December 17, 2008

Recent Temperature Trends in Context

Filed under: Surface, Temperature History

As 2008 nears an end, there are a lot of folks waiting to see where the final number is going to come in for this year’s global average temperature. It’s likely that the average temperature for 2008 will fall below the value for 2007 and quite possibly be the coldest year of the (official) 21st century. 2008 will add another to the growing recent string of years during which time global average temperatures have not risen. Does this mean that pressure of “global warming” fuelled by increasing greenhouse gas emissions from human activity has abated?

The answer is a qualified “no”—it seems that natural variations have been flexing their muscles and offsetting anthropogenic warming.

Figure 1 shows the time series of monthly average global temperatures from January 1850 to October 2008, as given by Britain’s Hadley Centre. Clearly, overall, temperatures have risen. (We have noted elsewhere that there is probably an overestimation of warming in recent decades).

Figure 1b shows the same history since 1965. The overall rise is still clear, but the temperatures during the last several years have just as clearly remained relatively level.

So what does this say about the rates of global warming? Is the lack of temperature rise in recent years unusual?

Figure 1a (top). Monthly time series of global average surface temperature, January 1850-October 2008. Figure 1b (bottom) same as Figure 1a except January 1965-October 2008 (data source: Hadley Centre).

To examine better this, we calculated the linear trend through the most recent 10 years of temperature data and compared it to the trend through all other 10-yr periods during the past 40 years. Our results are shown in Figure 2. As can be seen, recent trends are not particularly unusual in the broader context.

Figure 2. 10-yr linear trend (red) through successive periods in the surface temperature data (blue). The leftmost red box contains the 10 years’ worth of temperature data which are used to calculate the leftmost red point. The rightmost red box contains the most recent 10 years of data which was used to calculate the last trend point. The middle red box contains a recent 10 years’ worth of temperature data which has a lower trend than the most-recent 10years.

To see if our analysis was sensitive to the length of the trend we selected, we did the same thing for 8-yr trends and 15-year trends. The results we found were similar (Figure 3). In each case, recent trends are relatively low, but not unprecedented in the past 40 years.

Figure 3. Running 8-yr (blue), 10-yr (red), and 15-yr (green) trends through the monthly average global surface temperature data, January 1965-October 2008.

The reason for the rise and fall of the trend values is that, over these time scales, natural variations—primarily the cycle of El Niño/La Niña and volcanic eruptions—can and do impart a sizable signal to global temperatures.

Compare the variations in the strength of El Niño/La Niña as depicted in Figure 4 with the variations in the magnitude of the short-term (8- and 10-yr) trends depicted in Figure 3. Also notice that the trace of the 15-yr trend in Figure 3 is much less sensitive to El Niño/La Niña variations.

Figure 4. The strength of El Nino/La Nina as captured by the smoothed Multivariate ENSO Index (data source:

Some of you may have noticed that there is a statistically significant positive trend in each of the three (8-yr, 10-yr, and 15yr) trend time series (Figure 3). A trend of a trend is an indication of an acceleration taking place.

But don’t go and get too excited about this.

The reason for this acceleration is, as they say, “more apparent than real.”

It is not from increasing emissions of greenhouse gases, but instead, is an artifact of the eruptions of El Chicón and Mt. Pinatubo, which acted to cool temperatures in the early-to-mid 1980s and early-to-mid 1990s beneath where they otherwise would have been. To demonstrate their effect on the trend calculations, we statistically adjust the temperature record to approximate what it likely would have been had these large volcanic eruptions not occurred. Figure 5 shows our results. Note that the large 1982-83 El Niño signal seen in Figure 4 (a warming) becomes much more pronounced as does the warming signal of the extended moderate El Niño in the early 1990s. The warming influence of these El Niño events had been offset by the contemporaneous cooling of the volcanic eruptions.

Figure 5. Observed monthly average global surface temperature anomalies (blue) and the global surface temperature anomalies adjusted to remove the effect of volcanic eruptions (red).

So what effect does the removal of the volcanic signal have on the running trends? The putative acceleration of global warming vanishes.

Figure 6 shows the 10-yr running trend through the observed global temperatures along with the 10-yr running trend through the global temperatures with the effect of volcanoes removed. The trend through the trends is now gone, and the amount of variation is slightly reduced.

Figure 6. Running 10-yr trends calculated through observed global temperatures (blue), and the 10-yr running trend though temperatures with the effect of volcanic eruptions removed (red).

Also notice in Figure 6 that still, even after accounting for the onset and recovery from volcanic eruptions, the low trend values in recent years are still not particularly unusual.

But, there is an interesting recent development that may be an indication that the ongoing slowdown in the global temperature rise will last longer than other recent ones, and therefore become unprecedented in this era. This concerns the large-scale behavior of the sea surface temperature (SST) patterns in the Pacific Ocean—a phenomenon known as the Pacific Decadal Oscillation, or PDO for short. The PDO represents a flip-flopping of the persistent patterns of SST across the Pacific and seems to change its state every several decades (Figure 7). In its positive state, global temperatures seem to rise, while in its negative state, they seem to fall. There was a change in the PDO from positive to negative in the early 1940s and again from negative to positive in the mid-late 1970s. This matches well with the character of global temperature over the past century—an increase into the early 1940s, a cooler period from the early 1940s to the mid 1970s and increasing temperature again since the late 1970s (see Figure 1).

Figure 7. The monthly value of the PDO index from January 1900 through September 2008. For continued updates and to follow how the PDO progresses in the future, visit:

A close look at the recent values of the PDO index values shows that there is some indication that it may be in the process of switching from positive to negative. In fact, the folks at NASA’s Jet Propulsion Laboratory have begun to suggest that this is a real possibility. If it proves to be true, then we may see an extended period of time, perhaps even several decades, in which a natural cooling offsets at least a portion of the growing warming pressure of an increasing greenhouse effect. The rise in global temperature could be stopped for years to come.

If, on the other hand, the PDO is just flirting with negative territory, and instead returns back to the positive state that has marked the period since 1976, then global temperature may soon begin again their upward climb. Only time will tell.

The bottom line is this: the anthropogenic influence on global temperatures, while surely omnipresent, is not of a magnitude which prevents the influence of natural variations within the earth’s climate system from dominating the global temperature record for periods of years to perhaps even decades—with the downstream effects impacting the ultimate course that climate will take during the coming century.

While the anthropogenic pressure towards global warming has not stopped, it most definitely has been sidetracked.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress