Sunday, March 2, 2008
Natural Gas Analysis
First of all, one rarely finds this kind of data for crude oil, where each year gets tallied separately. I figure the reason it doesn't show up very often arises because the maturation of oil wells do not correspond to a given year very concisely. A variable maturation time essentially pushes the actual production to the next few years so that the chart wouldn't have the same columnar contrast. On the other hand, the draw of a natural gas reservoir occurs immediately and so the yearly data shows up very strikingly.
The Oil Shock Model handles the analysis fairly well. I basically set the maturation level to zero years and tried to emulate the chart's look.
Overlay below
Not having the actual discovery data available, I used empirical fits to generate the individual curves. The base curve for the years prior to 1980 results in a curve that follows the relationship:
Base = K * (0.55 exp(-T/2) + 0.45 exp (-0.15*T/2))
This essentially gives a fast slope and a slower slope which approximates a reserve growth component that I have reported on before. The individual yearly production curves also show a similar behavior (here small t corresponds from the date of the :
Base = Gain * (0.9 exp(-t/2) + 0.1 exp (-0.1*t/2))
The slow portion contributes only 10% of the bulk of the growth, so the reserve growth doesn't amount for much.
The other piece of the fit involved the contribution of the Gain which basically generates the envolpe of the curve, starting from the initial point in the data collection at 1980.
Gain = (1 - 0.7 exp(-0.02*T))
This gain function basically demonstrates that continually greater amounts of Natural Gas get extracted per year but the trend does not show that a peak will arrive any time soon. It instead suggests that new wells get constructed to meet the demand for Texas. The yearly time constant for each year's output remains a pretty quick 2 years, so that when a drop-off in production occurs, it will happen fairly quickly.