So how much natural gas is in the Marcellus Shale?

The U.S. Geologic Survey on Tuesday estimated the formation contains 84 trillion cubic feet (Tcf) of natural gas, up from a mere 2 Tcf in 2002. Headlines exploded across the Internet: "Federal report boosts Marcellus Shale estimate."

But on Wednesday another federal agency, the U.S. Energy Information Administration, which just a month ago estimated the shale contained 410 Tcf, announced it was revising its number downward in response to the USGS estimate. New headlines: "U.S. Slashes Marcellus Shale Gas Estimate 80%."

Up? Down?

For adversaries in the increasingly politicized and polarized world of shale gas, the USGS's new assessment cuts both ways.

Anti-drilling activists said the EIA's downward revision supported their view that the industry has hyped the new discoveries to generate political and investor excitement.

"I remain concerned about the processes which lead to the original estimates, and I have additional questions about how this change will impact the outlook for shale gas," U.S. Rep. Maurice Hinchey (D., N.Y.) said in a statement.

But the Marcellus Shale Coalition, an industry trade group, touted the USGS's upward revision as further proof of the abundance of shale gas.

The issue is important because of the growing controversy about shale gas, which the EIA says accounts for about a quarter of the nation's natural gas production. The nation consumes about 25 Tcf a year, mostly for heating and power production.

The Securities and Exchange Commission and the New York State Attorney General's Office are investigating industry estimates of gas reserves, which are more optimistic than the federal projections.

Indeed, during recent sessions with investment analysts, four big Marcellus operators - Chesapeake Energy Corp., Range Resources Corp., Ultra Resources Inc., and Cabot Oil & Gas Corp. - estimated their combined 2.9 million acres contain 76 Tcf, nearly as much as the USGS estimates for the entire formation.

The EIA says it is waiting to set its estimate once the USGS provides more information about its assessment to understand where the agencies diverge. "We will not be able to be more precise until that work is completed," said Jonathan Cogan, an administration spokesman.

Even at 84 Tcf, the Marcellus still contains a lot of gas, more than any of other shale-gas plays, according to the USGS.

Just three years ago, Pennsylvania State University professor Terry Engelder and a colleague, Gary Lash, estimated the Marcellus Shale could contain as much as 50 trillion cubic feet of recoverable gas, a number so astonishing that it triggered a land rush.

Engelder later increased his estimate to 363 Tcf and then nearly 450 Tcf, based upon actual production data.

The Marcellus Shale Coalition argues that the USGS numbers are low because its methodology discounts undeveloped parts of the shale.

"Hence, during early development of a gas shale play when there is very little production data anyway, the USGS numbers will be commensurately low as is the case now," said Travis Windle, a coalition spokesman.

USGS says there are many reasons that assessments might disagree - the use of different data, or proprietary information. Doug Duncan, associate program coordinator of the USGS's energy resources program, said the agency only makes its assessment after observing reliable production data over at least a 30-month period.

"We don't have a preconceived idea about what kind of answer we want to get," he said. "We try to get it right."

Without mentioning other estimates, the USGS asserted its primacy on the issue in its announcement Tuesday.

"USGS is the only provider of publicly available estimates of undiscovered technically recoverable oil and gas resources of onshore lands and offshore state waters," it said.