He is an influential climatologist and senior scientist at the National Center for Atmospheric Research. Wigley was part of a team whose analyses of the earth's average surface temperature, and how it has risen, has become the data most cited by climate experts. He also was a lead author of the landmark 1995 Intergovernmental Panel on Climate Change (IPCC) report which said that human activity is a likely cause of the warming of the earth's atmosphere. In this interview he summarizes the evidence pointing to the human factor in global warming, addresses skeptics' criticisms about this evidence, and explains how societies are vulnerable to even relatively small changes in climate.

It's an interesting distinction, really, because climate models are basically weather models that concentrate on different time scales. The standard difference between weather and climate is that climate is, in simple terms, the average weather. So climate is what we can expect to happen for a particular season or a particular year, but it's not going to tell us what's going to happen day by day.

It's interesting that sometimes people confuse. They think that a climate forecast is going to tell us what's going to happen in the year 1999 on January 3, or something like that. But, of course, we can't. We can only tell people what the average expectation is going to be. And even there, there's a lot of uncertainty.

So a weather forecast would talk about things that might happen over the next day or two. But climate would be something like a more average pattern over a longer period of time, like a season?

Or longer than a season, really. Just to give you an example, one of the concerns about future climate change, particularly in Europe, is changes in storminess. That doesn't mean we're going to be [able] to say, you know, there will definitely be more storms, they will be more intense, and they'll happen this time of the year, or they'll last for this certain period of time. But what we can say is that there may be, on average, more storms.

So if we looked at the average weather situation in the winter over a period of ten years, there would be more storms. That doesn't mean there are going to be more storms in any individual year. It means that the long-term prediction, the average prediction over a decade or more, is that there would be more storms. I'm not saying actually that there will be more storms, but that happens to be an area where there's a considerable amount of uncertainty. But that's just an example of the difference between weather and climate.

You've been talking about within a decade, from decade to decade. And we get bigger patterns going over longer periods of time as well? Century to century? Millennium to millennium?

Yes. Climate does affect all time scales, all time scales beyond the time scale of one year to another. An interesting example of the short end of the time-scale hierarchy is El Niño. And that's a quasi-oscillation of weather conditions that happens on a three- to eight-year time scale. And that means that, roughly, every three to eight years, it might be wetter in one region of the world than normal, and then at the other end of the cycle, it might be drier than normal.

The sorts of time scale that we're interested in for human influences on climate are longer than that. It may be that human influences will affect the way El Niño operates, but the primary concern is what's going to happen on time scales of 10, 20, 50, 100, or maybe many hundreds of years.

The issue at stake here is, before we decide whether or not humans are driving the climate into an abnormal state, we need, as scientists, to know what is a normal state in order to figure out whether we can detect a signal of a human influence which is, I guess you'd call it, abnormal, since it hasn't really happened before.

Yes. We do need to know the background against which the human influences might be superimposed. But in terms of identifying a human influence, the most important thing is not so much what the background is, but how much the climate normally varies from year to year and decade to decade.

If that variability were very large, then we would need a very big human-induced change in order to identify the effect of human activities. If that variability was zero, then any small change would be easily identifiable. So, really, it's the background variability or noise that is the concern. And this is, in technical jargon, referred to as a signal-to-noise problem. We need to know what the signal is, what the changes wrought by human activities might be--and we use complicated climate models to do that--but we also need to know what the noise is against which this signal is appearing.

Let's talk about some of the sources of data we have to determine this issue. When people talk about the instrumental record, what do they mean by that?

The instrumental record spans, roughly, the last few centuries. And the instruments that we're referring to here are instruments that measure temperature, instruments that measure precipitation or pressure at the earth's surface, sometimes instruments that measure temperature above the earth's surface. The first thermometers were invented many, many centuries ago. Apocryphally at least, Galileo was one of the first people to take temperature measurements and to develop a practical way of measuring temperature. Rainfall measurements go back even further than Galileo's thermometers. The earliest rainfall records, I think, go back to some very crude measurements made in Korea around about the fourteenth century or so. Still, quantitative measurements.

So the critical thing about instrumental climatology is that these are numbers. These are not just descriptions of what the climate or the weather was like, but they're actual firm numbers measured by scientific instruments.

But t's one thing to say that one person might have measured something at the end of the sixteenth and early seventeenth century, but you need to know: Was it done accurately? Was it done on a sufficient number of sites of the globe? Measured at the same time of day? There's all kinds of problems about using that data.

There are certainly many problems associated with using old instrumental records. An interesting example is the Galilean thermometer, because those thermometers have been preserved, or some of the thermometers he used have been preserved. And, in that case, it's possible to use them to measure temperatures now, and calibrate them or compare them with modern thermometers and see what the differences are, and see how accurate they are. And if there are differences, then those can be applied as corrections to the early records. We don't always have the luxury of being able to do that.

How accurate are they? They're not very accurate, are they?

Those thermometers are certainly accurate to better than one degree Celsius. The changes that we're talking about due to human activities, however, are substantially less than one degree Celsius. Or at least the changes that have occurred so far--may be half a degree Celsius. The future changes may be much larger. But you can always then ask the question: Well, if thermometers in the past were only accurate to one degree, how can we possibly identify a half a degree change in the temperature of the globe that people claim has occurred over the last 100 years? And the reason why we can do that is although individual thermometers in the past may not be terribly accurate, if there are enough of them, and if their inaccuracies are random, by averaging them all together, we can reduce that error substantially and get a fairly precise measure of temperatures over a large area, maybe of a substantial fraction of the globe.

