Study Suggests Weekly Hurricane And Rainfall Patterns Are Linked To East Coast Pollution

Examining some basic data sets in a way that has never been tried before, ASU climatologists Randall Cerveny and Robert Balling, Jr. have found proof for what many a weekend boater has secretly suspected: rain is most likely to occur along the Atlantic coast on the weekend and the weather is most likely to be better on a Monday, Tuesday or Wednesday. The most obvious culprit is the "natural" cloud-seeding effect created by the massive drift of East Coast pollution, which also follows a well defined weekly cycle.

The gray, smelly cloud of pollution has a strange silver lining, however. While pollution makes for more rainy weekends, it also apparently reduces the intensity of hurricanes that hit over the weekend, such that weekend hurricanes tend to be much weaker than, say, Tuesday storms.

"Hurricanes are the biggest storms that we have on this planet, in terms of energy and precipitation," noted Cerveny. "And what we've found is that we're having an impact on them. It's a little daunting, when you start to think about it."

Cerveny and Balling examined and compared three different data sets -- daily carbon monoxide and ozone measurements from a Canadian monitoring station on Sable Island off the coast of Nova Scotia, daily satellite-derived rainfall data for the Atlantic Ocean, and databases of coastal Atlantic hurricane measurements. In each case, when the two ASU scientists examined the data by day of the week, they found significant differences between days, and similar patterns of variation, with pronounced differences between beginnings and the ends of weeks. All three sets of climate data revealed a seven-day cycle.

"The human week is not a natural time period," said Balling. "Human effect on weather is the only explanation."

"If you're going to go out boating in the Atlantic, you're going to get wet if it's a weekend," Cerveny said. "And what we suggest is that this is probably linked to the pollution cycle."

In examining precipitation in the Atlantic, they found no daily variation when looking at the ocean as a whole, but a pronounced sine-wave pattern of variation for just the coastal areas, with average daily precipitation rising on Thursday and into the weekend and then dipping from Sunday through the middle of the week. Balling notes that when the team analyzed satellite data grid cells for an area a little further away from the coast, they found the same pattern, time-shifted in accordance with the rate of pollution drift.

Though the study does not directly address causation, a comparable fluctuation in the levels of East Coast air pollution points to an obvious connection. The fact that coastal hurricane intensity data taken from 1945 to 1996 follow a similar pattern (rather than being statistically uniform for each day of the week, as one would expect), supports this hypothesis.

"The fact that pollution can affect rainfall is actually well understood," said Balling. "We just had to look for the evidence in the right place. The hurricane data, though, surprised the heck out of me."

"We knew that cities have an effect on local weather with urban heat islands and so forth, and people are pretty sure that we're having a general global effect with carbon dioxide," said Cerveny. "But nobody had ever looked at the in-between area of large-scale regional weather. We appear to be affecting global weather on a scale that is comparable to El Nino."

The hypothesis is particularly important when applied to hurricanes, because of the destructive potential of the storms. Cerveny and Balling looked at 50 years worth of hurricane records, which include observations taken every 6 hours and found surprising statistical differences with important implications.

"Storms are substantially weaker during the first part of the week and stronger in the last part of the week," said Cerveny. "Pollution's thermal changes on the storm are apparently helping hurricanes blow themselves out. The difference is as high as 10 miles-per-hour wind speed, downgrading the storm almost as much as a Saffir-Simpson Scale category, meaning that if a hurricane were to hit on Tuesday or Thursday, it might be a Category 3, but if it were to hit on a Saturday or Sunday, it might be only be a Category 2."

Cerveny notes that the effect is similar to a weather-control method once attempted in a military experiment: "Back in the 1960s, the military had a project called Project Storm Fury that was developed to cloud-seed hurricanes. What we're suggesting here is that they were on the right track, but they just didn't do it on a large enough scale. We're looking at the combined pollution from the entire eastern seaboard -- that's what it takes to influence a hurricane."

Though the study has interesting implications, what most surprised the ASU scientists was the fact that no other researcher had ever attempted to analyze these major data sets in such a basic way.

"Interestingly, no one had ever looked at this pollution data from a daily standpoint before and, curiously, nobody has bothered to look at seven-day cycles in the weather data," said Cerveny. "Oftentimes the most fundamental research is that way you say to yourself 'why didn't anybody look at this' When we were putting this together, we went through every journal we could find, saying 'somebody has got to have done this before!' Luckily for us, no one had."

Cerveny and Balling's study can by found in the August 6 issue of Nature. A simplified version can also be found on the World Wide Web at http://www.asu.edu/clas/geography/nature/.