The Twentieth-Century (and beyond) Warming

Observed Temperature Changes in the 20th Century

We will now take a look at the global average surface temperature record since 1850. This record is constructed from actual measurements and thus can be considered more accurate than the proxy reconstructions we previously studied. However, there are issues with the measured temperature record that make it uncertain as well and you will sometimes see spirited debates about the details of the recent temperature record constructed from measurements. Several of these issues will simply be mentioned in the couple of paragraphs, though, we will not go into the specific details of the arguments about possible errors in the recent temperature record.

Ideally, we would like to have accurate thermometers whose measurement characteristics do not change with time and which remain in the same location and surrounding environment for a long period of time. This would allow us to measure real changes in temperature without influence from changes in the instrument or the local environment. Unfortunately, this is not generally the case. At most observing locations, the instrumentation is changed over time and adjustments must be made to account for differences in how different thermometers measure temperature. For many observing locations, urbanization (city building) has happened around the measuring site. Urban areas are generally warmer than nearby rural areas because many of the building materials effectively retain heat better than the natural land surface and also due to the heat generated within cities. This urban heat island effect can look like a real warming (or increase in temperature over time) unless adjustments are made to remove the effect. In fact any change in the local environment around a thermometer location needs to be considered, even something as simple as putting up new buildings or installing an air conditioner near the thermometer. In many cases, the actual location of the thermometer was moved (say within the same city) and corrections for things like changes in the altitude of the location must be made.

Another set of issues have to do with the geography of the available thermometer locations. Ideally we would sample the entire globe evenly since we want to measure changes in global average temperature. In actuality, there is a dense network of measurements in a few developed areas on Earth and rather sparse coverage over unpopulated land areas and oceans. Thus, adjustments must also be made to properly represent the global average from this observing network. In places where there is only sparse coverage of observations, it can be difficult to determine the true average temperature over a large region from just a few observations.

The figure below shows one reconstruction of the changes in global average temperature based on actual temperture measurements. According to the figure, the global average surface temperatures on Earth have risen by about 0.8°C (1.4°F) since 1850. The red bars indicate the global average temperature each year relative to the average temperature over the period 1961-1990. Thus, the plot shows the difference in temperature between each individual year and the 1961-1990 average. The blue line is a smoothed version of the plot drawn to highlight trends in temperature. There is some uncertainty in determining global average temperatures from the available measurements due to instrumental errors, changes in instrumentation over the record, changes in the local environment surrounding the thermometers, and the fact that measurements are quite sparse over some areas of the world, but even the most conservative estimates indicate that the surface temperature has risen at least 0.5°C (0.9°F) since 1850.

While there are arguments about how much global average temperature has increased over the period of the instrumental record, most everyone agrees that global average temperature has increased. One more point to make in this section. While we often talk in terms of global average temperature changes, keep in mind that there will be variation in the temperature changes at regional scales and over seasonal and yearly timescales at a particular location. It is important to understand this when considering the potential impacts of the temperature changes. As an example, the figure below shows the temperature differences for the year 2006 compared to the 1961-1990 average temperature. Even though the average global temperature for 2006 was about 0.4°C higher than the 1961-1990 average, note that some regions of the world showed much larger warm anomalies (over 4°C in some areas), while other regions actually had lower than aveage temperatures (over 1°C in some areas) for the year 2006. This again shows that even if the global average temperature goes up, there will be quite a bit of variability expected at regional spatial scales.

Implications of Temperature Changes in the 20th Century

We need to consider the question: Is the recently measured warming of global average surface temperatures caused by the increasing amounts of greenhouse gases in the atmosphere, or is it part of a natural cycle?

According to climate models the warming is largely due to the additional greenhouse gases. If greenhouse gas concentrations are left fixed at pre-industrial levels, the models show little if any warming during the twentieth century, while the same models do predict warming when they include the observed increases in greenhouse gases. However, climate models are unable to reconstruct the irregular pattern of the observed warming since 1860.

The irregular warming has been used by opponents of anthropogenic global warming (those who do not think the recent warming has much to do with human-added greenhouse gases) as evidence that natural climate fluctuations dominate over the effects caused by increased greenhouse gases. In particular, they point to two periods:

1. The rather large warming that occurred from 1910-1940. This warming is too rapid and too large to be fully explained by the rather small increases in greenhouse gas concentrations that occurred during this period.
2. The cooling trend that occurred from about 1940 to the late 1970s. This cooling occurred even though greenhouse gas concentrations continued to increase during this period.

Proponents of global warming caused by greenhouse gas increases contend that even during the general warming trend expected as greenhouse gases increases, there will be fluctuations along the way. They also claim that the cooling period from 1940-1970s can be explained by high aerosol concentrations in the atmosphere during this period related to "dirty" coal burning. Coal contains sulfur and when it is burned sulfur dioxide gas is released. The sulfur dioxide undergoes a gas to particle conversion to become a sulfate aerosol. Sulfate aerosols cool the Earth's surface because they reflect some of the incoming radiation from the sun toward outer space. Sulfate aerosols also disolve in raindrops forming a solution of sulfuric acid known as acid rain. Acid rain is quite damaging to natural lakes and forests and accelerates the decay of some building materials including culturally important statues. Prior to 1980 the US and Europe engaged in "dirty" coal burning, resulting in high amounts of sulfate aerosols, which may have acted as a cooling influence on global average temperature. After 1980 regulations were put in place to reduce acid rain and we started cleaning up our coal burning operations by reducing the amount of sulfur dioxide gas released. This reduced the amount of sulfate aerosols in the atmosphere after 1980 and according to this explanation, once the aerosols cleared, we started to again see global warming caused by increasing greenhouse gases.

