I will suggest and answer as to the "why" the US curve smooths out after 1960.

First, if you run a time series analysis on the variance(on the monthly data) you will see there is a periodicity in production of ~2.5 years. Even though they are much smaller in amplitude now, they persist to this day. What it looks like (and models out) is a damped sinusoid.

As the base magnitude of production increases (and "spare capacity" becomes challenged and decreases in both real and relative magnitude), the variance decreases.

Part of this change in amplitude may have been a policy change (and remember OPEC formed in 1960, though they didn't really flex their muscles until 1973). US imports grew significantly during the late 1940's and 1950's from 9% to over 20%, while oil exports fell essentially to "zero."

But in the period of the 1960's, oil imports also stabilized, as a percentage of the total, at around 20%. So, in a growing consumption modality, US production (or imports) could still act as the swing supplier though the degree of "spare capacity." was quickly being "consumed." And remember, the Texas Railroad Commission was also the 800 pound gorilla in the room with regard to US production.

A similar situation can be seen in the global data. If there is a great deal of "spare capacity" in oil production (and you have the storage capacity for what is produced), then you can see a great deal of variability both on the growth side and the downturn side of the production curves that are underpinned by economic conditions. This high degree of variability is seen in data all the way up to 1991.

But in the upturn in production that began in June 1993, the monthly variation became noticeably smaller (almost non-existent) as global oil production increased from about 60 million barrels per day in 1993 to about 67 million barrels per day just 5 years later in 1998. A short sputter between 1998 and 1999 and much monthly variation before resuming a runup into early 2001. Another short sputter into 2002 and then the last runup from 2002 into 2005 with somewhat larger monthly variation than seen in the late 90's run.

But since 1993, the variability (other than economic downturns) has become much smaller than in all previous data.

This diminishing variability looks like what happens when you have strong growth colliding with capacity issues and, in the case of the US in the 1960's, the coming peak. I suggest the same is likely to be true on the global scale, though we have many more "types of oil sources" than we had when the US peak occurred in 1970.

This is not to say we cannot have large spikes in production, but it seems less and less likely. Anyone know of any new Ghawar's planning to come on-line?

As an aside; in a class I teach, I start the class out on corn and wheat production with a progressive reveal of the curves and actually tell them what they are looking at. Then, I show this US oil production data and tell them "Here's 110 years worth of production data, and obviously we've gotten good at this. So, based on what we've done and been able to do so far, where do you think we will be in 40 years?" I also show them the growth variance curve that shows that while percentage growth is getting smaller, growth is still occurring. Then I give them the finer detail (monthly) from 1920 through 1970 so they can see the seasonal effect and the diminishing variability "for the past 50 years" prior to the 1970 peak.

Well, nearly all of the 40 students in the classroom just mentally extend the production out into the future and come up with an answer of "about 6" (representing the billions of barrels produced for a year, though they don't know that).

And then I put the graphs up with all the data since 1970 included and reveal what they are looking at. "Past performance does not guarantee future performance." Knowing what underpins the data, however, is useful.