If you could just address some of the criticism of the surface record that have been made...take the issue of sample bias. What do people mean by "sample bias"? Is this because it's easier to measure on land than sea?

Certainly there are a lot of problems with using temperature data from different sources, different countries, over land or over ocean, and so on. In fact, temperatures over the ocean, or the temperatures that they used in trying to get a global picture of temperature changes, are not air temperatures but usually sea surface temperatures, temperatures measured by ships, by dipping some sort of bucket into the ocean and then bringing the bucket up on deck and then letting it sit for a while, and measuring the temperature of that water. So there are all sorts of problems with different instrumentation, different methods of measuring temperature, different ways of exposing thermometers. The meteorological instrumentation exposure used in the nineteenth century was very different from what is used today.

There's another problem associated with temperature measurements, and that is urbanization. As cities have grown, so their own activity causes a so-called urban heat island, so temperatures in an urban area would be expected to increase with time, simply because of this urban heat island effect.

Is that because of the concrete?

Well, it's partly because human activities--vehicles, heating, industrial activity and so on--actually releases heat. And so there is an actual heat source that forms up a city or an urban environment, and that can spill over into the surrounding countryside, you know, maybe for tens of kilometers. Of course, another issue is that the concrete of roads and buildings and parking lots is able to absorb more heat from the sun's rays, and then give that heat off to the atmosphere in the urban environment and cause additional heating. So, there are a lot of reasons urban areas are warmer than outlying country areas.

When you add these up, there are a lot of problems. Are you satisfied that it's been possible to take the errors or control for the [errors] in these measurements, to get this number?

A fantastic amount of work has gone into trying to correct for all of these so-called biases, these instrumentation, exposure, and so on, effects. Part of the battle here is knowing, you know, what the problems are. And I think the problems are pretty easy to identify. Then the next step is to try and correct for those sources of uncertainty or sources of bias. And one can do that.

Just to give you an example: In measuring temperatures at the sea's surface or the upper layers of the ocean, many years ago people threw a particular type of bucket into the ocean and measured the temperature in that. It was small, not really like a household bucket, but a special instrument that was used to house the thermometer for measuring the temperature of the sea surface. And, over the years, the construction methods for those so-called buckets changed. And as they changed, bias crept into the measurement of sea surface temperatures, associated with the different shape and size and construction of the bucket. Well, in that particular case, we can actually look at the old buckets and measure that bias, [and] compare [it] with modern methods for measuring temperature in the ocean environment. We can also make little mathematical models of the buckets and determine how rapidly they cool if they're set on the deck and exposed to the wind. And we can apply both those empirical and theoretical correction methods, and we can compare them to see that they're consistent. We can apply those to the early instruments to make them compatible with modern instruments.

When you do all this and you end up with this curve that's sometimes called the Jones-Wigley curve, or whatever, what does it tell us? This is an average global temperature since about the late nineteenth century?

That record goes back to around about 1860 or so. And it is a record that combines temperatures over the oceans and temperatures over the land. What it shows most strikingly is that the temperatures during this decade are about six-tenths of a degree warmer than they were at the end of last century, or during the period 1860 to 1900.

So that's a bit more than half a degree Centigrade rise over 100 years.

That's right.

It doesn't seem much to people. So I might say that in the place where I live, Massachusetts, the temperature would vary ninety degrees maybe over the year, or vary a lot within a day. So half a degree Centigrade over a century doesn't sound like a lot, because my perception is of local weather. Explain to me why that should be cause for concern, or why that should be seen as a significant number.

The primary reason that's a cause for concern is that it seems to be outside the range of natural variability. So it's pretty clear that human activities are partly to blame for that rise--relatively small rise. Okay. So, how do you put that into a realistic context? Well, one way to do that is to look at records over past thousands or tens of thousands or millions of years, and see how much the climate of the globe or the global mean temperature changed on that very long time scale.

If we go back 20,000 years, a fair fraction of the world in the Arctic regions was covered by huge ice masses. That was the last glacial period. The temperature during that last glacial period was about four or five degrees Celsius less than today. And yet the environment was just radically different. Not that we're expecting such massive cooling to occur in the future. Quite the contrary. We expect warming of that order of magnitude to occur over the next few hundred years. If the difference between the Ice Age and the present was so large in terms of the physical environment, the vegetation, the amount of ice, the areas where people could live, the amount of rainfall, and so on, if there were such large differences between 20,000 years ago and now, and we anticipate similar differences--but in a different direction, the opposite direction--might occur over the next few hundred years, then I think that is cause for concern.

So it's not so much just the temperature changes, but it's the changes in all the other aspects of the environment: amounts of precipitation, the ability for vegetation to maintain its status quo, the amount of water that's available for agriculture and for water resources, and so on.

If we upped it from the half a degree to five or six degrees positive. Right? That's the logic?

Well, I thinks if we have a few degrees' warming, then that's half the amount of temperature change that occurred over the last 20,000 years, but it's going to occur in a few hundred years or even 100 years. So there's also an acceleration of the amount of warming, compared with what has occurred in the geological and historical past.

Now, a couple of things that sometimes skeptics say, I'd just like to get on record here. One of the things they say is that of this half a degree, a good fraction of it occurred earlier in the century, before 1945, and then followed by a sort of cooling period. And so it's difficult to make the argument that greenhouse gases are really responsible for all of this, or particularly the first part.

Well, it may be difficult for those skeptics to be able to make that association, but it's certainly not difficult for scientists who work in the area.

