The longer the time span, the more inaccurate the weather forecast. Why?

The longer the time span, the more inaccurate the weather forecast. Why?

Chaos is a common phenomenon in weather system. It is inherent randomness in deterministic systems, and its external performance is very similar to pure random motion, that is, unpredictable. Chaotic systems are very sensitive to small initial value changes or disturbances. No matter how small the disturbance is, it will completely deviate from the original evolution direction after a long time. Therefore, it is difficult to make an accurate forecast of the long-term weather process.

Its external performance is very similar to pure random motion, that is, unpredictable chaotic systems are very sensitive to small initial value changes or disturbances. No matter how small the disturbances are, they will completely deviate from the original evolution direction after a long time. The Meteorological Observatory also produces medium and long-term weather forecasts. For the public, short-term weather forecast is enough, but the main reason is that the accuracy of medium and long-term weather forecast is much worse than that of short-term weather forecast.

Long-term weather forecast has great uncertainty. Internationally, the weather forecast within 10 days has a certain accuracy at present, but the forecast over two weeks, that is, the long-term weather forecast, is very uncertain. This is because chaos is a common phenomenon in weather systems, and it shows an inherent randomness in deterministic systems. Its external performance is very similar to pure random motion, that is, unpredictable. Chaotic systems are very sensitive to small initial value changes or disturbances. No matter how small the disturbance is, it will completely deviate from the original evolution direction after a long time.

This feature of butterfly effect weather system was first discovered by Lorenz, a professor of meteorology at MIT. Lorenz has been using computers to simulate the weather process, so as to carry out numerical weather forecasting experiments. One day in 196 1, in order to save trouble, he kept the original initial data with 6 decimal places and only used 3 digits instead of calculation. He was surprised when he came back from a cup of coffee. At first, the calculation results were not much different from those before, but the later the calculation results were different, so that it was 108 thousand miles later. After checking every detail in the calculation, it is found that the small differences in the initial data will accumulate in the calculation and grow at an extremely fast speed, eventually resulting in huge differences.

In 1963, Lorenz expressed this conclusion as the butterfly effect, that is, "a butterfly flapping its wings occasionally in the Amazon rainforest of South America is likely to trigger a tornado in Texas within two weeks." Therefore, it is difficult to make an accurate forecast of the long-term weather process.