Climate models attempt to predict the future. They work like this. Climatologists must first consider umpteen climate variables which play a part in affecting long-term weather patterns (eg variations in the concentration of greenhouse gases in the atmosphere, the cloud cover, humidity, how heat energy from the ocean interacts with the atmosphere). These parameters are then fed into a computer model of the atmosphere, built from standard physics principles.
When all the parameters have been entered, the model begins number-crunching, letting every variable interact and 'stew'. The model typically calculates the outcome for a 30-minute interval. But obviously, predicting the climate 30 minutes from now is pretty pointless, so the computer crunches for 30-minute intervals over a longer simulated period of time, typically 30 years. It advances the new data from the last 30-minute run into the next 30-minute run, and so on. Hence the need for powerful computers – a 30-year time period means over half a million 30-minute solutions for every location on the planet! At the end of the simulation, climatologists can analyse what effect occurred over specific areas, known as grids.
When all the parameters have been entered, the model begins number-crunching, letting every variable interact and 'stew'. The model typically calculates the outcome for a 30-minute interval. But obviously, predicting the climate 30 minutes from now is pretty pointless, so the computer crunches for 30-minute intervals over a longer simulated period of time, typically 30 years. It advances the new data from the last 30-minute run into the next 30-minute run, and so on. Hence the need for powerful computers – a 30-year time period means over half a million 30-minute solutions for every location on the planet! At the end of the simulation, climatologists can analyse what effect occurred over specific areas, known as grids.
No comments:
Post a Comment