Anybody who works on trying to understand just, say, the temperature changes over the last 100 years is not going to attribute those temperature changes to a single causal factor. In fact, there are really three categories of factor that one has to consider.

One is the influence of human activities---the buildup of carbon dioxide in the atmosphere, and other greenhouse gases--and also the effect of sulfate aerosols. Aerosols are very small droplets, primarily made of sulfuric acid or ammonium sulfate, that are suspended in the air, and they arise from fossil fuel combustion, which produces sulfur dioxide. Sulfur dioxide is oxidized to these sulfates, and that produces these very small droplet particles in the atmosphere. And those particles have a cooling effect. They have a regionally specific cooling effect because they don't stay around in the atmosphere for more than a few days to a week. But they're continually replenished in areas where there's industrial activity.

So, essentially, from the human-influence point of view, we have two opposing effects. One is the fairly globally uniform effect of greenhouse gases--particularly carbon dioxide--and then counterbalancing that, but with an entirely different spatial pattern, the effect of sulfate aerosols. And the irony is that both the carbon dioxide and the source for the sulfate aerosols is fossil fuel burning. Okay. So that's just one side of the question.

The next important issue is, what about natural external factors that might cause temperature to change on a global basis? There are really two primary factors. One is the sun, and if the output of the sun were to change--and we do believe that it does change. In fact, we know from satellite measurements that the output of the sun changes on a decadal time scale. But we also suspect that the sun's output changes on times scales of 30 years to 100 years or more. Okay. So we have to account for that solar influence.

And then the other external forcing factor, or external influence, is the effect of volcanic eruptions. And a prime example there was Mt. Pinatubo, which erupted not so long ago and caused, for a year or two, a substantial amount of cooling.

Now, in addition to those two factors that we refer to as external forcing--driving mechanisms that are outside the climate system--the climate system itself can vary from year to year and decade to decade, as the amount of heat in the atmosphere is transferred to ice masses or to the ocean, and there's continual interchange between ocean atmosphere and the land surface and ice masses and so on.

These are kind of internal cycles knocking around?

Well, they're not cycles in the sense of having any strict periodicity, although El Niño is an example. I mean, that does have a reasonable periodicity. But there are other interactions that are going on, purely internally, within the climate system that could cause, for example, the globe to warm by one- or two-tenths of a degree, or cool by one- or two-tenths of a degree, over a 100-year period. So, what you have to do in unraveling the reasons for global warming is, you have to account for all of these factors. And at different times, different factors have more or less importance. So if I go back now and look at that record of global warming, it's true that over the period from about 1910 to 1940, there was very substantial warming, so much so that it cannot have been due only to human activities.

But there are two other possibilities. It could be due to changes in the output of the sun. And, in fact, we believe that this is the primary reason for that warming.

That it was reduced, you're saying?

The output of the sun during that period increased in the early twentieth century, and that added to a small amount of warming due to human activities. If we go then to the period from 1940 to 1970, when there wasn't much increase in global mean temperature, and then you say, "Well, well, at that time the emissions of carbon dioxide were going up very rapidly, so why didn't the world warm?" Well, part of the reason is that the output of the sun declined slightly during that period, so we believe, and also that the emissions of sulfur dioxide increased dramatically. So the amount, or number, of these little droplets increased, and they had a cooling effect that partially offset the warming effect that would have occurred just due to greenhouse gases.

And then since 1970, there's been a really dramatic warming, you know, more rapid than has ever occurred. And we believe that this is partly because the emissions of sulfur dioxide have been controlled in certain parts of the world, so that cooling effect has diminished slightly. There's tremendous growth in emissions of greenhouse gases, particularly carbon dioxide, over the last twenty years or so. And, interestingly, the output of the sun has increased a little bit over that period, too. So all those things together cause a very rapid warming.

So, basically, if you look at that whole record of global mean temperature changes over 100 years or so, then account for, or try to account for that, using all of these external factors, you can account for it tremendously well. In fact, the agreement between model expectations and the observed record of warming is spectacularly good. It's so good that it's clearly a bit of a fluke that it came out so well.

But one criticism could be that you could always, with so many factors--especially ones you don't have a good handle on, like the sun and the sulfate aerosols--you could really just cook the books. You can make it come out right.

Cooking the books would be a concern if that was what scientists were apt to do. But I don't think scientists generally operate that way. So, in explaining this past temperature record, the way it has been done is first to take independent estimates of all of the contributing factors, and put them together, and then run a climate model with an independent estimate of how sensitive that climate model is to external forcing, carbon dioxide, or whatever. So that all of the input information is essentially given to the climate modeler, and then you run the climate model, and then it comes out with this amazingly good agreement.

Now, after doing that, it's important to go back and say: Now, what are the uncertainties associated with these calculations? You can say, for example, that we don't really know what the cooling effect of sulfate aerosols is. We have a best-guess number, but we have a range of uncertainty. So, you consider and concatenate all of the upper and lower bounds of these uncertainties, and that gives you a range of possible temperature changes due to these factors. Well, the observed temperature change is right in the middle. But of course, you know, it's possible that all of the uncertainties acted in the same direction, and that the model calculation is overestimating the amount of the human influence, or the amount of the solar influence, or whatever.

But, of course, it's equally likely that it's in the other direction as well. So we, you know, we try to play the game as straight as possible, and then also to look at the range of uncertainties.

