Here on the Island of Doubt, climate change is a favored topic for two reasons. First, there is no more important a public policy challenge, no matter what Bjorn Lomborg says. Second, the role of uncertainty in climate science -- and science in general -- has been grossly distorted in many popular media, and this blog was established in part to combat such propaganda. So imagine my reaction to the appearance of a graph in last week's Science that addresses both climate change and uncertainty. At first glance, it may appear a little daunting, but it's worthy thinking about, so bear with me as I try to make sense of it all.
Here's the graph, titled "Minimizing uncertainties" (From "Climate Change: A Changing Climate for Prediction," by Peter Cox and David Stephenson. Science 13 July 2007: Vol. 317. no. 5835, pp. 207 - 208):
The graph is effort to pin just where (or more accurately, when) climatologists are able to best predict what's going to happen with the Earth's climate. The y-axis is "fractional uncertainty" which can be thought of as the amount of doubt. The x-axis needs no explanation
The reddish line, which starts at the top left and falls as we move into the future, represents our grasp of what's going on with the climate for any given starting point in time. Anything between now and 20 years from now, it's basically infinite. In other words, we really don't have enough information about all the various climate forcings (causes of temperature change) and feedbacks (melting ice and permafrost, cloud cover, etc.) The farther into the future we start plotting changes, the better idea we have of what those starting conditions will be, because we'll have more data and better models at our disposal.
The greenish line, which traces an opposite path, beginning at zero on the left and rising into the future, charts predictions of how much greenhouse gases (CO2, methane, nitrous oxides and a few other compounds) we will pump into the atmosphere. The farther into the future we look, the more uncertain are our predictions about the economic and technological conditions involved in generating those emissions.
The bluish line represents how complicated climate processes become as we move into uncharted territory. It's less important for our purposes, set it aside for just a second.
The combination of the red and green lines, represented by the black line, is the net level of uncertainty, or doubt, about just how warm the planet will get. It starts very high in the near-term of the next couple of decades, falls for a few decades and then starts rising again. It's a little higher in total uncertainty than the path traced by the green and red lines thanks to the effect of the blue line, which bumps up the doubt just a tad.
The upshot of this is that our best climate science produces the lowest level of uncertainty about how hot things are getting for the period between 30 and 50 years from now. "Fortunately," as the authors explain, "this is also the time scale over which most longer-term policy and business planning is carried out."
Keep this in mind when reading predictions of the planet's climate. The best we can manage is a level of doubt of 0.6 (or 60 % uncertain) at 40 years out. Any further into the future, and things get very murky indeed.
60% uncertain sounds pretty shaky, but if that's what we've got to work with then fine. I have long argued that the potential negative consequences of climate change are such that we can't afford to wait for the uncertainty to fall further before implementing policies to bring down our emissions.
To my mind there are two things to remember here. First, that the level of uncertainty represents the possibility of both over- and underestimating future temperature rise. Second, science is not about absolute certainty, it's about best guesses. The challenge for policy-makers is to know when the uncertainty has fallen sufficiently for decisions to be made. And most of the scientists involved now say, privately and/or publicly, that the the time for decisions has been upon us for some time now.
Given that the best estimates of climate change seem to fall into the same time frame as conventional long-term business planning, it would seem hard to come up with a good reason for waiting any longer.
Incidentally, the rest of the brief article deals with the negative consequences of this time frame, specifically that the policy-generating process used by the UN doesn't make sense. But that's another, albeit important, story. (Sadly, the full article is subscription-only, but here's the link for those with access.)
- Log in to post comments
More like this
This really bothers me as a scientist. I BELIEVE that global warming is occurring due to human activity. However, such high uncertainty (high p value) puts this in the realm of junk science. If I tried to publish anything in my biological field with such high uncertainty, I would be laughed at and rejected without even getting a review.
I just don't get that red line. It seems to assume we know next to nothing about how the climate system works, but presumes we will know a great deal in coming decades. I don't think climate scientists would say that today we know so little that uncertainty is (currently) greater than the magnitude of the predicted changes. If this had come out 40 years ago then the red line might make sense.
Mike: "If I tried to publish anything in my biological field with such high uncertainty, I would be laughed at and rejected without even getting a review."
Which is well and good assuming the consequence of your being wrong are not disasterous for the planet as a whole and life as we know it.
Since I don't have access to the article, can someone say a bit about the (sorry) uncertainty of that graph or the credibility of the article's authors? In so many ways this is the very crux of the public policy debate so I'd like more of an idea of how to weight these claims of uncertainty levels and how they change over time.
Trinifar: You are exactly right that my work does not have anywhere near the consequences. I have been asked by non scientists about global warming. I tell them I am not in that field. I also tell them that I BELIEVE humans are the cause of global warming. Unfortunately, as this graph shows, I tell them that the scientific proof has a lot of uncertainty. This is partly why it really bothers me that the media says that the scientific community is fully behind humans as the cause of global warming. As a human, I am. As I a scientist, I can not be.
Mike - the uncertainty represented there is not the uncertainty that climate change is due to human activity, it's the uncertainty related to the prediction of specific climatic outcomes. We know that things are going to change, but it's hard to predict exactly how they're going to work out.
Think of it like dropping a water ballon from a tall building - you know it's going to hit the ground, but you can't predict the burst pattern it will make. (Not the best analogy I've ever constructed, but it'll have to do...)
From a risk management perspective, uncertainty is the worst possible scenario. It's like driving blind.
