Clearing the Air

The Grit has helped me understand the importance of clarity in communication, and some of the ways in which I tend to be unclear.

While I do see some of your information as interesting, I still need to know more details. For instance, when you talk of global average temperature increases, what exactly do you mean. It could be that every place on Earth has an increased temperature, or one small area has a greatly increased temperature. This to me, and I suspect others, is confusing.

Excellent point. It highlights the fact that some things which are second nature to me are unknown (and therefore often confusing) to others, and that I often use terminology which is transparent to me, but not to readers.

Temperature anomaly is the difference between the temperature, and its average over some “reference period”; climate scientists generally analyze temperature anomaly rather than temperature. Why? Suppose there are two temperature recording stations in the same cornfield, but one of them is six feet off the ground while the other is elevated 50 feet in the air. They will record different temperatures! But generally they will indicate the same anomaly, i.e., when one says it’s hotter or colder than yesterday so will the other. By taking the difference between individual measurements and averages from the same station over a reference period, we separate the changes (which is what we’re really interested in) from the location-specific conditions. Also, anomalies are defined as the difference between temperature and the average for the same date, so we can remove the seasonal variation (unless of course we want to study the seasonal variation).

There are two principal “keepers of the thermometer,” the Hadley Center/Climate Research Unit (HadCRU) in England, associated with the U.K. meteorological office and the University of East Anglia, and the Goddard Institute for Space Studies (GISS), part of NASA. Temperature anomaly in HadCRU data is defined relative to a reference period 1961-1990, while that in GISS data use the 1951-1980 averages to define anomaly.

Global average temperature anomaly is defined as the area-weighted average of temperature anomaly over the entire globe. This is what you’ll see plotted in almost all graphs related to global warming in news and other media reports. “Area-weighted” means that each input to the average is weighted proportional to the size (area of earth’s surface) it represents. It’s area-weighted for two reasons. First, we don’t want small areas with lots of meteorological stations to dominate the global average. Without area weighting, the U.S. (and to a lesser degree, Europe) would dominate the world average simply because we have so many meteorological records. Second, area-weighting gives the closest approximation to the total thermal energy in the lower-atmosphere climate system.

Since it’s an average over all regions of the globe, it is indeed true that if one place gets hotter while another equal-size area gets colder by the same amount, global average temperature will not change. Likewise, if an area half the size cools by twice the amount, or an area twice the size cools by half the amount, the global average won’t change.

If global average temperature rises, is it because the whole globe has warmed, or just a few locations? The fact is, the global average doesn’t give this information. To answer that question, we need to know the distribution of warming; this graph shows the geographic distribution of temperature anomaly worldwide according to the 2005 report of the World Meteorological Organization.

Clearly the distribution of warming is not uniform. Small areas of the globe have cooled, particularly in the southern oceans. But vastly more areas of the globe have warmed, and by a much larger amount, so the overall average is considerable net warming. It’s also interesting to note the pattern of warming; it’s most extreme at far north latitudes, and in general land warms faster than ocean. This pattern was predicted by computer models.

Sorry, but you can’t have it both ways. If the corrected data from NASA is off, why should I trust any corrected data? Really, I don’t buy into that.

When it comes to trusting data, that’s risky business. The thing to do is to evaluate the reliability of the data as best one can. One of the best ways to do this is to look for the same information from independent sources, to see whether they agree. But even that isn’t infallible! That’s what lead to such high confidence in the claim that tropospheric temperature wasn’t warming; two independent sources (satellites and balloons) both gave the same result. In this case it turned out that they were both wrong; that’s rare — but it’s hardly the first time.