But some of the uncertainties, like the indirect effect of aerosols, are really quite large. And it's difficult to imagine how you're really going to narrow them, isn't it? Some of these questions are very difficult ones to settle experimentally.

Well, I'd like to quantify the uncertainty. I mentioned that the amount of observed global mean warming is roughly six-tenths of a degree Celsius over the last 100 years. If we put all of our best estimates of the parameters into a climate-model calculation, then the amount of warming is pretty close to six-tenths of a degree. The range of uncertainty in this model calculation is about plus or minus two-tenths of a degree. So that means...what the model says is that the warming due to all of the factors--and that includes the solar term, by the way--is, say, between four-tenths of a degree Celsius and eight-tenths of a degree Celsius. Now, part of that uncertainty is associated with the cooling effect of sulfate aerosols. And it's clearly important to try and reduce that particular uncertainty, because that will give us a better handle on how well the model and the observations agree.

But reducing many of these uncertainties is a really formidable task. Okay. So I'll go back to sulfate aerosols. The way sulfate aerosols have a cooling effect is very complicated. In fact, there are really three different components of that cooling.

The first is that in the clear sky atmosphere, in the absence of clouds, these little droplets reflect incoming solar radiation. And that obviously has a cooling effect. If less solar radiation gets to the surface, then the surface is going to be cooler. The second factor--which is, we believe, the most important factor--is what these aerosols do to clouds. They're sufficiently small that they can act as nuclei on which cloud droplets can condense. And that means that if you put more of these sulfate aerosols into the atmosphere, then clouds on average should have more droplets. If they have the same total amount of water, but it's distributed over a larger number of droplets, then the average diameter of a cloud droplet will be less.

Well, it turns out that when clouds have very small droplets, they're much more reflective than clouds that have big droplets. And you can actually see that when you go outside, because clouds...on a stormy day there are white clouds and gray clouds, or even black clouds, and it's usually the darker clouds that are about to rain. And the reason they're about to rain is they have bigger droplets. The white clouds are the clouds that have smaller droplets, and they reflect more sunlight. So we know that that's empirically correct.

The difficulty is in quantifying how much the reflectivity of clouds will be increased as cloud droplets become smaller. And in order to do that better, we need to have measurements of droplets in clouds--that's kind of a difficult thing to do--measurements of the sizes of these little aerosols, estimates of the physics of how the cloud radiation properties change, how the re--reflectivity changes and [so on]. So it's a tremendously difficult scientific problem to understand the processes that are going on, quantify them, and reduce the uncertainties involved.

Now, just to make things even messier: Not only do clouds become more reflective when there are more aerosols around, but also their residence times change. Quite clearly, if a cloud has a lot more small droplets, then it's not going to rain as readily. So if it doesn't rain as readily, it's going to stick around for longer. So what we say is that with a lot of aerosols in the atmosphere, the lifetime of a typical cloud will be longer. It'll stick around for longer. That means it can actually reflect incoming solar radiation for a longer period of time.

So, somehow we've got to put all of these factors together. It's a horrendous theoretical modeling computer calculation.

You were saying earlier that the six-tenths of a degree Centigrade is significant because it shows that something's happening at a faster rate than would be normal. But there are people who talk specifically about certain regions in the past, evidence of periods in, say, northern Europe that were called the Little Ice Age, or earlier, the Medieval optimum, when it was very cold or very warm, seemingly by more than this...stories which say that Greenland was colonized in the twelfth thru thirteenth century, then became unlivable by the fifteenth. So, if those things could happen before there was really any CO₂, isn't the natural variation bigger than what you have said we should be alarmed about?

These past changes in climate that we believe occurred, and which seem to have an influence on human populations, are certainly important. It's possible to look at those and say: Well, there was a lot of variability in the past. And how do we know that the present changes are not just another manifestation of similar types of natural variability? I'll come back to that in just a second. But you know, what's more important is that those past records show really how vulnerable societies are to relatively small changes in climate.

Now, let's get back to this issue of natural variability and those past records. What we have for the twentieth century is an estimate of how the whole global mean temperature, has changed. The past records are all from specific regions. For example, from Greenland or Iceland or parts of Scandinavia, and so on. We don't have a true global picture of how temperatures changed in the past. We have fragmentary records. We have records that are indirect and quite difficult to quantify in many cases. So it's a rather uncertain record, and I think it would be a very brave scientist who claimed that they knew what the past record of global mean temperatures was more than 100 years ago, or what the past variability in that record was.

You know, we have interesting evidence that suggests that the climate system is quite variable, and we take account of that evidence in trying to evaluate and assess the changes that have occurred in the twentieth century. And our judgment is that measured against that yardstick of rather imperfectly known past changes, the present changes are substantially greater than have occurred in human experience.

The global changes?

The global changes.

Here's another argument I've heard: If regional fluctuations can occur in the absence of a mean global change as significant as this--say, for instance, in Greenland or Mill Creek, in Iowa--if those things could occur naturally, does increasing the global temperature increase or decrease the regional variance? And make this more alarming or less alarming?

There's certainly evidence of societies being affected by changes in climate in the past,. And not just changes in weather from one year to another, but changes in climate that appear to have been prolonged. And there are some classic examples. I think when you actually delve deeply into these examples, you'll find that there were many factors other than just climate change that caused those societies to either just disappear, move, or change rather dramatically.

