Page 1
Page 3
Understanding Computer Models
Weather Models

Some of the most familiar models in our daily lives are the weather models used to predict the short term to ten day forecast. These models take observational data (such as wind speed, wind direction, air temperature, pressure, and humidity) collected from many locations and sources across a region, and use mathematical equations that represent the physics of the atmosphere to fill in the gaps between the measured points. Models then use these equations to predict what will happen in the future, including the development of storms or other weather related events.

Individual meteorological measurements of wind speed or temperature...
...can be combined to make a map of current conditions.
Models can be used to then predict what will happen in the future.

The forecasts made by weather models are most accurate for the first few days, then rapidly lose skill at lead times longer than a week. This mostly reflects the chaotic nature of weather, in which very small uncertainties in the current state of the atmosphere have a mushrooming influence on forecasts at longer time ranges. Even with a sophisticated measurement network including satellites, weather balloons, surface sites, planes and ships we cannot exactly specify every detail of today's weather, so we can never make a perfect forecast. Weather models have got much more accurate over the last few decades, but are also not perfect, adding to forecast uncertainty.

Weather vs. Climate Models

A model climate or earth system model is very similar to a weather forecast model, but with many added physical processes (such as ocean circulations, sea and glacial ice, vegetation and soil wetness, human emissions of greenhouse gases and other pollutants) to better simulate long-term climate variability and change. The climate model naturally simulates realistic storms and weather, whose statistics are used to answer climate questions.

The individual storms simulated by a climate model are not expected to be exactly the same as those that really occur, but they should have the same typical characteristics. Thus, asking climate questions take the information from weather models to look at averages. On average, how often does it rain in DC? And how much? In what months does it rain the most? Will it rain more often next year? What about ten years from now? Or 50 years from now?

For short-term climate predictions over periods of a few weeks to a few years, such as seasonal and El Nino forecasts, the model is started from our best guess at the current state of the ocean, ice, land and atmosphere. The ocean, especially, reacts rather slowly and allows such forecasts to have skill much longer than a weather forecast. For climate projections over periods of decades to centuries and longer, the model is started from a statistical mean state and spontaneously develops its own weather and climate.