Predicting future climates

Predicting the future has never been easy. It is clear that adding more greenhouse gases to the atmosphere will make the planet warmer, but how much warmer will it get, and what effects will this have on the world? There are a lot of factors to consider, and scientists try to take account of them all by building them into mathematical "models"of the climate, and using computers to see what happens when the figures for greenhouse gas emissions are increased. The resulting projections agree that if we do nothing to stop climate change, temperatures could rise by 5.4°F (3°C) or more by 2100—with potentially serious results.

Air in grid boxes interacts horizontally

Atmosphere divided into 3-D grid boxes, each with its own microclimate

Influence of_

vegetation and terrain included at surface

Water in oceanic _ grid boxes interacts horizontally and vertically

Air in grid boxes interacts horizontally

Atmosphere divided into 3-D grid boxes, each with its own microclimate

Air in grid boxes interacts vertically


Computers have been used for many decades to forecast the changes in the atmosphere that control day-to-day weather. Doing the same for long-term climate change is more difficult, because there are extra factors to consider such as changes in vegetation and ice cover. But supercomputers like these can process vast amounts of data, and every year the computers become more powerful and, more importantly, our understanding of climate improves.

Air in grid boxes interacts vertically

_ Oceanic grid boxes model currents, temperature, and salinity


A climate model is a computerized representation of the atmosphere based on a three-dimensional global grid. This is linked to submodels that represent other global systems such as the oceans, as shown here, or vegetation growth. The computer program allows some factors to be changed, such as the amount of carbon dioxide in the atmosphere or the degree of pollution from airborne particles or aerosols. The model then applies these changes to its virtual world, to see what effect they have on the climate.


When it is run on the computer, a climate model evolves in a series of short time steps, often of less than an hour each, and each step may take a few seconds to process. So achieving a projection to the end of the century takes several weeks, even when using one of the most powerful supercomputers. The results can then be used to generate graphs and images that show how temperatures and rainfall might change in different circumstances. This three-dimensional image has been generated from a model of changing global sea temperatures.

Scenario 1

Scenario 2

Scenario 1

Scenario 2


Each climate model is based on a different"scenario." These cover the various ways in which the world might develop. The basic scenario assumes steady economic growth and a corresponding increase in greenhouse gas emissions. Others might include a major switch away from them. These globes show predicted temperature rises by 2090-2099 based on two scenarios: a less industrialized world than today's world (1) and a world with rapid economic growth (2).


0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5


CONFIDENCE AND UNCERTAINTY Climate projections are a bit like long-range weather forecasts. However, they focus on the long-term averages of the whole system and how these change, rather than the day-to-day local variations that make detailed weather forecasting so difficult. So even if we cannot forecast the weather for next week, we can be confident that continued greenhouse gas emissions will lead to higher temperatures at the end of this century. But since it is hard to take account of all the variables, different modeling labs come up with different projections. This graph shows eight projections made by seven labs, using the same basic scenario that little is done to reduce greenhouse gas emissions. 1900


Deep Atlantic Ocean water




Each climate model takes weeks to run through to the end of the century and produce a climate projection. Models based on four or five scenarios take four or five times as long, and because there are so many different variables, there are thousands of possible scenarios. Faced with this problem, even the biggest supercomputers run out of capacity. In 2003 UK climate analyst Myles Allen came up with a solution—to offer thousands of home or school computer users a simplified version of a climate model that runs on a standard PC. Each user downloads a model based on a slightly different scenario. It runs as a background task whenever the computer is switched on, and can even display its current results as a screensaver. Running each model to the end takes several months, but since thousands of computers are working at once, they generate a lot of results more quickly than a few overloaded supercomputers.

-0.5 -0.4 -0.3 -0.2 -0.1 0.0 +0.1 +0.2 +0.3 +0.4 +0.5







0 0

Post a comment