Attribution: Starts with Temperature

Attribution (2)

This is the second in a series on the attribution of climate change; that is, how do we determine to what extent the observed warming is caused by humans? Here is the first entry Attribution (1) .

In the first blog a comparison was made with the ozone problem, and it was argued that the attribution of ozone depletion to chlorofluorocarbons had a “smoking gun.” The smoking gun was the ozone hole, and the observations and the modeling that described the ozone hole. Attribution is a matter of both observations and models, with the models representing the hypotheses based on the basic physical and chemical principles. The question was raised whether or not there was a “smoking gun,” an unequivocal observation that proves that global warming can be attributed to humans.

The short answer: There is not an equivalent smoking gun. Rather, like most criminal trials in the U.S., there is a collection of evidence. The question of attribution then becomes one of an accumulation of evidence and whether or not reasonable doubt can be eliminated. (Do you believe that smoking causes lung cancer? Do all people believe that smoking causes lung cancer? How is this problem the same, different, from the climate change attribution problem?)

Climate change predictions project with great certainty that temperature will rise, ice will melt, sea level will rise, and weather will change. The weather will change because there will be more energy (heat), the patterns of surface temperature will change, and the moisture content of the atmosphere will increase with the high temperatures. If the surface temperature of the globe is increasing, then first and foremost, we should be able to measure this increase. Here is a recent picture from the From the National Climatic Data Center showing temperatures for March from 1880 to 2008.




Figure 1. Time series of anomalies (anomalies = difference from an average) of monthly-averaged temperature for March, 1880-2008. From the National Climatic Data Center .

The top part of the figure is land and ocean combined, the middle part only the ocean, and the bottom part only the land. What does the figure show? In all instances the last three decades are the warmest part of the record. Many would look at this figure and identify a trend, but much more analysis is required. First, there is the length of the record, what has gone on before 1880? Second, there is the definition of the averaging period, 1961-1990, about which the anomalies (differences) are calculated. Third, this is only March, what about the rest of the year? And there are questions about the sampling of the observations and the quality of the observations. These and other questions are brought to bear on the conclusions that are drawn from the observations.

In Figure 1 there are a number of interesting attributes to note. First, over land and the land-ocean combination March 2008 is very warm. This continues a string of years that are quite warm, which is consistent with systematic warming of the planet. Second, in the past three years, the ocean surface has not been as warm as in the previous five. This defies the notion of a simple, always increasing surface temperature. It is worth noting, that despite the ocean cooling, the land has continued its warming. Also worth noting in the figures is the fact that the variability over land is more than twice of that over the ocean. There is a lot going here. It is not possible to convincingly conclude from this figure that warming is attributable to humans; in fact, from this figure alone it is not possible to conclude that the Earth is, on average, warming.

In one of my first blogs I showed a picture of temperature anomalies as they are geographically distributed. In the comments to that blog a couple of people stated that seeing that geographical distribution of warming was compelling evidence of not only warming, but human induced warming. Here is an updated version of that picture from 2007.



Figure 2. Difference of average surface temperature for 2007. From the National Climatic Data Center .

It is warm over the land masses of the northern hemisphere, especially in Russia. There are some cool patches, especially in the ocean, and in the southern part of South America. That the continents are warming faster than the ocean and that high latitudes in the northern hemisphere are warming fastest are, indeed, consistent with the predictions of greenhouse gas global warming. The consistency of this geographical distribution is a piece of corroborating evidence, but it is not sufficient to attribute warming to human activity. On the other hand, since this is a primary prediction of a simple variable with simple underlying physics, if this signal were absent, it would be a strong piece of contrary evidence. That is one of burdens of proof. The presence of certain signals is necessary but not sufficient for proof, but they may be sufficient to raise reasonable doubt.

The ocean: There is cooling over the ocean, and with the benefit of accumulated knowledge, we know that there is a source of internal variability in the ocean that is important on the global scale. That is the El Nino – La Nina atmosphere-ocean oscillation in the Pacific Basin. This oscillation is significant enough to alter the global temperature, and introduces the idea of internal variability. There are many important modes of variability in the climate system, and these oscillations must be accounted for in the attribution problem. The signal of greatest cooling at the equator in the eastern Pacific is indicative of a La Nina, and it contributes to a cooler globe. Similarly, when there is an El Nino, for instance the 1997-98 El Nino, the planet is warmer. This notion of natural variability makes the determination of trends more difficult, or maybe it makes the determination more interesting. It is, however, possible to extract the natural variability from the observations. That will be next.

r


Attribution of Climate Change to Human Activities:

WU Blog on Models and Attribution

Attribution (1)