Thursday, November 1, 2012

Earth Scientists Use Fractals To Measure And Predict Natural Disasters

Predicting the size, location, and timing of natural hazards is virtually impossible, but now, earth scientists are able to forecast hurricanes, floods, earthquakes, volcanic eruptions, wildfires, and landslides using fractals. A fractal is a mathematical formula of a pattern that repeats over a wide range of size and time scales. These patterns are hidden within more complex systems. A good example of a fractal is the branching system of a river. Small tributaries join to form larger and larger "branches" in the system, but each small piece of the system closely resembles the branching pattern as a whole.

At the American Geophysical Union meeting held last month, Benoit Mandelbrot, a professor of mathematical sciences at Yale University who is considered to be the father of fractals, described how he has been using fractals to find order within complex systems in nature, such as the natural shape of a coastline. As a result of his research, earth scientists are taking Mandelbrot’s fractal approach one step further and are measuring past events and making probability forecasts about the size, location, and timing of future natural disasters.
"By understanding the fractal order and scale imbedded in patterns of chaos, researchers found a deeper level of understanding that can be used to predict natural hazards," says Christopher Barton, a research geologist at the United States Geological Survey, "They can measure past events like a hurricane and then apply fractal mathematics to predict future hurricane events."
In the past, earth scientists have relied on statistical methods to forecast natural hazard events, but when Barton used fractals, he found that these patterns contain a level of information that has never been seen using statistical methods. Barton discovered that by comparing the fractal formulas of the size and frequency of a hurricane’s wind speed to the historic record of information about past hurricane landfall location and timing that he was able to predict the approximate wind speed of the hurricane when it made landfall at a given coastal location along the United States Atlantic and Gulf of Mexico coasts.
Forecasts of hazardous natural phenomena based on the application of fractals are now available to government agencies responsible for planning and responding to natural disasters such the Federal Emergency Management Association and other emergency personnel to be able to better forecast the size, location, and timing of future events. "Based on the fractal patterns seen over the past 100 years," says Barton, "We can better forecast the probability of a future event."