Now, a very good example is the settlements in Greenland. And they apparently died out quite dramatically, and people have suggested that that was associated with a change in climate. Now, when you actually look at the climate records, it's quite difficult to quantify those records with enough precision to associate cause and effect. And if you look at the archaeological remains, you can see that there are many other factors that could have contributed to the demise of those little outposts of civilization. Now, it might be that climate was the trigger. It might be that, in the past, those societies were vulnerable to relatively small changes in climate, and you know, that something came along and that was the end. But the societies themselves were vulnerable. I think that those past societies were much more vulnerable than most modern societies, certainly, you know, societies in Europe and North America and the developed countries. We would be able to stand those natural changes, I think, reasonably well.

So the concern is not so much the natural changes. I mean, that's the backdrop against which human-induced changes are going to occur. But the human-induced changes that are expected over the next 100 years are much, much greater than any changes that societies experienced in the past. Much greater.

So, given that we know these greenhouse gases have some radiative properties, how do we know that actually (1) it's increasing, (2) that the source for this increase is anthropogenic, and (3) by how much has it increased? Just give me those facts, the data.

We know, without a shadow of doubt, that the concentrations of the important greenhouse gases have increased dramatically over the last few hundred years. Those gases are carbon dioxide, methane, nitrous oxide, and other gases like the halocarbons or the so-called CFC's. The CFC's are interesting because, for most of those, they didn't exist before human beings came around. They didn't exist until the 1930's. And now they exist in considerable abundance, so we can be absolutely sure that they are due to human activities.

But what about carbon dioxide? That's the classic greenhouse gas. First, we know that the concentration of carbon dioxide has increased over the last forty years, because we have been measuring it with very high precision since 1957 or so, at initially just a couple of sites, but now there's a global network, and so we know that these changes are ubiquitous and very strong over the last forty years. When we first began these measurements, the concentration of carbon dioxide was 315 parts per million in the atmosphere, and now it's 365 parts per million. That's a really big increase.

Now, to go further back, what we have to do is go to Antarctica or Greenland--but the best records come from Antarctica--and drill a hole into the ice, and then within that ice core, we find trapped a fossil record, bubbles of air that were trapped when the ice was laid down, going back hundreds of thousands of years. And we can extract the bubbles in the ice, we can measure the atmospheric composition, we can measure the amount of carbon dioxide, and we can get a very accurate record of carbon dioxide changes going back hundreds-- literally hundreds of thousands of years, or more particularly, since preindustrial times.

Now, if we just look at the last 10,000 years, the level of carbon dioxide up to about 1700, or 1750 or so was reasonably constant...varied by a few percent, maybe 5 percent or so. Since the middle of the eithteenth century, there's been about a 30 percent increase in carbon dioxide. It seems like more than coincidence that this marks the period of industrialization and population growth and--

But couldn't that just have bubbled out of the oceans because it got warmer, because of the sun's heating?

Yes. So how do we explain this very rapid and large rise in atmospheric carbon dioxide concentration? There are a lot of possible explanations, and one is that it was just purely a natural phenomenon.

Well, we can determine that it's not natural by looking at different isotopes of carbon or carbon dioxide in the atmosphere. And the most important one of these is radiocarbon, or carbon-14. That's the constituent of carbon that is used for radiocarbon dating. There's a very, very small amount of radiocarbon in the atmosphere. It's produced by cosmic rays hitting the upper atmosphere. And normally the concentration of radiocarbon in the atmosphere is roughly constant. But we can measure how that has changed over the last 100 years. And we find that it's changed very dramatically, along with the increase in carbon dioxide. And the change is toward less radiocarbon. Now, how can that be?

Well, if the increase in carbon dioxide were due to burning fossil fuels, you have to remember that coal and oil and gas have been in the ground for a long, long time, and they have no radiocarbon left. That's all decayed away, due to radioactive decay processes. So if we burn fossil fuels, then what we're doing is, we're injecting carbon dioxide into the atmosphere that has very little or zero radiocarbon. And that would dilute the amount of radiocarbon in the atmosphere by a predictable amount. And lo and behold, the amount of observed dilution agrees with the prediction. In other words, because of this dilution of radiocarbon in the atmosphere, we can be very sure that a large component of the increase is due to burning fossil fuels.

So it's gone up by about a third, and the source is industrial fossil-fuel burning, primarily. This is new. So then we have to say: By itself, can the CO₂ really do that much? It's just one element in your climate system, right? You've talked about forcings and feedbacks. How do you distinguish between the two? How would we call a forcing? Forcings lie outside the atmosphere.

What I can do is give you a simple explanation of how the climate system works. And the analogy I like to use is of a motor vehicle. And the effect of carbon dioxide or the effect of the sun's output changing is a bit like putting your foot on the accelerator of a car. So that, in a sense, is an external forcing of the motor vehicle. By putting our foot on the accelerator, we're activating this engine, which drives the wheels and makes the car move. And in the same way, by turning up the output of the sun or by adding carbon dioxide, it's like putting our foot on the accelerator of the climate system. Now, one of the big uncertainties in the climate system is just how powerful that accelerator action is. Essentially, we don't know whether we've got a Volkswagen or a Porsche.

How sensitive it is, you mean?

We don't know how sensitive the climate system is to external forcing. I mean, we know that it's sensitive. We know that at least we've got a Volkswagen. But we might actually have a Porsche. And if we had a Porsche, then we'd have to worry a lot, because then a small amount of external forcing or a small amount of pressure on the accelerator would cause a very large climate change. So you know, my guess is, we're somewhere in between.

Given that you've got these forcings, what can either diminish or amplify the effect of you putting your foot on the accelerator? These are called feedbacks, right? So, that's your analogy of the car. So for instance, with a Porsche, presumably, that's where you see lots of positive feedbacks.

