I read Willis Eschenbach’s post last week on Trust and Mistrust where he posed several questions and challenged scientists to respond to the same questions. So, below is my take on these questions. There are a couple points I need to make up front. First, I’m speaking for myself only, not as a representative of the National Snow and Ice Data Center or the University of Colorado. Second, I primarily study sea ice; climate science is a big field and I’m hardly a specialist in the technical details of many climate processes. However, I will provide, as best I can, the current thinking of most scientists working in the various aspects of climate science. Except where explicitly called for, I try to provide only scientific evidence and not my beliefs or personal opinions.
Also, I use the term “climate forcing” throughout. I’m sure this is familiar to most readers, but for clarity: a climate forcing is essentially anything that changes the earth’s global radiation budget (the net amount of radiative energy coming into the earth) and thus “forces” the earth’s climate to change.
Preface Question 1: Do you consider yourself an environmentalist?
Yes. However, I’m no tree-hugger. I don’t believe the environment should be preserved at all costs. I love my creature comforts and I don’t think we can or should ask people to significantly “sacrifice” for the environment. My feeling is that the environment has value and this value needs to be considered in economic and political decisions. In other words, the cost of cutting down a tree in a forest isn’t just the labor and equipment but also the intrinsic value of the tree to provide, among other things: (1) shade/scenery/inspiration for someone talking a walk in the woods, (2) a habitat for creatures living in the forest, (3) a sink for CO2, etc. And I don’t doubt at all that Willis is an environmentalist. However, whether one is an environmentalist or not doesn’t make the scientific evidence more or less valid.
Preface Question 2: What single word would you choose to describe your position on climate science?
Skeptic. This may surprise many people. But any good scientist is a skeptic. We always need to challenge accepted wisdom, we need to continually ask “does this make sense?, does it hold up?, is there another explanation?, is there a better explanation?” – not just of the work of other scientists, but also of our own work. However, a good skeptic also recognizes when there is enough evidence to place confidence in a finding. Almost all new theories have initially been looked upon skeptically by scientists of the time before being accepted – gravity, evolution, plate tectonics, relativity, quantum mechanics, etc.
Question 1. Does the earth have a preferred temperature, which is actively maintained by the climate system?
Willis says that he “believes the answer is yes”. In science “belief” doesn’t have much standing beyond initial hypotheses. Scientists need to look for evidence to support or refute any such initial beliefs. So, does the earth have a preferred temperature? Well, there are certainly some self-regulating mechanisms that can keep temperatures reasonably stable at least over a certain range of climate forcings. However, this question doesn’t seem particularly relevant to the issue of climate change and anthropogenic global warming. The relevant question is: can the earth’s temperature change over a range that could significantly impact modern human society? The evidence shows that the answer to this is yes. Over the course of its history the earth has experienced climatic regimes from the “snowball earth” to a climate where ferns grew near the North Pole. Both of those situations occurred tens or hundreds of millions of years ago; but more recently, the earth has experienced several ice age cycles, and just ~12,000 years ago, the Younger Dryas event led to significant cooling at least in parts of the Northern Hemisphere. So while the earth’s climate may prefer to remain at a certain stable state, it is clear that the earth has responded significantly to changes in climate forcings in the past.
Question 2: Regarding human effects on climate, what is the null hypothesis?
I will agree with Willis here – at one level, the null hypothesis is that any climate changes are natural and without human influence. This isn’t controversial in the climate science community; I think every scientist would agree with this. However, this null hypothesis is fairly narrow in scope. I think there is actually a more fundamental null hypothesis, which I’ll call null hypothesis 2 (NH2): are the factors that controlled earth’s climate in the past the same factors that control it today and will continue to do so into the future? In other words are the processes that have affected climate (i.e., the forcings – the sun, volcanic eruptions, greenhouse gases, etc.) in the past affecting climate today and will they continue to do so in the future? A basic premise of any science with an historical aspect (e.g., geology, evolution, etc.) is that the past is the key to the future.
Question 3: What observations tend to support or reject the null hypothesis?
Let me first address NH2. We have evidence that in the past the sun affected climate. And as expected we see the current climate respond to changes in solar energy. In the past we have evidence that volcanoes affected climate. And as expected we see the climate respond to volcanic eruptions (e.g., Mt. Pinatubo). And in the past we’ve seen climate change with greenhouse gases (GHGs). And as expected we are seeing indications that the climate is being affected by changing concentrations of GHGs, primarily CO2. In fact of the major climate drivers, the one changing most substantially over recent years is the greenhouse gas concentration. So what are the indications that climate is changing in response to forcing today as it has in the past? Here are a few:
1. Increasing concentrations of CO2 and other GHGs in the atmosphere
2. Rising temperatures at and near the surface
3. Cooling temperatures in the stratosphere (An expected effect of CO2-warming, but not other forcings)
4. Rising sea levels
5. Loss of Arctic sea ice, particularly multiyear ice
6. Loss of mass from the Greenland and Antarctic ice sheets
7. Recession of most mountain glaciers around the globe
8. Poleward expansion of plant and animal species
9. Ocean acidification (a result of some of the added CO2 being absorbed by the ocean)
It is possible that latter 8 points are completely unrelated to point 1, but I think one would be hard-pressed to say that the above argues against NH2.
