Ensemble Forecasting ???

Initially using weather proverbs and today using super computers.

However, modelling the atmosphere and the interconnected oceans is so complicated that we still have a long way to go.

I remember many years ago when I was in Exeter someone said to me "I never watch the weather forecasts because they are always wrong".

I found difficulty seeing how if he never watched how he knew we were always wrong! I had even less chance of persuading him that if we were always wrong, if he always believed the opposite was the correct forecast, he could always be right!

Traditionally, meteorologists collect weather data twenty-four hours a day all around the world, which includes observations from land, sea and in the air. This information is fed very quickly into the super computers, which, by solving a series of mathematical equations, give the meteorologist a variety of weather maps up to a week or so ahead.


The supercomputers provide only one solution to the evolution of the weather patterns.
The supercomputers provide only one solution to the evolution of the weather patterns. This is called a deterministic forecast and gives little idea of the probability of certain events occurring. Scientists, therefore, tried to find a way of giving the meteorologist the probability of the deterministic forecast being correct; they called this an ensemble forecast.

In the atmosphere Chaos Theory reins supreme. It states that if a butterfly flaps its wings in one area of the world, it can have an effect on our weather in the UK. It is probably easier to understand, from the weather point of view, with a different example.

If a raindrop falling onto the top of the Andes in South America encounters a light easterly breeze, it will blow the rain onto the western side of the mountain. It will eventually reach the Pacific Ocean and interact with weather systems there.

If, however, there is a gentle westerly wind the raindrop will fall onto the eastern side and eventually become part of the Atlantic Ocean weather systems.

To cater for this very small change in the initial position of the raindrop above, ensemble forecasts run the deterministic forecast many times using slightly different starting points ending up with many different forecasts for a few days ahead.

While this may seem like a time consuming process, it allows the meteorologist to determine the probability of certain weather events happening.

If the results for most of the forecasts are very similar... the meteorologist has a high degree of confidence in the forecast.
If the results for most of the forecasts are very similar for a certain point in time, then the meteorologist has a high degree of confidence in the forecast. If, on the other hand, the ensemble forecast has a wide scattering of results, then the confidence in a particular forecast is much lower.

In this way the meteorologist can give the listener or viewer a probability of the forecast being correct.

Taken from BBC