Combining land and sea surface temperatures over the last 140 years
To calculate an average surface temperature for the Earth, we need to measure temperature evenly across its surface. We can, for example, divide it up into grid boxes of 5° latitude by 5° longitude and then average trends in temperatures across all 5184 grid squares.

The raw data for the calculation of mean surface temperature of the earth is measured from two main sources:

1. monthly readings from a network of over 3000 surface temperature observation stations
2. sea surface temperatures measurements taken mainly from the fleet of merchant ships, some naval ships and a network of data buoys

Satellite measurements of lower atmosphere temperatures are also available for the period of the last 30 years or so.

While thermometer readings go as far back as 1659 (the earliest are part of what is known as the Central England Temperature series) accurate and geographically diverse coverage on the instrumental record goes back around 140 years to 1860.

Over the years, measurement methods (and therefore accuracy) at both land stations and sea observations have changed. For example, urbanisation around land stations has had an effect on long-term land surface measurements in those locations that were once rural and are now urban.

Methods of collecting sea surface temperatures data have also changed over the years. Sea surface temperature is no longer measured by scooping up a wooden bucket of sea water and sticking a thermometer into it, but is now - believe it or not - usually made by measuring the temperature of cooling water entering merchant ships’ engine systems.

Stripping these and other weird and wonderful effects out of the raw historical data throughout the instrumental record has not been trivial. Methods to do this have themselves been the subject of much scientific debate.

Satellite observations of lower atmosphere temperature
In addition to the land and sea surface temperature series, during the last 30 years or so, we have also been gathering satellite observations of the temperature of  the lowest layer of the atmosphere (the lower troposphere). Meteorological satellites carrying microwave sounding units can remotely observe the average temperature of the lower troposphere.

Currently, scientists agree that the satellite data suggest the lower troposphere has been warming (as we might expect if the surface has been warming), although there is a debate among climatologists about the rate of warming (roughly in the range of 0.1-0.2 °C over 30 years).

Until relatively recently, while the surface temperature data from thermometers showed a significant warming, the satellite data showed a slight cooling. For many years this was a source of great interest for climate change sceptics who smelt a rat. However, the satellite cooling trend turns to a warming trend once scientists took account of subtle changes in the orbits of the satellites making the measurements.

Measuring changes over the last 1000 years
Temperature records before the mid-nineteenth century are rather sparse and inaccurate. If we want to deduce global mean surface temperature in the period before the instrumental record, we have to measure it indirectly using indicators from the paleoclimatic record.

Such ‘paleoclimatic proxies’ come from a variety of sources including:

• borehole measurements
• corals
• ice cores
• pollen distribution
• records of lake levels
• records of glacier advances and retreats
• tree rings

There are considerable methodological issues in interpreting data obtained from any of these proxies. Tree ring data, for example, are only available on land while coral data relate only to the tropical and sub tropical regions (you won’t be surprised to hear!).

Furthermore, historical documents, for example, tend to be biased towards describing more extreme events. So constructing global mean surface temperature as far back as 1000 years is a question of piecing together a story from lots of different sources of information. It is for this reason that it has been possible to deduce the mean surface temperature for the Northern Hemisphere for the last 1000 years, but not for the Southern Hemisphere.

As we travel back in time, the uncertainties in our measurements of global mean surface temperature cascade. So if you look at the hockey stick graphic from right to left it starts off as a neat, crisp single line which gradually thickens as we look further back in time. Eventually it resembles a sound waveform (like you might see in an audio editing application), a marked change from the crisp line on the right side of the graph. The thickness of our earlier estimates of global mean surface temperature represents the range of confidence we have in the scientific results of using paleoclimatic data.

Who knows, maybe next time you come across a news article on ‘global warming’ or ‘climate change’ it may raise a knowing smile?

  < previous   Page 2 of 2

Comments