Of course none of the above says anything about human influence, so let’s now move on to Willis’ null hypothesis, call it null hypothesis 1 (NH1). Willis notes that modern temperatures are within historical bounds before any possible human influence and therefore claims there is no “fingerprint” of human effects on climate. This seems to be a reasonable conclusion at first glance. However, because of NH2, one can’t just naively look at temperature ranges. We need to think about the changes in temperatures in light of changes in forcings because NH2 tells us we should expect the climate to respond in a similar way to forcings as it has in the past. So we need to look at what forcings are causing the temperature changes and then determine whether if humans are responsible for any of those forcings. We’re seeing increasing concentrations of CO2 and other GHGs in the atmosphere. We know that humans are causing an increase in atmospheric GHGs through the burning of fossil fuels and other practices (e.g., deforestation) – see Question 6 below for more detail. NH2 tells us that we should expect warming and indeed we do, though there is a lot of short-term variation in climate that can make it difficult to see the long-term trends.
So we’re left with two possibilities:
1. NH2 is no longer valid. The processes that have governed the earth’s climate throughout its history have suddenly starting working in a very different way than in the past.
2. NH1 is no longer valid. Humans are indeed having an effect on climate.
Both of these things may seem difficult to believe. The question I would ask is: which is more unbelievable?
Question 4: Is the globe warming?
Willis calls this a trick question and makes the point that the question is meaningless with a time scale. He is correct of course that time scale is important. For NH2, the timescale is one in which the effects of changing forcings can been seen in the climate signals (i.e., where the “signal” of the forcings stands out against the short-term climate variations). For NH1, the relevant period is when humans began to possibly have a noticeable impact on climate. Basically we’re looking for an overall warming trend over an interval and at time-scales that one would expect to see the influence of anthropogenic GHGs.
Question 5: Are humans responsible for global warming?
Willis and I agree – the evidence indicates that the answer is yes.
Question 6: How are humans affecting the climate?
Willis mentions two things: land use and black carbon. These are indeed two ways humans are affecting climate. He mentions that our understanding of these two forcings is low. This is true. In fact the uncertainties are of the same order of as the possible effects, which make it quite difficult to tell what the ultimate impact on global climate these will have. However, Willis fails to directly mention the one forcing that we actually have good knowledge about and for which the uncertainties are much smaller (relative to the magnitude of the forcing): greenhouse gases (GHGs). This is because GHGs are, along with the sun and volcanoes, a primary component that regulates the earth’s climate on a global scale. It might be worth reviewing a few things:
1. Greenhouse gases warm the planet. This comes out of pretty basic radiative properties of the gases and has been known for well over 100 years.
2. Carbon dioxide is a greenhouse gas. This is has been also been known for well over 100 years. There are other greenhouse gases, e.g., methane, nitrous oxide, ozone, but carbon dioxide is the most widespread and longest-lived in the atmosphere so it is more relevant for long-term climate change.
3. The concentration of CO2 is closely linked with temperature – CO2 and temperature rise or fall largely in concert with each other. This has been observed in ice cores from around the world with some records dating back over 800,000 years. Sometimes the CO2 rise lags the temperature rise, as seems to be the case in some of ice ages, but this simply means that CO2 didn’t initiate the rise (it is clear that solar forcing did) and was a feedback. But regardless, without CO2 you don’t get swings between ice ages and interglacial periods. To paraphrase Richard Alley, a colleague at Penn State: “the climate history of the earth makes no sense unless you consider CO2”.
4. The amount of carbon dioxide (and other GHGs) has been increasing. This has been directly observed for over 50 years now. There is essentially no doubt as to the accuracy of these measurements.
5. The increase in CO2 is due to human emissions. There are two ways we know this. First, we know this simply through accounting – we can estimate how much CO2 is being emitted by our cars, coal plants, etc. and see if matches the observed increase in the atmosphere; indeed it does (after accounting for uptake from the oceans and biomass). Second, the carbon emitted by humans has a distinct chemical signature from natural carbon and we see that it is carbon with that human signature that is increasing and not the natural carbon.