Yes, that's right. So the difference between the Porsche and the Volkswagen is related to these processes called positive feedbacks. And I'll give you one example of a positive feedback, and that is that if we were to add carbon dioxide to the earth's atmosphere and cause warming, then the oceans would warm, and the amount of water evaporating from the oceans would increase. And it happens that water vapor is also a power greenhouse gas. So that by putting carbon dioxide into the atmosphere, we increase the amount of water vapor, and so we increase the total amount of greenhouse gases in the atmosphere and amplify the effect of carbon dioxide alone. And that amplification is called a positive feedback. And there are a number of different feedback processes in the climate system. And the difficulty is knowing just how strong those feedback processes are. Some of them are relatively easy to quantify. The water vapor feedback is one.

But other feedback processes are much more difficult to quantify. An example is the effect that warming might have on clouds. You can make up a very simple argument. You could say: Well, more warming would mean more water vapor in the atmosphere, therefore more clouds, and clouds reflect incoming solar radiation, and that should have a cooling effect. And that would be a negative feedback. But, of course, clouds affect the climate system in very, very complicated ways. And you just have to look at the difference between day and night. In the daytime, clouds might make the day cooler. But in the nighttime, the presence of clouds makes the night warmer. So it's not immediately clear whether clouds cause warming or cooling, or whether they're a positive or a negative feedback effect.

Critics such as Richard Lindzen say that you have to have water vapor feedback in order to get up to the big numbers of concern. They say, if you couldn't depend on that, then CO₂ by itself wouldn't really do all that much. Is that true?

Lindzen has this argument that says that all of the other feedback process--all the positive feedback processes--are dependent on the water vapor feedback having a certain magnitude. So it's really important to quantify the water vapor feedback. And Lindzen has a theory that says that this is less than other people have suggested. It's a rather complicated theory, but essentially it boils down to this: More warming causes more evaporation, and therefore the clouds become more active, and stronger, convective clouds, or cumulus clouds or storm clouds, have greater, stronger updrafts, and inject water from the surface up into the upper atmosphere.

Now, what Lindzen says is that there's a counteracting effect that...okay, so the clouds are bringing water into the upper atmosphere, water vapor into the upper atmosphere, and that is acting as a greenhouse effect and amplifying the warming due to carbon dioxide. But, in addition, the updrafts... eventually they come out of the top of the clouds. The dry air then is denser than the surrounding environment, and so that air--that dry air--sinks down into the atmosphere, eventually back down to the surface. In fact, this is a very important process. These downdrafts of dry air outside of clouds can be extremely serious for aviation, for example. So there's no doubt that they exist. Now, what Lindzen says is that this causes a net drying effect that is greater than the effect of the moisture injected into the upper atmosphere by the clouds. And the drying effect, he claims, will reduce the water vapor feedback, and then that will have a knock-on effect and cause all the other feedback effects to be smaller.

So it's a nice hypothesis, and it's something that we can test. We can test it observationally in very simple ways by actually looking at water vapor changes in the atmosphere, or we can do things rather more sophisticated and actually quantify the effect of water vapor changes on an intra-annual time scale. In other words, we can look at the variability of water vapor in the atmosphere throughout the year and see whether, on that short time scale, the Lindzen hypothesis works.

Well, it turns out that, in both cases, the effect that Lindzen claims to be important is, we believe, not important. You know, we've tried to test his hypothesis in different ways, and we find that it actually doesn't work. It's physically realistic, but it is outweighed by other processes.

So the climate is more sensitive than he claims.

The climate's a lot more sensitive than he claims.

There's another argument, made by the coal industry, that another feedback is just the greening of the world. Basically, you put the CO₂, plants grow, trees grow; it will take it up. What's the problem with this? That would be a negative feedback. They argue in their extreme version that it's a good thing.

I've been talking about feedbacks within the climate system. But it's possible to think of feedbacks that are within the whole global environmental system. And one of those is the fact that plants grow better with higher amounts of carbon dioxide and, that as plants grow better, they absorb more carbon dioxide from the atmosphere. So that one might expect that if you burn fossil fuels, the amount of carbon dioxide stimulates the growth of vegetation and that slows down the rate of increase of carbon dioxide in the atmosphere, and that will therefore slow down the rate of climate change.

Now, that's a process that's been know for, you know, 100 years or so. And it's a process that can be reasonably well quantified. And it is incorporated into the models that we use to predict how much carbon dioxide will increase in the future. In other words, I mean, we have fully accounted for that so-called "CO₂ fertilization process" in projecting CO₂ increases in the future.

But...I was just going to say that that's actually not their argument. Their argument is that CO₂ is good for you, and you know, it's the impact on the vegetation--

Given that you can have positive and negative forcings and positive and negative feedbacks, is there any way, logically, that you can escape from this process? Given that we accept greenhouse gases are positive forcing, could ever the indirect effect of sulfate aerosols...could you offset that perpetually? Can you escape from the dilemma?

That's a pretty broad question.

Because you could say, for instance, even if the climate was less sensitive than you said, in time, if you pumped enough in, the figures would come out. You couldn't escape from it unless you had negative forcings which could keep up. Can negative forcings ever keep up? The natural ones, obviously, fluctuate both ways, but I'm talking about the anthropogenic forcings.

It's difficult to answer that question. But I can say something else that's kind of off the track a little bit but...

