Chaos theory and Michael Fish | Tim Palmer

Quick Overview

The speaker explains that the limitations of long-range weather forecasting, exemplified by Michael Fish's famous prediction failure before the 1987 storm, are due to the inherent chaos in atmospheric systems, leading to the modern probabilistic approach in meteorology, where multiple model runs are used to quantify uncertainty.

Key Points: Michael Fish famously told viewers not to worry about a hurricane on the evening TV forecast before the devastating 1987 storm hit Southern England, causing billions in damage. The speaker attributes the difficulty in long-range deterministic forecasting to the work of Ed Lorenz and the development of Chaos Theory in the 1970s/1980s. Chaos theory demonstrates that even minuscule differences in initial conditions (like a butterfly effect) cause predictions of chaotic systems, such as weather, to diverge completely over time. The visual representation of the Lorenz attractor shows how two nearly identical starting points eventually diverge and become uncorrelated. The solution to this inherent uncertainty in weather prediction is to shift from deterministic forecasts (a single answer) to a probabilistic science, running many models to gauge the range of possibilities. Modern weather forecasts, like those shown for the 1987 storm probability, use ensemble forecasts (e.g., 51 members) to show the probability of hurricane-force winds, replacing the old single-prediction philosophy. The speaker argues that this probabilistic approach correctly reflects the nature of the laws of physics governing chaotic systems, showing how small-scale variations affect large scales.

Context: The video features a speaker, likely a meteorologist or climate scientist, discussing the historical context and theoretical underpinnings of modern weather forecasting, specifically referencing the famous incorrect forecast by UK weatherman Michael Fish before the Great Storm of 1987. The discussion centers on how the mathematical concept of Chaos Theory, pioneered by Edward Lorenz, fundamentally changed the practice of meteorology from attempting single, precise long-range predictions to embracing probability.