If nothing is done to reduce emissions, current climate models predict a global warming of about 1.4 – 5.8°C between 1990 and 2100. These projections are based on a wide range of assumptions about the main forces driving future emissions (such as population growth and technological change) but do not assume any climate change policies for reducing emissions. Even a 1.4^oC rise would be larger than any century-time-scale trend for the past 10,000 years. These projections takes into account the effects of aerosols and the delaying effect of the oceans. Oceanic inertia means that the earth's surface and lower atmosphere would continue to warm for hundreds of years even if greenhouse gas concentrations stopped rising in 2100.

The average sea level is predicted to rise by 9 to 88 cm by 2100. This would be caused mainly by the thermal expansion of the upper layers of the ocean as they warm, with some contribution from melting glaciers. The uncertainty range is large, and changing ocean currents, local land movement and other factors could cause local and regional sea levels to rise much more or much less than the global average. Slightly faster melting of the Greenland and Antarctica ice sheets is likely to be balanced counteracted by increased snowfall in both regions. As the warming penetrates deeper into the oceans and ice continues to melt, the sea level will continue rising well long after surface temperatures have leveled off.

Regional and seasonal warming predictions are much more uncertain. Although most areas are expected to warm, some will warm much more than others. The largest warming is predicted for cold northern regions in winter. The reason is that snow and ice reflect sunlight, so less snow means more heat is absorbed from the sun, which enhances any warming: a strong positive feedback effect. By the year 2100, winter temperatures in northern Canada, Greenland and northern Asia are predicted to rise by 40% more than the global average.

Inland regions are projected to warm faster than oceans and coastal zones. The reason is simply the ocean delay, which prevents the sea surface from warming as fast as the land. The size of this delay depends on how deep any warming penetrates into the oceans. Over most of the oceans, the uppermost few hundred metres do not mix with the water beneath them. These upper layers will warm within just a few years, while the deep ocean stays cold. Water mixes down into the ocean depths in only a few very cold regions, such as the Atlantic south of Greenland and the Southern Ocean near Antarctica. In these regions, warming will be delayed because much more water needs to be warmed up to get the same temperature change at the surface.

Global precipitation is predicted to increase, but at the local level trends are much less certain. By the second half of the 21^st century, it is likely that wintertime precipitation in the northern mid- to high latitudes and in Antarctica will rise. For the tropics, models suggest that some land areas will see more precipitation, and others less. Australia, Central American and southern Africa show consistent decreases in winter rainfall.

More rain and snow will mean wetter soil conditions in high-latitude winters, but higher temperatures may mean drier soils in summer. Local changes in soil moisture are clearly important for agriculture, but models still find it difficult to simulate them. Even the sign of the global change in summertime soil moisture - whether there will be an increase or a decrease - is uncertain.

The frequency and intensity of extreme weather events such as storms and hurricanes may are likely to change. With increasing global temperatures the world is likely to experience more hot days and heat waves and fewer frost days and cold spells. Climate models also consistently show extreme precipitation events becoming more frequent over many areas and the risk of drought becoming greater over continental areas in summer. There is also some evidence to show that hurricanes could be more intense (with stronger winds and more rainfall) in some areas. There is little agreement amongst models concerning changes in mid-latitude storms. There are also other phenomena, such as thunderstorms and tornadoes, where knowledge is currently inadequate for making projections. However, models still cannot predict how. The models used to simulate climate change cannot themselves simulate these extreme weather events, so the evidence is indirect. There is some concern that patterns of extreme weather may change because the models predict changes in ocean surface temperatures and other factors that are known to affect storm and hurricane development. However, it will be many years before scientists can predict whether individual regions will become more or less stormy.

Rapid and unexpected climate transitions cannot be ruled out. The most dramatic such change, the collapse of the West Antarctic ice sheet, which would lead to a catastrophic rise in sea level, is now considered unlikely during the 21^st century. There is evidence that changes in ocean circulation having a significant impact on regional climate (such as a weakening of the Gulf Stream that warms Europe) can take place in only a few decades, but it is unknown whether or not greenhouse warming could trigger any such change. Climate models that do show a weakening in the Gulf Stream still project warming over Europe.