Great work Wes. For me log normal plots tell the truth as best can be learned from the limitations. I know you understand the significance of a straight l/n plot but many don't understand the physical natural laws driving the reservoir dynamics. I'll bet more than a few are suspicious of such a straight and predictable projection. Once, on another thread, I tried to explain the accuracy of production forecasting when the decline curve reaches this point but I doubt I did much good.

westexas, what was the historical decline rate of Texas and the Lower 48 & how closely do you think Saudi Arabia may follow this? Sorry if this has already been mentioned on prevuous threads.

It is self evident that the logistic method for swing producers has problems, but does that make it useless?

Depends on how big the problems are. You and Laherrère are clearly pretty far apart here.

The 2006, 2007 and 2008 Saudi data are going to show a logistic plot that is consistent with an URR of less than 200 Gb, (C+C).

But, the URR is going to show that it is creeping higher as you add those numbers in. You will also find % of Qt moving backward, not forward as you would expect if the HL was correct. That's what I discovered both when I backcast Texas from about 1960 forward, and when I did the HL for Saudi and added in the years following 2005. In other words, incorporation of new data indicates that the URR is not static per the HL.

It might be useful to update the paper, incorporate the data through 2007, (you could even estimate 2008 at this point) and indicate what is changing - and by how much. But what, I mean URR, % of Qt, etc.

Jeff,
You should use this result instead:

It leads to the following forecast which has a wide uncertainty interval in terms of flow rates but the upper limit is close to what Jean is saying:

The HSM is also pointing to a constant production scenario until 2020:

I tend to concur with your view as the graphics below. In the first graphic, the data is annual data AND for comparison I show where the linear flow would have to be to achieve a URR of 260 GB. The last "tail" data is what would be projected if the total year average for the KSA was 9.3 MMBPD, a number that has not yet been achieved in 2008 as an average for the YTD. As Robert points out elsewhere, it has the effect of shifting the curve to the right but only slightly. Flat production eventually shows up as a declining curve as the P/Q ratio gets smaller with each additional data point. And to continue to shift the curve "to the right" one has to have continual growth in production for quite sometime. Mathematically, what compounds the issue is that larger the production the faster the increase in sum value (Q) and the more difficult it becomes to sustain or increase the P/Q ratio.

Notice that this graphic projects to a value of ~213 GB. the arrow to the 2005 peak also moved in the translation from Excel to a JPEG. Thats the point just above the arrowhead.

The second graphic is the same data except using monthly as opposed to annual data. The curve has much more noise to it, but the linearization covers the same time frame (January 1990 to present). Note that the monthly data suggests a somewhat lower URR (~205) AND because of the larger number of datapoints is less influenced by a single or a small collection of production increases or decreases on a monthly basis. To shift this curve substantially, as noted above, would require production increases that are not characterisitic of any production curve that we have seen.

For fun, I increased the KSA output by an average of 100,000 BPD/month from now until the end of 2009 (so that production is over 11 million barrels per day and beating their November 1980 output record. It does shift the curve to a URR of ~240GB or roughly the amount used globally in a little more than one year. But as soon as this increase "stops" the curve begins it's decline once again. So, it is reasonable to ask, in the face of what we do and do not know, does any of this increase seem reasonable or even plausible?

And for real fun if you increase at a rate of 200,000 BPD/month so that production at the end of 2009, production is more than 12.5 million BPD, it shifts the curve a little further to the right to the value of 260 GB URR, but only 20 GB greater than the hypothetical immediately above even though the rate of increase is twice as high over the same period. And this production curve does not look like anything we've seen recently from the KSA.

"Ultimately," the big numbers get you!

I not know much about this subject but I like to share some simple thoughts, please forgive me for my ignorance:

First, a swing producer exist when there is an abundance (supply > demand at a reasonable price) so in order to stop price from falling the swing producer cut production, vice versa for the opposite. As soon as abundance is gone and tightness comes, when demand is greater than price there is no longer a swing producer because all the producers are producing at their maximum level to take advantage of the price.

Since the hubbert's linearization is I think based on the assumption that a producer will produce at its maximum level geologically possible therefore it not work for swing producers. There is no use of linearizing a swing producer. In order to get some useful information about flow rates and reserves must only linearize non-swing producers only. The swing producers must only linearize when they are no longer swing producers.

Second, all oil exporters which also have vast natural gas reserves (I think they all have that) would ultimately liquify their natural gas to be sold as liquids. So, in order to get useful information about prices, flow rates etc we must take in account natural gas reserves of an oil exporter as well as its oil reserves. Luckily there is no opec like thing for natural gas liquids so we can assume that govt declared reserves are correct. This thing help reducing chaos in our data and predictions that come due to uncorrect reporting of oil reserves by some govts.