Gail - I'm having trouble understanding the graphs.

What is the "baseline"?

Does the graph presume an amount of production over baseline at year 1?

Help!

Imaging you have graphs for a number of countries, all expressed in barrels per day. For each of the graphs, chop off the bottom piece of the graph, that shows no variability. For example,

In other words, subtract an amount equal to the lowest production during the period shown on the graph for each country. In the graph above, it occurs when hurricanes interrupt production. That is the baseline production.

Then produce a graph by adding up the tops of the various graphs. The band width shows how a particular country's production has increased or decreased. So what a person is looking for is changes in band widths--narrower widths mean less production for that particular country.

Matt has some groupings of countries he uses. All countries that he think peaked in 2005, for example. The bigger countries are shown separately, though.

Perfect.

Now that you've pointed it out, it's crystal clear!

Thanks.

Since it is an incremental view, the base production does not really matter and is somewhat arbitrary. I think he set it up so that it sums to match the current total.

It would be really slick to be able to adjust the baseline with a Flash interface or the like; this is one of those fancy web presentations I can easily visualize but have no idea how to implement.

The base production is not arbitrary. It is the sum of all minimum crude production by country, starting with January 2001. Of course base plus incremental production must add up to the total as shown in Rembrandt's curves. Iraq shows up with a large amount of incremental production because of the production drop during the Iraq war. The same is true for Venezuela which had a strike causing a big decrease in production.

The graphs are an update to:

Bumpy Crude Oil Plateau in the Rear View Mirror
http://www.theoildrum.com/node/3793
or
http://sydneypeakoil.com/matt/Crude_Oil_In_Rear_Mirror.pdf

It could be arbitrary. I did say in this case that it added up to current full production. But it could be arbitrary, and that knowledge separates people that understand calculus from those that don't. :)

Maybe I was too flippant. The base production is an invariant and therefore it can be removed. The calculus comment means that you can also called it a differential production curve.
P(time) = P0 + dP(time)

in which case you can see that P0 can be removed with no loss of generality in the time dependence.

...the base production is invariant

Wrong. If for example in the next month UK or Norway drop to a new minimum the base production will change.

It is being re-calculated every month.

It is invariant over the span of the graph.

What you are describing is something he has not plotted.

I don't know why we are arguing over this. Let me just state the algorithm:

1. Scan through production levels for each country over a time range
2. For each country i, store the minimum production level Po(i)
2. Sum Po(i) over all countries i, call this total the invariant Po
3. At each time and country, store the differential of dp(i)=(P(i)-Po(i))
4. For each time value, create a stacked bar graph of the sum of dp(i), using Po as the invariant baseline

Let us argue over the algorithm so we don't get lost in the ambiguousness of the english language.