You're asking inconvenient questions. See my Socratic Economics V: Supply and Demand

In that little diagram, the "equilibrium" or crossing point is observed, namely a (price, volume) point. The rest of the diagram is counterfactual. I suspect people are using consumption as a proxy for demand and production as a proxy for supply. The difference between consumption and production is the change in inventory. But ignoring that difference for a minute, consumption is just the demand at the "equilibrium" point. Nobody is counting all the times someone goes and says "gee, I want to buy some more of this but it's too expensive" or "gee, I wanted to buy some more of this but the shelves were empty" and so the demand curve is unobserved (unobservable?). Similarly with the supply curve.

Does one have to look at the order book at a commodities exchange (or the open interests in futures) to get supply and demand curves for oil?

And, of course, if the supply and demand curves are counterfactual it is going to be rather hard to measure their slopes (the "price elasticity" of supply and demand).

Not that I claim that diary or the comments has an answer - all I'm suggesting is that the concepts are nonsense. the IEA is representing production, not a supply curve as a function of time. And they're, of course, assuming a particular level of satisfied demand.

Of all the ways of organizing banking, the worst is the one we have today — Mervyn King, 25 October 2010