Temperature
Measurement
Recent past
Before scientists can say that there is global warming, they need to be able to estimate the Global Mean Surface Temperature (GMST) both now and well into the past and study how it changes.
Using a local temperature will not do as there may be local factors affecting this. For example, due to plate tectonics the UK has wandered considerably over time (it was once south of the equator), and the south of England is still sinking now that glaciers are not weighing down the north. Also, the centres of cities (urban heat islands) tend to be warmer than the surrounding countryside, so if the temperature has been recorded as a city growed it will show a higher rise than average.
There have been many temperature scales proposed and used over the years. One commonly used is the Celsius scale (ºC) where there are 100 degrees between the boiling (100ºC) and the freezing (0ºC) of pure water at normal pressure.
It is important that the measurements around the world are taken in a consistent manner, as the temperature can vary by the height of the thermometer from the ground, and whether it is in the shade. Thermometers are therefore kept about 1 meter from the ground in a white louvred box called a Stevenson screen, like the one illustrated on this Irish Meterorological Service page. The locations where the measurments are taken need to be representative of the world as a whole, and measurements should be accurate. Often the avearage over 30 years is taken, to reduce the effects of unusual years and bring out the long term trend.
A meteorologist called Gorson Manley collated measurements taken in various places in England from January 1659 onwards and attempted to standardize them. Others have continued the work since, and the results are reported quarterly in the Weather monthly magazine of the Royal Meterological Society. Similar work has produced an annual GMST figure from 1861 onwards. The Intergovernmental Panel on Climate Change uses figures produced by the Meterological Office and the University of East Anglia.
GMST is not the whole story, as the rate of change at the Equator has been considerably less than that in the Arctic and Antarctic.
Earlier
To estimate the temperature earlier than 1659, scientists use their knowledge of which flora and fauna occur with which climates, and look for fossil records. They also look at the geology, for example for signs of past glaciation - signs that match the efects of current glacial action.
Glaciers ground out U-shaped valleys compared to the V-shaped valleys created by fast-flowing rivers, and a lot of rocky debris is deposited at the end of glaciers. This enables us to tell that all of Scotland and Canada, parts of Northern USA and Northern Britain and Wales have been under ice sheets at some time, for example.
Preserved plant and animal fossils (including pollen grains) which we recognise can tell us what the climate was like when they lived, by comparing them with the kind of climates in which they or similar species live today. Cores can be taken in peat bogs and sediments which show many thousand years of history, dated using radiocarbon dating, and the pollen will show how the climate changed over time. Other factors could also affect the types of pollen, however - for instance agricultural practices, and (in recent decades) acid rain. The oldest core, covering over 140,000 years, came from Garnd Pile in the Vosges, France.
Fossil records go back much further - as far as 2.4 million years ago.
Effects
For most places, a very slow rise (i.e. slow enough for nature to react and migrate) in average temperature of a few degrees Celsius would result in a natural environment similar to that currently experienced by places several hundred metres closer to the Equator. It would become impossible to continue to grow some temperature-sensitive crops (such as maize) in some areas, but newly possible in other areas. Some sea routes might cease to be solid ice. If the temperature fell, the opposite would happen, glaciers and ice sheets would grow.
Sea levels would rise. Firstly, the volume of the water in the ocean increases with temperature rises (thermal expansion). Secondly, any melted ice from glaciers and ice-sheets currently on land would run into the seas. As a rough guide, an increase in 3ºC might result in a sea level raised by roughly 30cm. This would be enough to make many low-lying areas such as the Maldives (highest point 2.4 m) and parts of Bangladesh, southern China, the UK, the Netherlands and southern USA more susceptible to sea flooding.
Some places would need more fuel for heat in winter (and more insulation), and others for air conditioning in summer. This could affect deforestation and pollution locally, and/or fuel imports. Precipitation might increase or decrease, also affecting crops.
Under construction
Acknowledgements
The Open University course: S190 Global Warming - the Science Behind the Headlines
References
S103 Discovering Science, Block 2: A Temperate Earth? ISBN 0 7492 8188 X, Open University, 1998