As for the surface temperature record from thermometers, there are lots of reasons to have very high confidence. For one thing, we’ve been at it (measuring temperature with thermometers) for a long time — nearly four centuries in some locations. Also, it’s been studied by scores, perhaps hundreds of science teams over the last fifty years, while the satellite data have really only been scrutinized by a half dozen or so for a few decades. For another thing, reading a thermometer is a pretty simple procedure (compared to measuring microwave brightness!), and the history of science shows that the simpler a measurement is, the (dramatically) more likely it is to be correct. Also, there are lots of independent, corroborating evidences that the globe is warming, including:

1. Steady decrease in polar ice extent and thinning of the polar ice cap.
2. Retreat (or disappearance) of glaciers worldwide.
3. Collapse of the Larsen B ice shelf (and others) on the Antarctic peninsula, and the breaking of the Ward Hunt ice shelf in the Arctic.
4. Earlier arrival (by about three weeks) of spring and later arrival of winter in mid-to-high northern latitudes. This is most strongly indicated by the earlier arrival of spring snowmelt runoff, a crucial source of water for much of human society.
5. Melting of permafrost in Alaska and Siberia.
6. Migration of species habitat, most dramatically the burgeoning of invasive species in northern regions where they used to be killed off by winter freezing (like the pine beetle that has destroyed millions of acres of pine forest in regions which no longer have enough “killing frosts” to eliminate them).
7. Dramatic increase in wildfires (in the western U.S. and many other areas of the world).
8. Dramatic increase in the severity of heat waves. Most notable is the European heatwave of 2003, which killed 30,000 people. Statistically this was “5-sigma” event, which means that if climate had not warmed, the chance of its happening naturally is less than one in a million.
9. Large-scale bleaching of coral reefs.

Also, if the climate change in the 40-50 years was negative, then the greenhouse gas theory must be tossed out the window, since we did churn out a large amount of the stuff during that time span.

On the contrary, it merely highlights the fact that greenhouse gases are not the only factor affecting global climate. It’s true we did “churn out a large amount of the stuff during that time span,” but it’s also true that we churned out large amounts of other stuff, in particular sulfate aerosols.

Aerosols are tiny particles that tend to stay in the atmosphere for a long time. They also tend to reflect sunlight back to space, decreasing the amount of solar energy reaching the surface, and this cools earth’s climate. This is why massive volcanic eruptions cool the planet; they inject aerosols into the atmosphere, and if they’re emitted in high enough quantities high enough into the air they disperse and cause global cooling.

There was slight global cooling from about 1940 to 1975, and I say “slight” because it’s so small that it’s not really “statistically significant” (i.e., it could be just random “natural variation” rather than a genuine trend). But this much is abundantly clear: we didn’t see global warming during that time period!

That’s because the industrial activity of World War II and the post-war era created so much in the way of man-made aerosols that their cooling effect negated other factors. Keep in mind that although there was man-made CO₂ (and other greenhouse gases) during this period, there wasn’t as much as there is today (CO₂ stays in the atmosphere for a long time, so it builds up — the stuff from the WW2 era is still there) . Even so, it had an impact; if not for the greenhouse gases, the aerosols would have caused global cooling; the effect of both canceled each other making global temperature reasonably stable.

There are three main “episodes” of temperature change in the last century. First is a period of warming from about 1915 to 1940, at a rate of about 0.13^oC/decade. This was followed by a period of (very slight) cooling from about 1940 to 1975 at about -0.02^oC/decade. Finally is a pronounced warming from 1975 to the present at a rate of 0.18^oC/decade. In the following graph, each black dot shows the global average temperature anomaly during one month; the graph shows values from January 1856 to December 2005. There are also two “smoothed” versions: the blue line is annual averages and the red line is 10-year averages. By smoothing, we suppress the effect of month-to-month fluctuations and accentuate the longer-term trends.

We can also see the effect of other climate factors, including the cooling just after 1992 caused by the explosion of the Mt. Pinatubo volcano, and the extreme warmth of 1998 caused by the strong el Nino of that year.

The early-20^th century warming is due to the combination of several factors: increase in solar irradiance, a period of unusually low volcanic activity (hence a lower-than-average concentration of cooling aerosols), and greenhouse gases. The mid-century levelling off is due to the effect of man-made aerosols canceling the effect of increasing greenhouse gas concentrations. The recent warming is because aerosol emissions were severely limited in the 70s by law (they’re responsible for acid rain), and because greenhouse gases have now reached levels where they dominate other factors.

However, I did read somewhere that the hottest decade was somewhere around then.

As the graph above shows, this is simply not the case. The hottest decade so far is 1990-2000, but it’s a pretty sure bet that when this decade (2000-2010) is complete it’ll surpass the previous one.