6. Given the above points and NH2, one expects the observed temperature rise is largely due to CO2 and that increasing CO2 concentrations will cause temperatures to continue to rise over the long-term. This was first discussed well over 50 years ago.
If you’re interested in more details, I would recommend the CO2 page here: http://www.aip.org/history/climate/co2.htm, which is a supplement to Spencer Weart’s book, “The Discovery of Global Warming”.
Of course, there are other forcings so we don’t expect an exact match between temperatures and GHGs with a completely steady temperature increase. Periods of relatively cooler temperatures, more sea ice, etc. are still part of the natural variations of the climate system that continue to occur. Such periods may last for months or years. The anthropogenic GHG forcing is in addition to the natural forcings, it doesn’t supersede them. And of course, as with any scientific endeavor, there are uncertainties. We can’t give the precise amount warming one gets from a given amount of CO2 (and other GHGs) with 100% certainty; we make the best estimate we can based on the evidence we have. And that tells us that while there are uncertainties on the effect of GHGs, it is very unlikely the effect is negligible and the global effects are much larger than those of land use changes and soot.
Question 7: How much of the post-1980 temperature change is due to humans?
Here Willis says we get into murky waters and that there is little scientific agreement. And indeed this is true when discussing the factors he’s chosen to focus on: land use and soot. This is because, as mentioned above, the magnitudes of these forcings are small and the uncertainties relatively large. But there is broad scientific agreement that human-emitted CO2 has significantly contributed to the temperature change.
Question 8: Does the evidence from the climate models show that humans are responsible for changes in the climate?
Willis answers by claiming that climate models don’t provide evidence and that evidence is observable and measurable data about the real world. To me evidence is any type of information that helps one draw conclusions about a given question. In legal trials, it is not only hard physical evidence that is admitted, but information such as the state of mind of the defendant, motive, memories of eyewitnesses, etc. Such “evidence” may not have the same veracity as hard physical evidence, such as DNA, but nonetheless it can be useful.
Regardless, let me first say that I’m a data person, so I’ve always been a bit skeptical of models myself. We certainly can’t trust them to provide information with complete confidence. It may surprise some people, but most modelers recognize this. However, note that in my response to question 6 above, I never mention models in discussing the “evidence” for the influence of human-emitted CO2 on climate. So avoiding semantic issues, let me say that climate models are useful (though far from perfect) tools to help us understand the evidence for human and other influence on climate. And as imperfect as they may, they are the best tool we have to predict the future.
Question 9: Are the models capable of projecting climate changes for 100 years?
Based on Willis’ answer to Question 1, I’m surprised at his answer here. If the earth has a preferred temperature, which is actively maintained by the climate system, then it should be quite easy to project climate 100 years into the future. In Question 1, Willis proposed the type of well-behaved system that is well-suited for modeling.
However, Willis claims that such a projection is not possible because climate must be more complex than weather. How can a more complex situation be modeled more easily and accurately than a simpler situation? Let me answer that with a couple more questions:
1. You are given the opportunity to bet on a coin flip. Heads you win a million dollars. Tails you die. You are assured that it is a completely fair and unbiased coin. Would you take the bet? I certainly wouldn’t, as much as it’d be nice to have a million dollars.
2. You are given the opportunity to bet on 10000 coin flips. If heads comes up between 4000 and 6000 times, you win a million dollars. If heads comes up less than 4000 or more than 6000 times, you die. Again, you are assured that the coin is completely fair and unbiased. Would you take this bet? I think I would.
But wait a minute? How is this possible? A single coin flip is far simpler than 10000 coin flips. The answer of course is that what is complex and very uncertain on the small scale can actually be predictable within fairly narrow uncertainty bounds at larger scales. To try to predict the outcome of a single coin flip beyond 50% uncertainty, you would need to model: the initial force of the flip, the precise air conditions (density, etc.), along with a host of other things far too complex to do reasonably because, like the weather, there are many factors and their interactions are too complex. However, none of this information is really needed for the 10000 toss case because the influence of these factors tend to cancel each other out over the 10000 tosses and you’re left with a probabilistic question that is relatively easy to model. In truth, many physical systems are nearly impossible to model on small-scales, but become predictable to acceptable levels at larger scales.
Now of course, weather and climate are different than tossing a coin. Whereas coin flips are governed largely by statistical laws, weather and climate are mostly governed by physical laws. And climate models, as I mentioned above, are far from perfect. The relevant question is whether climate can be predicted at a high enough confidence level to be useful. As mentioned in NH2, we find that climate has largely varied predictably in response to past changes in forcing. This is clearly seen in ice core records that indicate a regular response to the change in solar forcing due to changes in the earth’s orbit (i.e., Milankovitch cycles). If climate were not generally predictable, we would expect the earth’s climate to go off into completely different states with each orbital change. But that doesn’t happen – the earth’s climate responds quite regularly to these cycles. Not perfectly of course – it is a complex system – but close enough that the uncertainties are low enough for us to make reasonable predictions.