While this explanation for the "unexpected" cooling that took place from about 1940 through 1980 seems plausible, it is difficult to believe that most of the recent changes in global average temperature are mainly due to increases in greenhouse gases and changes in sulfate aerosols. Similar cooling and warming periods are seen all throughout history regardless of human acitivity. Additionally, even though western world cleaned up its coal burning after 1980, new coal fired power plants continue to be built throughout China and other parts of the developing world. The increase in sulfate aerosols produced by the developing world more than makes up for the reductions in the western world. By the way the cooling during the period 1940-1980 is "unexpected" only if you believe that recent changes in global average temperature are dominated by increases in greenhouse gases. If you believe that natural fluctuations in temperature are dominating over human factors, then the cooling period is not unexpected. Finally, the acid rain issue provides another example of a large-scale environmental problem that was improved through regulation (at least in the western world), similar to the international agreements to eliminate production of ozone destroying CFCs. In the case of acid rain, cause and effect was easily established and it was relatively easy to reduce sulfur emissions by burning low sulfur coal or scrubbing the emissions. This is different from the global warming issue where we do not have a clear cause and effect relationship and we have no current substitute to using fossil fuels to meet the current energy demands of the world.

Based on the historical temperature record, some claim that the warming since 1980 appears to be the fastest rate of increase over the last 10,000 years, but we cannot prove that since we would be unable to resolve all short term temperature variations that have occurred over the last 10,000 years. In other words, because we cannot resolve year to year variations in global avearge temperature from proxy records, temperature reconstructions based on proxy records are a "smoothed" version of the real temperature changes. (Also, note in the global average temperature reconstruction shown above that the rate of increase in global average temperature between 1910 and 1940 is not that much different from the rate of warming between 1980 and today). Thus, there is no way to conclude that the rate of temperature increase today is faster than any other time in the last 10,000 years. However, we can say with some confidence that through the year 2008, the warmest 13 years in the measured temperature record (since 1860) have been recorded since 1995 and the 15 warmest years have all occurred since 1990 (see Rank diagram warmest to coldest years from Hadley Centre). For many, this is enough evidence that the recent temperature increases are not part of some natural cycle, and must be due to human activity, caused by increasing greenhouse gas concentrations in the atmosphere.

However, if we go back 800 years to the Medieval warm period (or Little Climatic Optimum), current global average temperatures may be only slightly warmer than that period. (If we go back 6,000 years to the Holocene Optimum, global average temperature was most likely warmer than it is today.) Again, a possible difference today is that the rate of warming (0.8°C in 100 years) may be faster than the rate of warming that lead up to the Medieval Warm Period, although as stated above we cannot confirm this.

Another issue, which has been gaining attention lately, is that the rate of warming of global average surface temperatures has certaintly slowed in the last 10 years. There is even a hint that the smoothed annual average global temperature may have topped out around 2002 and is beginning to trend downward. Keep in mind that we will need more years of data to see if this does indeed happen. Note also that the year 2010 was a warm year, which hints that we may not be entering a cooling period. Nevertheless, after about three decades of warming global average temperature, it is entirely possible that the Earth is now be entering another multi-decadal period of cooling global average temperature. There is evidence for a multi-decadal oscillation in global warming and global cooling since 1860. In fact, several multi-decadal natural oscillations have been identified in climate records in recent years. For example the Pacific Decadal Oscillation (PDO) contains identifiable warm and cold phases in sea surface temperature in the north Pacific Ocean (a similar North Atlantic Oscillation [NAO] has also been identified). Several researchers have gone out on a limb to predict global cooling over the next 20-40 years based on a reversal of the PDO from warm phase, which began in 1977, to cold phase, which may have begun around 2000 (see another perspective on recent temperature changes.) Supposing that we are entering a period of cooling global average temperatures, what might this mean in terms of answering the question of whether or not global average temperature is significantly warmer due to anthropogenic greenhouse gas increases?

For those convinced that our emissions of greenhouse gases are significantly warming global averge surface temperature, a cooling period is simply explained as a natural climate variation sitting on top of a longer term rise in global average temperature due to increased greenhouse gases. It is difficult to dismiss this possibility.

For those convinced that recent increases in greenhouse gases are not significantly changing global average temperature, this indicates that natural variations dominate over higher greenhouse gas concentrations. The longer term rise in global average temperatures on top of the multi-decadal oscillation is simply a recovery from the little ice age cold. Some point to a possible longer term multi-century oscillation between warm and cold peroids, i.e., medieval warm period, little ice age, current warm period, etc.

You should realize that even if we enter a period of 30 years of global cooling, this does not prove or disprove anthropogenic global warming. However, 30 years of global cooling is not predicted by climate models or the 2007 IPCC report, so if it does occur, it may be seen as a failure of current climate models to predict future climate changes.