The Case For AGW And Why I Am Skeptical
The best real argument for anthropogenic global warming appears to me (and I don’t think I am constructing a straw man here, but feel free to correct me if I am – or if I am missing anything) to be:
1. There is a historical correlation between atmospheric CO2 and temperature. This correlation is probably mostly feedback from temperature change (more CO2 comes out of solution as oceans warm and more goes in as it cools), but…
2. We know that CO2 is a greenhouse gas, so even if we don’t know exactly how much an increase in CO2 will drive temperatures, it is also a forcing factor according to some curve. Now, this curve is certainly not linear and we don’t really know where we are on it, and the majority of correlation being ocean feedback makes it really complicated, but…
3. We have put together computer models that account for everything we know about, and the only way we can get them to fit historical data is to apply certain CO2 forcing formulas. If these formulas are accurate, then the current rise in CO2 that we are seeing will continue to produce greater heat, and eventually this will reach a level of doing more harm than good. And…
4. The only formulas we can figure out that work are based on effects from human generated CO2, so if human beings could just reduce their CO2 output, we might not have to worry about future, potentially harmful temperatures.
* * *
Now, here are the scientific problems I have with this story (the economic issues would require another post):
A. Having a formula that turns CO2 into higher temperatures and successfully maps past temperatures, in no way guarantees that it correctly projects future temperatures.
I know enough math (or could look it up – because it has admittedly been many years) to generate a formula to produce a curve to give a good fit to almost any data set and then have it do whatever I choose after that fit. So, even if I allow for no agenda on the part of the climate modelers, I have to think about their reasons for choosing one formula over another (in this case, a curve that continues linearly, or takes off exponentially, rather than the curve that rises slower and slower logarithmically). Their choice in this regard is based on some hypothesis about how CO2 produces greater temperature, but this hypothesis (there are actually more than one, because different modelers get very different future projections) does not seem to be experimentally testable.
B. There is still a lot of room for doubt that the climate modelers have accounted for every other forcing effect besides CO2 that might have produced recent historical temperature records.
Solar forcing is a strong candidate for being such an unaccounted factor. There are strong historical fits with temperature (even stronger than CO2 correlation on longer time scales) and absolutely no reason to believe that the correlation is feedback rather than forcing – while the mechanism by which earth temperatures can drive atmospheric CO2 is obvious, the mechanism by which earth temperatures could drive sun spots is unimaginable (at least not by me and I have a pretty good imagination).
[Here is a link to a video talking about the most recent science I can find on solar forcing: http://cdsweb.cern.ch/record/1181073/]
C. There is a lot of room for doubting that the modelers have properly accounted for every possible negative feedback effect that might completely mitigate the problem.
It is intuitive that there are such negative feedback mechanisms, or we would probably see many more historical swings from ice-house to hot-house planet and back again. There are only a few in hundreds of millions of years, with our current ice age state exhibiting many many smaller swings that have never broken through the ice age temperature ceiling that we are currently closely approaching. Something happens at that temperature that keep us in an ice age cycle. I am not sure what that is, and I don’t think anyone else is either. But the fact that we have been in an ice age for millions of years, with none of the interglacial periods (like now) breaking a certain temperature level, indicates some downward temperature effect probably kicks in at this level.
This would probably be both some feedback on the shorter time scale of interglacial peaks and a separate downward temperature pressure from whatever slowly sent us into the ice age, as shown by the long term declines of both high and low maximums in the graph below. The later might be the continental drift of Antarctica into a polar position to hold extra ice and act as a reflector.
My favorite candidate for the former shorter-term feedback (unaccounted for in climate models) is cloud cover. Greater heat produces more clouds. More clouds reflect more light. More light reflected causes cooling. Negative feedback.
But that is just a guess – it could be anything or nothing. The IPCC reports I have read basically say [paraphrasing] “We don’t know if there will be more or less clouds or if this would reflect more incoming light or absorb more outgoing heat.” There is a lot of room for uncertainty here.
D. The temperature records upon which the climate models are based, and therefore the degree that CO2 has been implicated in global warming (as the default candidate), may not be correct.
There is an ongoing disagreement between satellite and surface based measurements. The surface based measurements seem to show more recent warming. The argument in favor of the surface based temperature record has been that the satellites do not monitor the poles, and that most of the warming is happening there. But there is no known mechanism (at least not any that I have been able to find an explanation for) by which CO2 increase would mostly heat the poles .