It is worth mentioning here that while the general response of climate to forcing is steady and predictable, there is evidence for sudden shifts in climate from one regime to another. This doesn’t invalidate NH2, it merely suggests that there may be thresholds in the climate system that can be crossed where the climate transitions quickly into a new equilibrium. When exactly such a transition may occur is still not well known, which adds uncertainty suggest that impacts could come sooner and be more extreme than models suggest. On the other hand, as Willis mentions there may be stabilizing mechanisms that much such transitions less likely.
Finally, Willis says that climate model results are nothing more than the beliefs and prejudices of the programmers made tangible. But if Willis stands by his answer to Question 1 that the climate stays in preferred states, it should be very easy to create a new climate model, without those biases and prejudices, and show that humans aren’t having a significant effect on climate
Question 10: Are current climate theories capable of explaining the observations?
Willis answers no, but he doesn’t answering the question he poses. He instead discusses the climate sensitivity of to CO2 forcing, i.e., 3.7 Watts per square meters leads to a temperature change between 1.5 C and 4.5 C. These numbers are simply a quantitative estimate of NH2, with an associated uncertainty range. Not being able to narrow that range certainly indicates that we still have more to learn. But it’s important to note that as computing power has increased and as our understanding of the climate has increased over the past several decades that range hasn’t shifted much. It hasn’t gone to up to 6.5-9.5 C or down to -4.5 to -0.5 C. So this is further support for NH2. While perhaps we haven’t been able to narrow things down to the exact house in our neighborhood, we’ve gained increasing confidence that the hypothesis that we’re in the right neighborhood is correct.
But getting back to the question Willis posed. Yes, current climate theories are capable of explaining the observations – if one includes GHGs. Increasing GHGs should result in increasing temperatures and that is what we’ve observed. The match isn’t perfect of course, but nor should it expected to be. In addition to anthropogenic GHG forcing, there are other natural forcings still playing a role and there may things we’re not fully accounting for. For example, Arctic sea ice is declining much faster than most models have projected. Remember, where models are wrong does not necessarily provide comfort – things could ultimately be more extreme than models project (particularly if a threshold is crossed).
Question 11: Is the science settled?
This isn’t a particularly well-posed question, for which Willis is not to blame. What “science” are we talking about? If we’re talking about the exact sensitivity of climate to CO2 (and other GHGs), exactly what will be the temperature rise be in the next 100 years, what will happen to precipitation, what will be the regional and local impacts? Then no, the science is not even close to being settled. But if the question is “is NH2 still valid?”, then yes I would say the science is settled. And as a result, we also can say the science is settled with respect to the question: “have human-emitted GHGs had a discernable effect on climate and can we expect that effect to continue in the future?”
Question 12: Is climate science a physical science?
Willis answers “sort of” and that it is a “very strange science” because he defines climate as the “average of weather over a suitably long period of time” and that “statistics is one of the most important parts of climate science”. Our description of climate does indeed rely on statistics because they are useful tools to capture the processes that are too complex to explicitly examine. This is not unlike a lot of physical sciences, from chemistry to biology to quantum physics, which employ statistical approaches to describe processes that can’t be explicitly measured. But statistics are merely a tool. The guts of climate science are the interactions between elements of the climate system (land, ocean, atmosphere, cryosphere) and their response to forcings. This isn’t really all that different from many physical sciences.
Question 13: Is the current peer-review system inadequate, and if so how can it be improved?
There is always room for improvement and Willis makes some good suggestions in this regard. Speaking only from my experience, the process works reasonably well (though not perfectly), quality papers eventually get published and bad papers that slip through the peer-review process and get published can be addressed by future papers.
Question 14: Regarding climate, what action (if any) should we take at this point?
This is of course an economic and political question, not a scientific question, though the best scientific evidence we have can and should inform the answer. So far there isn’t any scientific evidence that refutes NH2 and we conclude that the processes that influenced climate in the past are doing so today and will continue to do so in the future. From this we conclude that humans are having an impact on climate and that this impact will become more significant in the future as we continue to increase GHGs in the atmosphere. Willis answers no and claims that the risks are too low to apply the precautionary principle. The basis for his answer, in practical terms, is his conclusion that NH2 is no longer valid because while GHGs have been a primary climate forcing throughout earth’s history, they are no longer having an impact. This could of course be true, but to me there doesn’t seem to be much evidence to support this idea. But then again, I’m a skeptic.