The leading issue here is not really how much has the climate changed in the past. Of course, we have to understand what those changes have been caused by. But the real issue is what the changes are going to be in the future. And not only what our best estimate of those changes might be, but what the range of uncertainty is. And it's possible to run a climate model and make the positive feedback effects very small--minimize them, essentially--and then get a very low estimate for future warming. And if you do that, over the next 100 years, with some reasonable estimate of what the emissions of carbon dioxide and other greenhouse gases might be, and what the emissions of sulfur dioxide might be, the warming over the next 100 years is still pretty substantial. The lowest possible amount of warming would be 1 to 1.5 degrees Celsius.

In other words, at the real bottom end of the range of possibilities, we anticipate a warming that would be at least double the rate of warming that has occurred over the last 100 years. At the other end, though, the warming could be as much as five degrees Celsius over the next 100 years. And that would be horrific, I would say.

Are you saying that's the best you can do? You can't, say, invoking the indirect effect of aerosols, knock it back even lower?

Let's just consider the effect of aerosols in the future. In the past, what's happened is, because of burning fossil fuels, the loading of these aerosols in the atmosphere has increased tremendously. The emissions of sulfur dioxide have increased tremendously over the last 100 years, globally. Now, in North America and Europe, those emissions have declined over the last 10 to 20 years. The real growth in the emissions of sulfur dioxide today is in countries like China and India. Now, what's going to happen to those countries in the future?

Now, as we become richer as a society, we tend to become much more environmentally conscious. We have different priorities, and we tend to be more concerned about issues like acid rain, urban air quality, and so on. And basically it's because we can afford to do something about sulfur dioxide emissions and sulfate aerosols that we have done something about it. Those economies of countries like China and so on are growing very rapidly. The societies are becoming richer, and we expect that they soon will be able to afford to do something about the emissions of sulfur dioxide. So, we don't expect their emissions to continue to go up. What we anticipate is that over the next twenty to thirty years or so, they will have the concern about the environment, coupled with the wealth to be able to do something about it, and they'll do what we've done. They will reduce the emissions of sulfur dioxide.

So what's going to happen in the long term, then, is that the emissions of sulfur dioxide and the loading of sulfate aerosols in the atmosphere will go down. It might go up for another twenty years or so, but after that, it's going to go down. Now, these aerosols have a cooling effect. If there are less aerosols, the net effect has to be warming. So, what's going to happen in the future is that sulfur dioxide will have the opposite effect to what [it has] had in the past. In the past, sulfur dioxide emissions have offset greenhouse gas warming. In the future, as they reduce, they will add to greenhouse gas warming.

They'll reveal it.

They will just make the warming stronger.

Given that, then, if we reach a position where you've convinced people there's cause for concern, and then the discussion goes on to asking: What do we do about it. You've mentioned residence time. This CO₂ we've put in the atmosphere stays in the atmosphere for quite a long time, doesn't it? So, it's not enough to simply cap emissions at current levels if we have a world that uses 90 percent of its energy from fossil energy. Really, to change these buildups in atmospheric gases, you have to cut massively these emissions, don't you?

In deciding what to do about future global warming, the motor car analogy's actually quite useful, because we're in this car, it's zooming along at a fairly high speed, and we're worried that there might be a cliff down at the end of the road there that we're going to go over. So what do we do about it? Well, we don't just leave the accelerator pedal at the same position. In other words, for the climate system, we don't just keep emissions at their present level. We actually have to take our foot off the accelerator. And doing that in a car isn't going to stop the car immediately. The car has a tremendous amount of inertia and is just going to keep going.

In the same way, even if we reduce the emissions of carbon dioxide and other greenhouse gases--we take our foot off the accelerator--the climate system's going to keep on going. It's going to keep on warming for a number of years. I don't mean number of years, by the way. I mean decades or centuries. So we have to do something pretty dramatic in order to stop global warming.

We don't necessarily believe that all global warming is bad, by the way. We don't necessarily believe that the increase in carbon dioxide is uniformly bad, because carbon dioxide accelerates plant growth and can be good, potentially, for agriculture. But what we're afraid of is that if the planet warms too much, we're going into unknown territory. We can't predict the climate well enough to know what to expect. So we certainly don't want to go too far down the road, down the pathway of global warming. We have time to think about what to do. But eventually, we have to do something dramatic.

We really have to move completely away from using fossil fuels for energy. Either that or we have to somehow sequester the carbon dioxide that's being emitted into the atmosphere, which would be a very costly thing to do. I think that, you know, the only solution really is to, in an economically sensible way, reduce our dependence on fossil fuels as sources of energy. You know. I think that might be quite difficult to do, but--

When you look at something like a Kyoto proposal, which is politically unacceptable to many people, how much difference would it make to the problem if it were enacted tomorrow?

If everybody were to come on board with the Kyoto protocol, then that would slow down the rate of warming of global mean temperature, slow down the rate of climate change by a very, very small amount. So we have to clearly, in the future, do something a lot more dramatic than just the Kyoto Protocol. It's a very small step along a pathway that is long and difficult.

And it's a step we're almost certainly not going to make, the way things are going. When you look at the scale of the problem from a carbon point of view.

What do we have to do? We have to stop as much getting into the atmosphere. Otherwise it will double, treble. What will happen if we just carry on exhausting fossil supplies?

Okay. If we were just to follow the Kyoto Protocol...if everybody came on board--and .I think it's unlikely that the United States will--if all countries did, that would not stop the increase in the growth of carbon dioxide emissions.