Another issue is that it has recently come to light that there have been some odd choices made in compiling data for the temperature record upon which the models were based. For example, only one site out of dozens was chosen to represent the entire Antarctic continent, and that chosen site showed the greatest warming signal of all stations – an 8x greater warming signal than the raw data from all those stations. This sort of thing could explain the greater polar warming and the land versus satellite data discrepancy.
[Here is a link to an article about temperature data selection in Antarctica:
http://noconsensus.wordpress.com/2009/12/13/ghcn-antarctic-warming-eight-times-actual/]
And if the land based temperature record is bad, then the models built to fit it are all bad.
So, when the modelers say things like, “Only human action can explain the temperature record” they may be absolutely correct – its just may be a different kind of human action than what they are considering…
E. And last – but perhaps most importantly – the projections made in the IPCC 2001 models have not proven accurate.
There may be better models now, and they may still predict immanent warming that may reach dangerous levels, but the projections that have been around long enough to be tested at all, have failed that test. With no increase of the hypothetical problem over 10 years, either the hypothesis is wrong, or it is less of an immediate threat than was predicted.
I find it hard to be convinced that there is an urgent need for immediate political action to stop a problem that moves so slowly as to not show itself for a decade.
We hold science in such high regard because the scientific method has proven to work so well in producing useful new knowledge. But the success of this method relies on experimental confirmation. The less experimentation that can be done, the less you can consider a given conclusion to be based on science.
Consider the number of ways that one can do smaller scale experiments to test an evolutionary hypothesis. Each time you put a shovel in the ground you are engaged in an experiment that might unearth fossils that demonstrate new evidence for or against evolution. Each new fossil that has been discovered after the theory of evolution was published can be considered as new evidence. So even before science advanced to the point of understanding DNA or observing cultures of bacteria react to induced environmental changes, a lot of experimentation was actually going on.
Now consider the very complex hypothesis that the average temperature of the earth will increase to specific high levels over decades due to increased CO2 enhancing the green house effect.
What are the smaller scale experiments we can do to verify this claim? The scope and time frame of this hypothesis make it almost unverifiable by the scientific method. The only thing we can really do, by way of collecting new experimental evidence, is to wait decades and see what happens.
Now, the experimental data collected over the past decade has not supported AGW. Overall temperatures have not risen as expected. Where temperatures have increased, it has not been at the latitudes or altitudes predicted by the hypotheis of CO2 enhanced greenhouse effects.
But I have to be skeptical even about this non-confirming data, because any given experiment can produce anomalous or coincidental results. It is only through significant repetition that we gain scientific confidence, and repeating these experiments (even once) requires another 10 years (or more), its not something that is fast or small scale enough for the repetition of experimental results that gives us scientific confidence.
Although conclusions based on AGW hypothesis have been published by people with degrees in science, they have not arrived at these conclusions via the the scientific method. They have skipped the most important step in that method. So, these conclusions should not be given the level of respect that you usually reserve for scientific conclusions.
No, but what other model of prediction are you going to use? Tea leaves? Sarah Palin twitters? Haruspicy? (Combining the latter two might not be a bad idea, actually). Seriously, looking for a guarantee from scientific findings is never going to happen, and there’s a much lesser degree of certainty in climate science, given all the variables.
B) Solar activity has been thoroughly discredited as a source for the climate change activity we’ve seen over the last 200 years: http://www.independent.co.uk/environment/climate-change/solar-activity-not-the-cause-of-global-warming-456785.html
C) There is room for uncertainty here, I agree. However, studies of Venus have shown the cloud cover there traps heat on the planet’s surface, so this could go either way. If there are other negative feedback effects (hurricanes, etc), I’d be interested in seeing them, but I don’t think any of this relates to human-caused global warming. If there’s a study that has shown significant changes in cloud cover vis-a-vis climate changes, I’d be interested to see that as well.
D) “The argument in favor of the surface based temperature record has been that the satellites do not monitor the poles, and that most of the warming is happening there. But there is no known mechanism (at least not any that I have been able to find an explanation for) by which CO2 increase would mostly heat the poles.”
So a) why would that mean that the data from the poles was irrelevant and b) how is this an argument for using the satellite data? Minus the data from the poles, how closely do the ground and sat readings match? are there any other reasons either one would be more accurate?
E) Define “urgent” in regards to political change. I think we can agree, that’s the subject of some debate as well. However, given the glacier pace at which the widespread reform we would need to stop the level of human activity is perhaps the only thing slower than climate change, I would say it does demand “urgent” attention.
The fact that earlier models didn’t work out is only testament to the size of the problem they’re dealing with.
