15 October 2010

Is Your "b exponent" Hyperbolic or Exponential?

In the FT yesterday, John Dizard explains an important debate over the future of natural gas, with far reaching implications for investors, governments and you and me.  Like many debates of relevance to policy, the debate over production decline curves in natural gas wells from shale gas formations hinges on expectations of the future that can only partially be addressed via scientific predictions.  Uncertainties and ignorance are fundamental.  The market has been rife with optimism that there is plentiful, recoverable natural gas available via new technologies for drilling.  But not everyone agrees.  Dizard explains the debate:
[R]right now in the US and European energy world, there is one number that does tell you the story. It’s just not a number most investors, and the policy tribes, ever heard of, let alone used as a guide to action.

I’m speaking of the “b exponent”. You are much safer questioning the occupation of a gas promoter’s mother than doubting the “b exponent” of his wells. After all, the former is only about the past; the latter tells his future. And your future, since you’re paying for his future.

The b exponent is a term in the equations that define curves on a chart that describe the rate at which the production of gas or oil wells declines over time.

These equations have been tweaked and elaborated over the past 60 years, incorporating the slow increase in geological science, along with the cyclical requirements of gas and oil promoters.

The decline curve equations all incorporate terms for initial production rate, the initial rate of production decline, and the degree to which that initial decline rate flattens out over time.

The b exponent is a way of getting the curve generated by the equation to fit that rate of flattening.
He explains the debate over the "b exponent" in terms of two views on how it will play out in the context of shale gas:
The debate, which is turning into an argument, over the potential for shale gas development turns on that rate of flattening. The two schools of thought are the “hyperbolic decline” people and the “exponential decline” people, referring to two curve shapes.

The hyperbolics believe that decline curves for shale gas flatten out over time, much like conventional, vertical wells drilled into sandstone or carbonate formations. For them, the b exponents of their wells would be 1 or greater. They can take the high initial production rate of shale wells, or “IP”, and show financiers how this translates into high recoverable reserves or “EUR”.

The exponentials, a so-far smaller group, say that shale wells decline quickly after their initial high production, so the b exponents would be, say, 0.5. On hearing this, the hyperbolics will break a long neck beer bottle on the bar, and a fight will start.
Why does this debate matter?
If the hyperbolics are right, then shale gas wells in Louisiana or, prospectively, Poland will produce gas at a reasonably high rate and low cost over a long period of time.

That is what has been sold to Wall Street and is being used as a negotiating position with Russian gas officials. Up to now, the hyperbolic argument has worked with Wall Street. Unfortunately for Europe, the Russians think the production and unit cost numbers are an empty bluff.

I have seen very smart exploration and production people using the best available science and data make large losing bets on the size and location of hydrocarbon deposits. Unlike Wall Street or political people, they eventually face the facts, admit when they’re wrong, and use the information from the failures to do better next time.

If this were only an insider industry debate, then the rest of us could ignore it, or buy some popcorn and watch the show.

But it isn’t. The US, Europe, and now China are making huge investments in switching from coal-fired power to gas-fired power, and if there isn’t enough gas at a low enough price, they have a problem.

For example, if the pessimists/exponentials are right, then the ultimately recovered gas reserves from, say, the Haynesville deposits in Louisiana and Texas could be closer to 2bn cu ft for the average well, rather than the 6 bcf some operators project. If so, then the wells on most of their land would need a gas price that is at least twice, and perhaps three times, what is on offer in the spot and futures markets.
Keep your eye on the evolution of knowledge of the "b exponent" -- this is not a debate that will be settled via theoretical arguments, but through actual experience in natural gas recovery.  Stay tuned.