Weather is the archetypical example of a chaotic system. Indeed, it was his calculations of weather models that led Edward Lorenz to his landmark 1963 paper that helped launch modern chaos theory. (See the article by Adilson Motter and David Campbell, Physics Today, May 2013, page 27.) Lorenz famously gave a 1972 talk titled "Predictability: Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” Chanh Kieu and Zachary Moon at Indiana University Bloomington now report on a related question: How well can we forecast the intensity of a hurricane? For steady conditions, Kerry Emanuel has shown that thermodynamic arguments yield a wind speed—the standard measure of hurricane intensity—that's a reasonable upper bound for a majority of observed storms (see Emanuel's Quick Study, Physics Today, August 2006, page 74). Using a minimal dynamic model, Kieu and Moon found that the limiting speed is stable: Simulated hurricanes approach that equilibrium value regardless of their initial conditions. Yet the errors in forecasts of hurricane intensity don't go away as one might expect; rather, speed errors level off at about 8 m/s (18 mph) after four to five days in real-time intensity forecasts. Through full-physics simulations, the pair discovered that those errors arise from a so-called chaotic attractor at the maximum potential intensity limit. That finding implies that improvements in intensity forecasts are more likely to come from better modeling of the large-scale environment than from better knowledge of the storm's initial state. Even so, the researchers note, the maximum range of predictability is only about three days, and likely shorter for mature hurricanes. (C. Q. Kieu, Z. Moon, Bull. Amer. Meteorol. Soc., in press, doi:10.1175/BAMS-D-15-00168.1. Hurricane Ike image courtesy of NOAA Environmental Visualization Laboratory.)
