The albedo of a surface is the ratio of reflected to incoming radiation. This ratio depends on the wavelength of the light: unqualified, it usually means visible light. Fresh snow albedos are high: up to 0.9. Powdered charcoal has a low albedo. Earth has an average albedo of 37-39%.
According to the National Climatic Data Center's GHCN 2 data, which is composed of 30-year smoothed climatic means for thousands of weather stations across the world, the college weather station at Fairbanks, Alaska, is about 3°C (5°F) warmer than the airport at Fairbanks, partly because of drainage patterns but also largely because of the lower albedo at the college resulting from a higher concentration of pine trees and therefore less open snowy ground to reflect the heat back into space. Neunke and Kukla have shown that this difference is especially marked during the late winter months, when solar radiation is greater. To give another example, the entire nation of Belgium, which is among the most urbanized in the world, is considerably warmer than the mostly open, unforested French countryside that lies to its immediate southwest, perhaps also because of albedo effects.
Although the albedo-temperature effect is most famous in colder regions of Earth, because more snow falls there, it is actually much stronger in tropical regions because in the tropics there is consistently more sunlight. When Brazilian ranchers cut down dark, tropical rainforest trees to replace them with even darker soil in order to grow crops, the average temperature of the area allegedly increases by an average of about 3°C (5°F) year-round (refs???).
Svalbard is a group of islands extending from about 75°N to 80°N in the very warmest part of the Arctic Ocean, the part that is bathed by the long stream of warm water coming from the Caribbean. In the wintertime, the coastline of most of the island remains free from ice, which gives towns on the most prominent headlands incredibly warm winters, with temperatures averaging as high as -7°C (19°F) at one weather station. But just a very short distance away, the northeastern part of the island often does freeze over. And here, the winter climate is a beautifully unstable oscillation between the very furthest reaches of the Gulf Stream penetration and the featureless cold Arctic winter. Temperatures in the northeast average about -29°C (-20°F) but have fallen as low as -57°C (-70°F). This would be unheard of in the ice-free areas of the islands.
Albedo works on a smaller scale, too. People who wear dark clothes in the summertime put themselves at a greater risk of heatstroke than if they wear white clothes.
The albedo of a pine forest at 45°N in the winter in which the trees cover the land surface completely is only about 9%, among the lowest of any naturally occurring land environment. This is partly due to the colour of the pines, and partly due to multiple scattering of sunlight within the trees which lowers the overall reflected light level. The ocean's albedo, because light penetrates, is even lower, at about 3.5%, though this depends strongly on the angle of the incident radiation. Dense swampland averages between 9% and 14%. Deciduous trees average about 13%. A grassy field usually comes in at about 20%. A barren field will depend on the color of the soil, and can be as low as 5% or as high as 40%, with 15% being about the average for farmland. A desert or large beach usually averages around 25% but varies depending on the color of the sand. [Ref for all this: Edward Walker's study in the Great Plains in the winter around 45°N].
Urban areas in particular have very unnatural values for albedo because of the many human-built structures which absorb light before the light can reach the surface. In the northern part of the world, cities are relatively dark, and Walker has shown that their average albedo is about 7%, with only a slight increase during the summer. In most tropical countries, cities average around 12%. This is similar to the values found in northern suburban transitional zones. Part of the reason for this is the different natural environment of cities in tropical regions, e.g., there are more very dark trees around; another reason is that the tropics are very poor, and city buildings must be built with different materials.
Snow albedos can be as high as 0.9. This is for the ideal example, however: fresh deep snow over a featureless landscape. Over Antarctica they average a little more than 0.8.
Because trees are such strong absorbers of heat, it seems logical that removing forests would tend to cool Earth as a whole. This is an extremely complex subject, and there are large amounts of evidence both for and against this view. Some reasons that it may not be true are the fact that in tropical parts of the world, the soil underneath trees often absorbs more heat than the trees themselves, that forested areas tend to have higher cloud cover, and thus reflect away more heat anyway, and the belief that Earth is regulating its own temperature, and thus that deforestation has essentially no lasting effect on global temperature. But the fact remains that tree-free environments where deep snow covers the ground during the winter have winter albedos 10% to 50% higher than nearby forested areas, and temperatures in the middle latitudes are as much as 11°C (20°F) colder over the barrens. Near the poles, the difference approaches zero because less sunlight is coming in the first place (add Hadley study of value of reforesting; albedo vs co2).
If a marginally snow-covered area warms, snow tends to melt, lowering the albedo, and hence leading to more snowmelt (the ice-albedo feedback). This is the basis for predictions of enhanced warming in the polar and seasonally snow covered regions as a result of global warming.
In fact, those few remaining supporters of the old Soviet plan to melt the Arctic Ocean rest their arguments largely on the fact that once the ocean is melted, temperatures there would never become cold enough for the ice to build up again. Their theory is probably false, as climate models have shown that temperatures would still be below freezing for most of the year, and then, because of albedo, would soon be below freezing all year because the ice would remain throughout the summer. (The average temperature at the North Pole in July is estimated to be about -1°C (31°F), and it is estimated that without an icecap that July would average around 6°C (42°F).)
Clouds are another source of albedo that play into the global warming equation. Different types of clouds have different albedo values, theoretically ranging from a minimum of near 0% to a maximum in the high 70s (source?). Climate models (which?) have shown that if the whole earth were to be suddenly covered by white clouds, the surface temperatures would drop to a value of about -151°C (-240°F). This model (is it singular now? they were plural above...), though it is far from perfect, also predicts that to offset a 5.0°C (9°F) temperature change due to an increase in the magnitude of the greenhouse effect, "all" we would need to do is increase the earth's overall albedo by about 12% by adding more white clouds.
It is unknown, however, whether injecting dark-colored particles into the atmosphere would have a conversely warming effect. During the Kuwaiti oil fires in 1991, temperatures in the usually hot desert kingdom cooled down by an average of 7°C (13°F) during the period of greatest obstruction. Thus it may be impossible to warm the planet in a controlled way by clouding up the atmosphere with dark-colored materials, and thereby reflecting less heat back into space.