Monday, 21 February 2011

The Global Cone of Uncertainty

When weather forecasters predict the path of a hurricane, they use what is known as a "cone of uncertainty" to show the area where the storm could possibly hit. The cone gets wider as it moves farther from the hurricane's present location.
Technology has made us very good at measuring current weather conditions, but our ability to forecast storms still has room for improvement. Theoretically, if we could measure every element affecting the weather at a given moment, we could reliably extrapolate into the future. But given the complexity of the atmosphere, there are always a few factors that are either unaccounted for or represented by estimates and assumptions. This introduces uncertainty in our calculations. With time, the effects of these factors multiply, so the cone representing possible outcomes must expand.
When it comes to short-term weather trends, we all have a pretty good grasp of this concept; the further in the future we try to predict, the less reliable the prediction will be. We understand that if we're planning an outdoor gathering, we should keep checking the forecast as the day approaches to find out if we'll need to set up a tent.
Yet, when it comes to long-term trends, where the greater number of potentially important factors and the longer time spans make projections even less certain, we often view the world as being more predictable than it really is. Especially when doing so generates good headlines.
After last year's the oil spill in the Gulf of Mexico, the press was quick to pick up on claims that oil could soak the entire eastern coastline, washing shores in an inky tide. The New York Daily News noted, with barely contained excitement, that one government report "included a dramatic computer animation showing tendrils of oil coiling like a nest of snakes in the Gulf of Mexico, then winding around Key West into the Gulf Stream and shooting up to North Carolina by July before flowing east out to sea toward Europe."(1) The National Center for Atmospheric Research, which produced the report, cautioned that it was based on a simulation involving dyed water, not oil. But this didn't prevent the Daily News from running a brightly colored map with the caption "A map predicting just how far north and how [far] into the Atlantic the oil spill could reach."
Those predictions of an oily Armageddon never came true. Much of the oil was cleaned up by deep-sea microbes before it ever reached the surface. Due to prime conditions, the bacteria multiplied rapidly, and within 24 hours of the spill, the number of oil-eating bacteria around the wellhead had grown tenfold. John Kessler, a chemical oceanographer at Texas A&M University told The Christian Science Monitor that the speed of the microbial activity "was a surprise for us."
If we understand so clearly that weather forecasts for a week from now have a large margin for error, why do we treat climate predictions for decades into the future as though they are infallible prophecies? I have written here before about our tendency to assign more weight to predictions of global climate consequences than seems to be justified by scientists' demonstrated skill at forecasting the state of the atmosphere or of the seas.
This view does not make me a skeptic of the global climate change that has already occurred, which can be measured with some precision. But I think it makes me realistic about the limitations of those measurements, and of our ability to forecast future changes with any useful degree of reliability.
The world is, of course, going to change. It always has. The ancient city of Alexandria, which once boasted one of civilization's finest university systems and a 440-foot-high lighthouse that was a marvel of the world, is now accessible only to those with scuba gear. According to Jean-Daniel Stanley, a geologist with the Smithsonian Institution's National Museum of Natural History, the city slid into the sea over the course of centuries. Earthquakes and a tsunami speeded the process, but the gradual subsidence of the land and rise of the sea also played a role.
A more recent story concerns the orange groves of St. Augustine, Fla. The trees were planted by Spanish settlers and flourished there for many years. In the early 1800s, the area surrounding St. Augustine exported between 2 million and 2.5 million oranges each year. But a series of freezes culminating in the 1890s pushed commercial citrus cultivation south to the Orlando area. More freezes in the 1980s sent growers even further south. Now only a few intrepid gardeners raise oranges in St. Augustine. The nearest commercial groves are nearly 200 miles away, in warmer climates.
Yet, in the mid-2000s, when warm weather plants started to creep northward, horticulturists and the media responded as if such a change was unprecedented and reported the shift as a sign of long-term global warming. The National Wildlife Federation released a report entitled "The Gardener's Guide to Global Warming" that boldly predicted that by the end of the century, the climate would no longer be favorable for the official state tree or flower in 28 states.
It was, therefore, a breath of fresh air to read a recent article in The New York Times on just how "topsy turvy" the weather can be.(2) The article discussed the past two years of heavy snows in the Mid-Atlantic and northeastern United States and in western Europe, which occurred while winter warm spells gripped northeastern Canada and Greenland. Such simultaneous warm and cold spells across regions undermine claims that global climate change is simple and unidirectional. Unpredicted cooling in regional weather patterns has taught scientists and journalists to be more cautious in labeling similar warm periods elsewhere as evidence of "global warming."
Global warming is certainly occurring, as it has since the peak of the last Ice Age. Human activity most likely plays a role. The point is that we do not yet understand exactly how warming is occurring, or how it might affect the weather in the future.
John M. Wallace, an atmospheric scientist at the University of Washington who wrote some early papers predicting that warming would continue uninterrupted, told the Times that he will not be so eager to publish on minimal evidence in the future. "Nature," he said, "has a way of humbling us."
(1) New York Daily News: BP Oil Slick Could Hit East Coast In Weeks: Government Scientists
(2) The New York Times: Cold Jumps 'Fence,' Stoking Winter's Fury
For more articles, please visit the Palisades Hudson Financial Group LLC newsletter or subscribe to the blog.