An unexpectedly small gas storage injection of 53 Bcf reported last week by the Energy Information Administration (EIA) for the week ending Aug. 22 initially sent gas futures prices climbing higher last Thursday, but the market then reversed course partly on speculation that the agency might have to revise its storage numbers upward.

The weekly storage injection came in well below even the lowest predictions and increased the storage deficit compared to levels last year for the first time in months. It fell short of last year’s injection during the same week, and the five year average of injections. Part of the reason for the small injection apparently was related to the sharp increase in demand from gas fired generators that were called upon to fill the void left by downed nuclear units and coal plants following the Aug. 14 blackout. Tropical Storm Erica also wiped out at least 334 MMcf of gas production, possibly much more, in the Western Gulf at the beginning of the storage week.

However, some observers have been concerned about a growing discrepancy between EIA’s weekly estimated storage numbers and the actual storage numbers from EIA’s monthly data for earlier this year.

As the EIA completed its monthly data for February, March and April, it became clear that the weekly data estimates for those months were falling short in increasing amounts each month. EIA’s weekly numbers fell short of its complete monthly data by 26 Bcf for the end of February, by 29 Bcf at the end of March and by 85 Bcf at the end of April, according to EIA’s Bill Trapmann.

Observers now wonder whether those differences increased over the summer and whether the market should be looking for a substantial upward revision to weekly storage levels, which show 2,319 Bcf of working gas in storage as of Aug. 22, or about 187 Bcf less than the five-year average and 397 Bcf less than at the same time last year.

“If, and that’s a big ‘if,’ any revision to the weekly numbers does occur, that revision is likely to be an increase in inventory levels rather than a decrease,” CitiGroup analyst Kyle Cooper said in a recent note to clients.

But Trapmann, who manages the weekly storage survey, said there’s no reason to be alarmed; no revisions are expected and the discrepancy probably worked itself out over the summer.

“Conceptually [the monthly and weekly numbers] should line up, but in practice, given that one is a sample survey involving some estimation, it isn’t guaranteed that you would get the same number as you do for the monthly,” he said in a recent interview with NGI.

Trapmann said it also isn’t correct to assume that the current weekly storage estimate is lower than actual storage levels. “I don’t think so. We’ve done some analysis to date to look at this, and the seasonal differences that we think affected our estimation process are probably transitory.” He said the differences between the old American Gas Association weekly series and the EIA’s month data often peaked around March and April and then diminished.

“Someone asked me if they should simply add 85 Bcf (the difference at the end of April) to the current weekly total,” said Trapmann. “I don’t think that would be appropriate. There is a very good chance that the difference has worked its way through.”

However, he noted that the estimated standard error in the latest weekly report is 52 Bcf.

Trapmann also noted that some of the April weekly numbers also probably should have been 11 Bcf/week higher due to a reclassification of base gas to working gas that occurred sometime during that month but that was reported to EIA by storage operators much later and then incorporated into the weekly survey in June.

There also may be some methodological reasons for the discrepancies that may need to be worked out, he said. “We feel that while the weekly storage methodology probably performs well and has performed well in more normal circumstances, this past winter was rather unusual with the tremendous draw downs. The methodology is predicated on the working assumption that [storage] firms will tend on average to perform similarly with respect to their uses of storage.

“We’ve cross checked the individual submissions between the two surveys and they are consistent for the information we received, so we believe the issue is within the methodology [for the weekly survey]. We are looking into [whether changes are needed]. Before we would introduce any significant change to the system, it would certainly be announced in advance to allow people to prepare themselves for it,” he said.

Ideally, EIA would like to refine its methodology for the weekly survey so that it could capture changes in the industry/market as they happen, according to Trapmann. “I’m sure we all hope that the events of the past winter don’t repeat themselves with these massive drawdowns, but if they do we will have to be prepared.”

However, he added that the EIA certainly wants to avoid making any changes to the storage methodology that would lead to frequent revisions.

©Copyright 2003 Intelligence Press Inc. Allrights reserved. The preceding news report may not be republishedor redistributed, in whole or in part, in any form, without priorwritten consent of Intelligence Press, Inc.