The one point you seemed to have missed in all of this is the main thrust of the AGW evidence–that of the vast disparity between pre- and post-industrial levels of CO2 and the effect on the climate.
A. Indeed, there is “a much lesser degree of certainty in climate science, given all the variables” (known and potentially unknown), hence my skepticism that anyone can accurately predict global temperature on a decades scale. And the hypothesis for the CO2 forcing curve for global temperature is not even testable on a smaller time scale. If you can’t do an experiment to demonstrate your hypothesis, any conclusion you reach can not be considered scientific.
B. Solar forcing does indeed not perfectly correlate to recorded historical temperatures – but if that is your standard for “thoroughly discrediting” a potential variable from being included in a complete climate model, then you have to throw out CO2 forcing as well – as CO2 has rising significantly from 1940 – 1970 while temperatures declined, and temperatures have failed to rise since the late 1990s while CO2 has continued to rise.
Solar forcing could explain the 1940-1970 decline – if it was overcoming the effect of CO2 increase at that time, because there was less CO2 present. Then from 1980 – 2000 it would be the other way around – maybe higher CO2 levels were then overcoming the expected decline in temps that Solar forcing would have been predicted in the early 1990s. If so, this makes CO2 forcing even more impressive.
But then what has happened 2000 – 2010? temps have stalled. Does Solar forcing finally overcome CO2? Does the CO2 only drive temps so high and then fade off logarithmically? Is there some third forcing factor at work here? Or is none of what we know true, and the climate is a true chaotic system with such extreme sensitivity to initial conditions that the attempted models will never be correct? I don’t know – nor I believe, does anyone else.
You should really follow the link in the article above and watch the video about the experiment they are doing at CERN right now with the brand new super collider, with the intention of demonstrating a proposed mechanism for solar forcing. There are also shorter term solar magnetic field drop outs that happen all the time, allowing experiments to be done when these occur. Some (but not all) of these experiments that have been done have shown regular temperature correlation with about a one week lag for ions created by cosmic rays to attract other aerosol particles until they are large enough to act as a seed for water vapor condensation and increase cloud cover.
Interesting and exciting stuff. Brand new testable climate science being done right now.
C) I believe that when and where clouds form will turn out to be the key factor. They block incoming sunlight but also outgoing infrared. So at night they will raise temps and in the day they will cool. And interestingly, this pattern shows up in the warming signal shown in recent temperature records – greater warming at night than in the day – pointing to cloud cover as a factor in the recent temp rise. Also, land absorbs and emits more IR than Sea, so clouds over sea are likely to cool more in the day and warm less at night. Lots of interesting potential additions to the models – and not present in any form in the climate models I have seen information on (pre-2001), although I bet such things are being included in more recent models.
D) The satellite data and ground station data do track pretty well, but with less sharp or recent warming shown in the satellite data. I have no specific argument for either, I am just pointing out uncertainties in the data upon which models have been based. And depending on how chaotic climate really is on these scales (unknown) bad initial data will have more or less of an effect. There are also some troubling issues with “corrections” of the raw land data, both at time of initial collection and even years later with historical revision of the data. These corrections always seem to be towards greater recent warming.
The Antarctica thing is just one of several odd choices made by those constructing the temperature record that was used in the models. I don’t put this down to conspiracy, but there is a reason why, in most fields of science, data collection is usually done on a double blind basis. Also, raw data is not usually withheld from independent review and comparison to the finally chosen data sets – with people filing FOIA requests to try to get it to independently examine the methodology. Unconscious confirmation bias has long been known to pollute scientific studies when proper safe guards are not in place.
Here is a link that touches on these issues (probably written from a biased viewpoint, as a lot of register articles seem to be anti-AGW, but the facts given look correct, even if the conclusions are open to opinion):
http://www.theregister.co.uk/2008/05/02/a_tale_of_two_thermometers/
E.) One could just as easily use that “snails pace of government” argument as a caution for passing such laws. Harmful laws seem to stick around forever. The only thing slower than getting the government to pass a good law is getting it to admit a mistake and repeal a bad one.
If you are implying that somehow failure of the hypothesis implies that it is more likely to be true, then that is just silly. But I agree that it is indeed a hard problem – lots of room for uncertainty – lots of room for doubt.
Not missed at all – that is the basis for the AGW argument I described at the top. But historical CO2 levels and historical temperatures are the initial observations upon which the hypothesis was formed – they are not evidence. New data must be gathered (preferably in a controlled experiment) to see if it matches the hypothesis. The new data is what we call experimental evidence. The data used in the formulation of the hypothesis can not also be used as evidence to confirm the same hypothesis.