Philadelphia's snow is measured in National Park, Gloucester County, 3 miles from the old site, by an observer who declined to be interviewed or even identified. (In fact, the observer told the weather service that if the duties included talking to a journalist, the agency could reclaim its ruler. No offense taken.)
Among reports from 34 counties in Pennsylvania, New Jersey, Maryland, and Delaware, the airport/National Park site was the only one to log two 8-plus snowfalls - 8.4 on Feb. 2-3 and 8.3 on March 3-4.
The reasons for the change were complicated - and a source of consternation to some of those keeping track of climate change. But it was part of a national trend, and in Philadelphia, an overriding issue was an obvious one.
Snow totals were coming from the roof of a terminal building, "a pretty atrocious way to measure snow," said Gary Szatkowski, head of the weather service's regional office in Mount Holly.
The weather station was moved to the airport in 1940. At first, the snow could be measured on grass, he said, but the airport became ever more paved over.
Over time, the observers were forced to measure on the roof, and that became difficult when the snow wasn't sticking well or it was windy. So they sometimes estimated the snowfall by using a formula based on the temperature and the snow's liquid content.
"It's an accepted method," said Jim O'Sullivan, a weather service meteorologist and data-management specialist in Silver Spring, Md. "It's not the preferred method."
In 1995, the weather service installed an automated measuring device, but the gadget couldn't measure snow. So the agency continued to pay the observers and used them even after the Federal Aviation Administration took over the contracts several years ago.
Ultimately, the weather service decided to find its own snow observer. It advertised for a reliable person who lived near the airport and had measuring places away from trees and roofs. It found the right candidate in National Park, and the change took effect three winters ago.
Szatkowski said that while measuring snow "is not real high-tech," he is satisfied that the observer is following government standards, which include taking readings every six hours and averaging several measurements when it's windy.
Those methods are likely to yield different results from liquid estimates, said David A. Robinson, a Rutgers University professor and snow specialist.
"With snowfall, it's the measurement practice that makes the difference," Robinson said.
"I wouldn't disagree with that," Szatkowski said, but he added that any differences between the airport and National Park should be minimal. "Even if you put two observers on the same field, you would get different results."
Contact staff writer Anthony R. Wood at 610-313-8210 or twood@phillynews.com.
