Interesting story in today's Philadelphia Inquirer, especially with how the
Weather Channel obtains its temperatures.
Gauging the temperature: Degrees of uncertainty
By Anthony R. Wood
Inquirer Staff Writer
A chilly rain is falling on a late-autumn morning. The temperature is 33 in
Center City, perilously close to freezing, or so says your favorite Web
site. But, wait, the radio says it's 35; weather.com, 32; AccuWeather, 34;
WeatherBug, 31; the car dashboard, 36; and don't even bother looking at
that
bank thermometer.
So which would be correct?
Perhaps none would be exactly right. The simple question - what exactly is
the temperature? - turns out to be a complex one to answer. For example,
just before 8 on Saturday, the chilliest morning of the season, the
weather.com reading for Center City was 32, AccuWeather's was 34, and
WeatherBug's was 39.
Never have so many thermometers prodded the atmosphere over the United
States of America, and never have so many quibbled with one another.
And while the issue of a few degrees might seem less than monumental in
these troubled times, in the age of global warming, the climate community
frets mightily over it.
"It's the kind of thing that we climatologists think about all the time,"
said Nolan Doesken, president of the American Association of State
Climatologists, "but most people think we're nuts for thinking about it."
A veritable blizzard of temperature measurements confronts the consumer
daily, hourly, even minutely - on TV, on radio, online, on bank buildings
(very suspect), on car dashboards. They come from sources reliable and
unreliable. On any given morning, you could get seven different opinions on
the temperature outside before the first jolt of caffeine takes hold.
"It's out of hand - and wonderful," said Doesken, also the state
climatologist for Colorado. Even poor data, Doesken said, are useful for
checking against good and bad sensors.
In figuring out the global temperature to within tenths of a degree
Fahrenheit, it is up to the climatologists to figure out which readings are
valid and which to disregard, and how to fill in the gaps all over the
planet.
The keepers of the world-temperature databases have done a decent job of
prevailing above the chaos, said David Robinson, the New Jersey state
climatologist and measurement expert. But he says the proliferation of
conflicting readings has supplied fresh fuel for those who question just
how
much the Earth is warming, and how accurately it is being measured. "It's
fed right into the naysayers," he said.
Most thermometers today are electronic devices that produce a digital
reading by measuring the resistance to air temperature. Generally, they are
fine on the technical end, Doesken said, but they too often are badly
sited,
set up poorly, improperly maintained, or all the aforementioned.
A common issue is excessive sun exposure that can make thermometers jumpy,
said Bruce Rose, senior meteorologist at the Weather Channel, which has
attempted to smooth out inconsistencies with its own temperature program.
The universe of quality varies from the elite government stations to the
suspect systems used at the local bank. Usually, the bank sensors work,
Doesken said, but they tend to be located above warm, paved surfaces and
places where the air cannot circulate.
He and his colleagues hope banks do a better job with money than with
weather.
"Bank thermometers were always a source of entertainment," he said. "We
sort
of liked them. It made us laugh. It couldn't possibly be 112."
The most expensive and reliable thermometers are the automated sensors at
about 1,500 government stations, including the one at Philadelphia
International Airport. Unfortunately, that network is spread out widely,
and
as we all know, no one lives at the airport.
Those readings are supplemented by other government sensors and readings
from cooperative observers.
In recent years, the private-instrument universe has expanded
exponentially.
In large degree, that's the work of AWS Convergence Technologies Inc., a
Maryland outfit better known as WeatherBug, which has become a national
sensation.
In September, 23.5 million people visited the WeatherBug site, nudging both
weather.com and Twitter, according to Andrew Lipsman of the Internet
tracking firm comScore, based in Chicago.
WeatherBug has set up a network of several thousand sensors across the
country, including 284 in the Philadelphia region, said company president
Mark Hoekzema, formerly a meteorologist at a Washington television station.
Most of them are at schools, but some are the source of those "game time"
temperatures at sporting events. The sensors transmit data, and the
readings
are available with the click of a mouse, or to anyone who downloads the
WeatherBug program.
He said a national staff of about 25 tries to keep track of all the
instruments, but it's next to impossible to keep all of them in constant
working order. "Over time, instruments do go bad," he said. "There might be
a wiring issue. We have a limited number that get damaged due to weather."
Doesken and Robinson said the network has some siting problems because so
many of the instruments are atop roofs and susceptible to building heat.
"It's difficult to find the perfect site," Robinson said.
Rather than having its own instrument network, the Weather Channel relies
on
a system called HiRAD, for high-resolution aggregate data. When someone
punches in a zip code on weather.com, the temperature that appears is based
on an average of available readings in the vicinity, tweaked for factors
such as elevation.
WeatherBug directs the user to the company's nearest actual measuring site.
"We're giving real data," Hoekzema said. "There's nothing interpolated or
computer-generated."
A WeatherBug station is at Lincoln Financial Field, and another stadium
WeatherBug became the source of controversy a few weeks back when the
Phillies were on their way to the World Series.
On the night of Oct. 11 in Denver, as the Phils were getting ready to play
the Colorado Rockies, the official scorer posted a game-time temperature of
35. That set off yowls of protests in the press box. It had to be colder
than that, skeptics complained. Where was that thermometer - inside a
hot-dog bun?
The scorer, Dave Einspahr, went with the reading from the WeatherBug
thermometer, 84 feet above street level at Coors Field. Einspahr also
happens to be an official National Weather Service observer, and he said he
always checks other thermometers on the way in to make sure the WeatherBug
reading is in the ballpark, so to speak.
He logged the temperature for the official box score about 10 minutes
before
game time, he said, recalling that it was 34.5, which rounded to 35. Other
thermometers in the vicinity showed readings from 29 to 34.
Evidently, however, the temperature was dropping quickly; the WeatherBug
archive shows it falling to 30.7 by 9 p.m. That might explain the chill
that
prompted the press-box protest.
Even if the reading was off a tad, it was still cold. But Doesken noted
that
with winter approaching, precision of the thermometer you're consulting can
be critical.
Say rain, snow, or that ever-popular "wintry mix" is falling. If the
temperature from the source you're consulting reads 34, and it's actually
31, Doesken said, "it may have a big effect on your day."
http://www.philly.com/inquirer/front_page/20091109_Gauging_the_temperature__