In other words, of course the Kyoto Protocol would do something: It would slow down the rate of growth. But don't forget that the Kyoto Protocol only applies to the developed countries, and right now, the other countries of the world are growing quite rapidly, and their emissions are growing, and that will continue over the next century unless something much broader than the Kyoto Protocol is enacted. If emissions of carbon dioxide continue to grow, then the level of carbon dioxide in the atmosphere will continue to grow.

Following the Kyoto Protocol would reduce the level of carbon dioxide in the atmosphere in the year 2100 by a relatively small percentage. Just to give you some numbers, what we think as a best guess, in the absence of any policies to reduce the missions of carbon dioxide...as a best guess, we think that the concentration of carbon dioxide in the year 2100 would be about 700 parts per mission, which is roughly double what it is today. If we enact and follow the Kyoto Protocol, then we might reduce the buildup by a few tens of parts per million carbon dioxide. In other words, instead of doubling the amount of carbon dioxide, we might only increase it by 90 percent. So it's clearly a very small effect.

But we still have molecules of CO₂ in the atmosphere that were produced in James Watts' steam engine, don't we? The stuff hangs around for a long time. So the question is, to turn the curves of concentration, what do you have to do?

So the Kyoto Protocol will cause a small slowdown in the rate of buildup of concentration in the atmosphere.

It won't stabilize it though, you're saying?

It will not stabilize it. Now, the guideline for the Kyoto Protocol...the background document is the United Nations Framework Convention on Climate Change. And that has as its ultimate objective the stabilization of the concentrations of greenhouse gases in the atmosphere. There are a lot of conditions associated with that stabilization. You know, for example, stabilization should be done in a way that is not economically disruptive. That may not be spelled out exactly that way in the Framework Convention, but nevertheless it's quite clear that we've got to make sure that the economy of the world isn't damaged by our desire to save the environment. There's got to be a balance between these things.

Stabilization of carbon dioxide concentration in the atmosphere...how do we do that? What level do we want to stabilize carbon dioxide in the atmosphere at? You know, clearly, the lower that stabilization level, the more we've got to do. If we were to stabilize the level of carbon dioxide in the atmosphere at, say, 50 percent above what it is today, then we could allow emissions of carbon dioxide to grow over the next ten or twenty years--or maybe the next ten years, I should say.

But after that, the emissions of carbon dioxide worldwide would have to decline rapidly. And over the next fifty years, you know, from 2010 to 2050 or 2060, the emissions of carbon dioxide would have to drop down to, say, about a third of what they are now. That's an overwhelmingly difficult task. How on earth do we achieve that? What sort of technology is required in order to go away from the burning of fossil fuels? How do we develop that technology? I mean, these are massive problems that we face.

Just what scenario is more likely? I'm just trying to get the sense of how difficult this is. What other way is there to drive a car, apart from gasoline? Hydrogen economy? That's not going to come in ten years. Right?

How much time do we have to do these things? Is it one year? Ten years? Fifty years? I think that the planning has to begin now, and it already has begun now. I mean, the Kyoto Protocol is a small step. But the Kyoto Protocol doesn't say anything about how we're going to actually get to this ultimate objective of stabilizing the level of carbon dioxide in the atmosphere. Somehow we've got to build in some incentive for technological development in order to move away from fossil fuels. That's my view, and that's the view of some of my colleagues.

And you know, that raises the issues of how on earth do we do that? How do you stimulate innovation? Technological growth? Do you do it by so-called command and control, by taxing fossil fuel usage? Or do you do it by providing incentives for technological innovation? I really don't know the answers.

I think that we probably have another ten or twenty years to figure out how to get on the pathway that will eventually lead to stabilization. It's going to require a global effort. It's not just going to be the developed world. We've got to do this in a way that embraces all of the countries of the world, because climate change is not just going to affect us; it's going to affect everybody. In fact, ironically, it's probably going to affect North America and Europe less than these developing countries. They're the people who are going to suffer.

But, on the other hand, you know, we're the people to blame, because we're the people who've put most of the carbon dioxide into the atmosphere. So, you know, somehow we've got to get together and realize that, you know, we're to blame, and the developing countries may not be so much to blame for where we are now, but the people are going to suffer, those in the developing countries. So that we and them, we together have a responsibility to do something about it.

Last question. If nothing was done, and global temperatures went up by as much as five or six degrees, are we talking about something pretty extreme there, in your view?

If the world were to warm by five degrees Celsius, and there were a really big increase in the level of carbon dioxide in the atmosphere...I mean, suppose it got up to three times the preindustrial level, which is not impossible. I mean, if we do nothing about it, that's almost sure to happen. What will that do to the environment? Well, a five-degree warming will just change the whole climate system radically. Precipitation patterns will be entirely different. The amounts of precipitation will be entirely different. You know, storm tracks would be unrecognizable from what they are today. You know, it might take a few hundred years to get there, which would give us time to plan for the changes, but the changes would be really enormous.

The worst prospect, however, is sea level rise. If we were to warm the world by five degrees, then I strongly believe that large parts of the Antarctic ice sheet would flow into the ocean and melt, and cause sea level to rise by potentially many meters. And even if that didn't happen, a five-degree warming would cause all of the smaller glaciers in the world to melt, would cause substantial melting from Greenland. It would cause the ocean to warm up and expand by a large amount. Even without parts of Antarctica falling into the ocean, sea level would still rise by meters. And that means that a lot of coastal communities would suffer tremendously. Island communities in the Pacific and the Indian Ocean would be dramatically affected. People may not be able to live in those low-lying areas.