Engineers work with uncertainty all the time, so perhaps we should be looking to them for more guidance. Thinking here of building in California with risk of earthquakes or designing a bridge or building to withstand a certain amount of wind. Certainly (pun intended) we can design our way of life, energy production systems, etc. in the face of uncertainty. As the graph above shows, we aren't dealing with complete uncertainty, and as well all know the consequences of significant climate change are extreme.
As a scientist I am constantly amazed that AGW is posited as "settled science" by those that want to make radical changes to the economy and enrergy policy of the developed world.
The uncertainties surounding climate science are larger than the measured values in most cases. I would never even present a paper for review that had uncertainties of the magnitude regularly bandied about, or more often swept under the rug, by most "climate scientists".
To say that we don't know the future consequences of our current actions but that the dangers are so great that we must make drastic change is illogical at best and foolhardy at worst.
To pretend that such changes are justified by "science" is just dishonest.
"The uncertainties surounding climate science are larger than the measured values in most cases."
So you don't think we should listen to the IPCC and it's recommendations? And you base this on what?
"To say that we don't know the future consequences of our current actions but that the dangers are so great that we must make drastic change is illogical at best and foolhardy at worst."
No one is saying we don't know the future consequences. Uncertainty is not the same as lack of knowledge.
"So you don't think we should listen to the IPCC and it's recommendations? And you base this on what?'
A careful review of the peer reviewed literature.
"No one is saying we don't know the future consequences. Uncertainty is not the same as lack of knowledge."
The people who claim that we face a "dangerously warming" future based on the available data are either self promoting charlatans, like Al Gore, or grossly misinformed.
If you actually read the IPCC report and not just the highly politicized "Statement for policymakers" you will see that there is little "consensus" that we face a "climate catastrophe".
Hansen's mythical "tipping point" is pure fantasy. Actual studies of the Greenland ice sheets show that there is nothing alarming or related to CO2 going on at all.
If you would like a list of peer reviewed articles showing that conditions in the arctic, and antarctic for that matter, are completely consistent with natural variability I would be glad to point you to them.
Yeah right: Oreskes, 2004 (pdf)
Summary
A 2004 article by geologist and historian of science Naomi Oreskes summarized a study of the scientific literature on climate change.[8] The essay concluded that there is a scientific consensus on the reality of anthropogenic climate change. The author analyzed 928 abstracts of papers from refereed scientific journals between 1993 and 2003, listed with the keywords "global climate change". Oreskes divided the abstracts into six categories: explicit endorsement of the consensus position, evaluation of impacts, mitigation proposals, methods, paleoclimate analysis, and rejection of the consensus position. 75% of the abstracts were placed in the first three categories, thus either explicitly or implicitly accepting the consensus view; 25% dealt with methods or paleoclimate, thus taking no position on current anthropogenic climate change; none of the abstracts disagreed with the consensus position, which the author found to be "remarkable". It was also pointed out, "authors evaluating impacts, developing methods, or studying paleoclimatic change might believe that current climate change is natural. However, none of these papers argued that point."
Yeah right.
Yeah right.
The IPCC's current estimate for sea level rise this century is 18 to 59 cm, or probably around 40 cm. This will result in salt water intrusion into aquifers in some areas in addition to other problems.
But note that this projection explicitly assumes no major change in the large polar ice sheets! Why? Because ice sheet dynamics are not known well enough to model. This is not the same as "there is nothing to worry about". What do we know about ice sheets? They form very slowly but may break down rapidly. Melt water forms on top in the warmer months. Some of it flows through cracks or melts it's way down to the bottom of the ice, where it lubricates the sheet's advance. According to NASA climate scientist James Hanson (1):
"The area with summer melt on Greenland has increased from around 450,000 square kilometers when satellite observations began in 1979 to more than 600,000 square kilometers in 2002. Seismometers around the world have detected an increasing number of earthquakes on Greenland near the outlets of major ice streams. The earthquakes are an indication that large pieces of the ice sheet lurch forward and then grind to a halt because of friction with the ground. The number of these "ice quakes" doubled between 1993 and the late 1990's, and in has since doubled again. It is not yet clear whether the quake number is proportional to ice loss, but the rapid increase is cause for concern about the long-term stability of the ice sheet."
1. James Hanson, "Climate Catastrophy" in New Scientist July 28 - August 3 2007.
Pete, thanks for the supporting data!
Lance said: "To say that we don't know the future consequences of our current actions but that the dangers are so great that we must make drastic change is illogical at best and foolhardy at worst."
This statement would be true in many circumstances but I have a problem with it as applied to the GCC issue. The main reason being that even if greenhouse gasses produced by burning fossil fuels is not causing/amplifying global warming, it is adding to the pollution levels that we all live with. Lance's statement seems to imply that we ought to continue with the present policy rather than make any attempt to correct the situation. I'm not buying into that despite the fact that there is uncertainty about our predictions and uncertainty about the results of our actions. But maybe we shouldn't worry about it since in the end we are all going to die anyway. :-(
The graph is being framed as an argument of scientific uncertainty (which is really more than an order of magnitude less than the 10% claimed by the watered-down IPCC report's conclusion), when in fact the graph will be presented to policy-makers to ratchet down their global warming mitigation efforts. The graph itself will be falsely used in the context of expressing Global Warming uncertainty, when it is really expressing uncertainty in policy-makers judgement over whether they want to ensure the survival of the human race. If Americans continue to elect (mildly genocidal) leaders who won't respect science, I'm sure they will respect tariffs imposed by the educated leaders of the rest of the world. Don't forget to tell Cox and Stephenson to shrink the "total uncertainty" portion of their loaded statistics graph to respect this obvious conclusion.