Whenever I see a plot of the logistic curve, I always like to see the expression with values for the coefficients next to it. I believe that this consists of an overall scaling parameter, an exponential factor, and a time shift on the exponential for the simplest version.

In this case, it's a random sum of truncated logistic curves that can produce any kind of curves (even multimodal). Therefore, it's not a parametric model with a particular set of parameters.

That's an artifact produced by your model assumptions. The lag between discovery and production means that the production curve for 2000-2010 was 99.99% determined by discoveries before 2000. Thus, changing discoveries after 2000 should have no discernible effect on a peak of 2006. Eyeballing the two curves, the differences only appear on the backslope of the curve, a few decades hence.

Your assumption is not correct SilentE. At the moment the average life after discovery when an oil field comes on-stream is approximately 4-6 years. Some oil fields come on-stream 2 years after they have been discovered!

I assumed a uniform distribution between 6 and 25 years to model the delay between time of discovery and time of production start. The economic context has obviously a huge impact on this time lag but it's a challenge to derive a statistical relationship oil price, available capital and decision to invest in new production.

I didn't know that it could be as fast as two years! I assume it must depends on the size of the field and its location (offshore fields must take longer). Do you have any data on that?

Average field distribution from 100 projects announced to be taken on-stream between the period 2005-2010 (taken from my own report, launch date now set at the end of October due to media-related reasons).


This analysis showed that:

• 23% of the projects were EOR projects
• 12% of the projects were discovered before 1980
• 7% of the projects were discovered between 1980 and 1989
• 29% of the projects were discovered between 1990 and 1999
• 18% of the projects were discovered between 2000 and 2004
• 11% of the projects were unconventional projects (tar sands and orinoco heavy oil)


Some Specific oil projects discovery - on stream time lag

Golfinho

Country: Brazil
Type: offshore
Discovery date: 2003
Reserves: unknown
Production starting date: 1st half 2006
Amount of production: 100.000 b/d
Operator: Petrobras

Constitution

Country: Gulf of Mexico
Type: Offshore
Discovery date: 2002
Reserves: unknown
Production starting date: 2006
Amount of production: 70.000 b/d
Operator: Kerr-Mcgee

Muniteer-Exeter

Country: Australia
Type: offshore
Discovery date: 2003
Reserves: unknown
Production starting date: 2005, full production expected in 2006
Amount of production: 100.000 b/d
Operator: Santos

Luda

Country: China
Type: offshore
Discovery date: 2000-2002
Reserves: unknown
Production starting date: 2005
Amount of production: 40.000 b/d
Operator: ConocoPhilips

Tui

Country: Maui
Type: offshore
Discovery date: 2002-2004
Reserves: 20-30 million barrels
Production starting date: 2006
Amount of production: 30.000 b/d
Operator: Taranaki


As an average distribution we see that newly discovered fields are at the moment coming on-stream on average 4-6 years after they have been discovered. Sometimes this is 3 years and in very special cases 2 years.

The fields coming from the "existing reserve base" or in other words, discoveries made before the year 2000 are rapidly being brought into production due to economic factors (high price of oil) and technological possibilities. Or in other words we are eating the "lower fruit"

Almost all projects coming on-stream are offshore oil projects (except for EOR and Unconventional). This is a serious issue since offshore oil fields rapidly decline on average after 10-15 years of production.

Do you mean that only oil fields that entered production in 2005 were on average discovered 4-6 years ago? Or that all oil fields discovered in 2000 will be placed into production within (average) 4-6 years? That is, does that 4-6 year period only apply to the subset of oil fields that are actually put into rapid production, while other fields are deliberately held in reserve for far longer, and thus because they have yet to be produced they are not included in the average?

Any insight into what accounts for the wide range of differencs - is it national, regional, or type (offshore)?

No



Yes



No, because of economical and technological differences and incentinves when compared to the past any oil field discovered at the moment is rapidly being brought into production. Whether it is an old oil field (discovered 30 years ago) or a new oil fields (discovered 5 years ago)



most oil comin on-stream is offshore nowadays, allmost every on-shore field is already in production to my knowledge (except for a large amount of barren Russian fields, they are also coming into production). This is a worldwide trend (except maybe for the "missing" 130 billion barrels not in production in Saudi Arabia).

Thanks Taskforce for these details!

It gives the following pdf for the lags (excluding unconventional projects):

This is very different from a uniform density, it looks more like a Gamma law.

question:
1- do you know the percentage of offshore projects?

Which is what I have been saying with my model. I assumed a convolution of three exponential phases, Fallow, Development, and Maturation. This just happens to give a gamma distribution of order 3. Which gives ~ x-cubed * exp -x/3. Drop one of the phases, and it is still a gamma but of order 2.

Ill look it up later today or tomorrow.

Besides that you should be aware that 90% of all oil fields every discovered were in production in 2003 according to IHS. That means that approximately around 2010 no oil fields will come on-stream anymore from before 2000 (or very little). Given the current rate in which oil-fields are coming on-stream.

It's not an exact figure but i guesstimate 60% from going through 40 projects and also including unconventional oil and reworking old fields.

WebHubbleTelescope I am looking through your shock oil model. Looks really interesting and it is even so simple that i can understand parts of it .

Anyway might be interesting to insert some new variables like reserve increases over time due to technology. Or faster pumping up of oil due to technology whereafter the decline curve is far steeper. Or overproduction...

I assume you are talking about the legions of analysts working for the oil industry at a salary, sequestered in their office buildings, unable to say a peep about the reality of the situation? And then the paid consultants, told to muddle the arguments when someone comes close to the truth?