of data.
Finally, this brings us to climate modeling. Some critics claim that the entire skeptical case against anthropogenic global warming is built around the flaws found in GCM climate modeling software—this is not true. Just as the output from climate models does not provide proof of global warming, the use of computer models in climate research does not invalidate that research. Indeed, computer modeling can be an important and useful tool in most any scientific endeavor. Unfortunately, is seems that many climate modelers have committed the greatest sin a modeler can commit—believing that their models and the system being modeled are one in the same.
As most experienced modelers know, all models are wrong but some are useful—climate models, GCM, are no exception. As previously reported, GCM are inherently inaccurate (see “Climate Models Irreducibly Imprecise“). I have also highlighted the errors and sensitivity of climate models to small changes in configuration in “Extinction, Climate Change & Modeling Mayhem.” And this is only scratching the surface. By their very nature computer are simplifications of the real world. Simplifying assumptions must be made to make the models computationally tractable, otherwise not even the largest supercomputers in the world could run them. So all climate models are, by necessity, dumbed down representations of an incomplete theory calibrated and fed with uncertain data.
Even when asked to make near term predictions—essentially extended weather forecasts—climate models are woefully inaccurate. As reported in Geophys. Res. Lett., David Lavers of Princeton University and colleagues tested eight seasonal climate forecast models for their skill at predicting temperature worldwide and precipitation over land masses. The problem, according to this new analysis, is that existing climate models show very little accuracy more than one month out. Even during the first month, predictions are markedly less accurate for the second half than the first. According to the researchers, current models simply cannot account for the chaotic nature of climate. Now extrapolate that inaccuracy to predictions decades, even centuries into the future.
A new article in the Journal of Climate by Knutti et al. discuss some major sources of differences between models,. According to a review of the paper by H. Jesse Smith in Science the work is important “because it is not normally clear which models’ scenarios are likely to be the most realistic, the question arises of which specific models to believe and why.” Knutti et al. themselves state: “there is little agreement on metrics to separate ‘good’ and ‘bad’ models, and there is a concern that model development, evaluation and posterior weighting or ranking are all using the same datasets.” In other words, they are tuning and then judging their models using the same restricted set of uncertain and error prone data that has been used in the past. No wonder their models don’t work.