Last month, the Canadian Association of Petroleum Producers organised a field trip for a group of economics professors to see a couple of the oil sands installations. I once wrote somewhere that the oil sands are the most important Canadian economic fact of our time, and I jumped at the offer to join the tour. I'm glad I did. There are many very good articles out there describing what is going on in northeast Alberta - much better than I could write - but reading those accounts is most emphatically not the same as seeing the installations in person.
Other people who have gone more often and seen more will have longer and better stories to tell. What follows are my impressions of what I saw and learned.
Here is the first picture I took - not by a professional photographer, nor with a high-quality camera, and from a moving bus - as we drove towards the Syncrude north mine operation:
Desolate, barren plains with towers emitting smoke in the far distance (it's actually almost all steam, but I didn't know that at the time). I felt like we were driving into Mordor.
The people who run oil sands operations take safety very, very seriously:
At times, this seemed a bit comical - I was once scolded for not holding onto the banister of the 5-step staircase leading into a building - but it's not hard to see why they would be so fanatical about it. Even if you set aside the usual concerns about workers' welfare, these installations operate 24/7: they really, really don't want to be obliged to shut things down in order to deal with an accident.
The Syncrude mining operation is astonishing in pretty much every dimension you can think of. The north mine - which was opened by Jean Chrétien in 1996 - is now a hole about 200-300m deep and 8 km across, and I if I recall correctly - and I may not - is due to remain active for another 10-15 years or so.
The extraction bit here is pretty simple: the diggers drop a couple of scoops onto the haulers, and the haulers trundle for the 15 minutes it takes to dump their loads at that crusher in the forefront. It's the scale that's incredible: each load is about 400 tonnes, which is in turn transformed into 200 bbls of synthetic crude. And the heavy haulers - the word 'trucks' seems inadequate - are the size of a house.
The next step is to separate the bitumen from the sand (the stuff that isn't bitumen is the tailings). The upgrading process itself consists of breaking the bitumen - which consists of molecules of long carbon chains, and with not much hydrogen - into the shorter hydrocarbon chains that make up synthetic crude. (You pick up a lot of this talk around Fort McMurray.) This goes on in one of the many, many buildings that look like this:
The elapsed time from shovel through the upgrading process and to the pipeline is about four hours.
One of the things that happens during the upgrading process is that sulphur is extracted. There's no way to get the sulphur to market (there is no railroad, and trucking is too expensive), so Syncrude has accumulated enormous stockpiles of the stuff; here's a picture of a portion of one of the dozen or so piles that I saw:
Apparently the stockpile is so large that if they ever did figure out how to get that sulphur to market, it would squash the world price like a bug.
Syncrude would be very disappointed with me if I didn't point out that when when an ore body is exhausted, they don't simply walk away from a gigantic hole in the ground and move on to dig another. The next stop on our tour was this reclamation project at the former East Mine:
That little piece of Mordor was in fact a former mine that is in the process of being reclaimed. This portion is further along in the process:
This is where we talk about the famous tailings ponds. The tailings - the oil sands with the bitumen removed - are roughly the consistency of yogurt. The basic strategy of dealing with them is to let it settle so that the sand sinks to the bottom and the water rises to the top. When the sand has settled out and the water is removed, you get that picture of Mordor I posted above.
Oh yes, while we were there, explosions were going on all the time. After the disaster of all those ducks trapped in the tailings ponds at the Aurora mine, Syncrude set up a system of radars (yes, really) to detect any waterfowl that might be in the area and cannons to scare them away.
The next step involves taking the layer of dirt that was dug away to expose the ore (they call it the 'overburden') and using it to do some serious landscaping to recreate the sort of terrain that had been in place before the digging began - muskeg and all.
Our Syncrude guide even made us go out and walk around in the newly-reconstructed muskeg, and it was just about as much fun as you'd imagine a stroll in the northern woods in 30+C weather would be. Okay, my grumpiness was partly my own fault: I had forgotten to wear a long-sleeved shirt and they made me wear a windbreaker to conform with their safety regulations. On the upside, I did get a picture of this bison, part of a longer-term project to re-introduce them to the area:
One the important things to understand about the Syncrude operation is that time there is measured in decades: two or three decades of extraction, another couple of decades of reclamation, and then another couple of decades to make sure that the ecology of the reclaimed land is sustainable.
That was the first day's outing. The next day was at the ConocoPhillips-Total Surmont in situ site. Surmont 1, to be precise: construction of the Surmont 2 expansion project was in progress a couple of kilometers off in the distance. The ore bed there is too deep to get at using mining techniques, so they use steam-assisted gravity drainage (SAGD). This involves drilling pipeline pairs that enter the deposit horizontally. The top pipe blasts steam into the deposit, making it sufficiently liquid to flow by pipeline, and the bottom pipe draws the steam and bitumen mixture away. This is what a typical well pair looks like above ground:
There are a score or so of these at each 'well pad' and Surmont 1 has 18 well pads. That makes for a lot of pipes heading back to the main installation:
Steam will get the bitumen to the main installation, but it's not going to get it to world markets. The next step is to take out the water (it's recycled) and to dilute the bitumen (which can't flow; it's basically solid tar at room temperature) with a solvent. The resulting mixture is called dilbit - diluted bitumen - which can be shipped south by pipeline. The solvent is recovered during the refining process at the other end, is sent back and recycled.
And that was my tour. What did I learn, beyond a bit of conversational chemical engineering? One of my first reactions was to wonder why the production of synthetic crude and dilbit is classified by the NAICS as 'oil and gas extraction' while oil refineries are part of the manufacturing sector. Maybe much of our much-discussed "crisis in manufacturing" would go away if upgraders were classified in the same way refineries are.
The question I kept asking revolved around this graph from last year:
If we took multifactor productivity seriously as a measure for technical progress, we'd (wrongly) conclude that firms in the oil and gas sector were using technology three times more advanced in the early 1960s than they're using today. So my question was how technical progress in the oil and gas sector should be characterised, because clearly the 'producing more with fewer inputs' MFP measure was the wrong way of going about it. These were the answers I got:
- Technical progress has expanded the number of projects that are economically viable. For example, before the development of in situ technologies, ore deposits that were too deep to get at using mining techniques would have been left alone. The counterfactual of no technical progress would be a decline in output.
- Much/most of the investment is front-loaded: oil sands projects require years of construction before production starts. According to this list, current oil sands production capacity is 2.25m bbl/day, and another 788,000 bbl/day - some 35% of existing capacity - is under construction. Since current investment expenditures are applied to Statistics Canada's perpetual inventory model, the projects under construction show up as a current increase in the capital stock with no corresponding current increase in production. This would drive estimates for MFP down, ceteris paribus.
- Environmental regulations are more strict. This is the only point I hadn't thought of before. The idea here is that expenditures that reduce environmental degradation but do nothing to increase oil production will show up as a reduction in measured productivity. I don't know how important a role this has in explaining the oil and gas producutivity puzzle, but it's not something to dismiss out of hand.
And that's my report. My advice for anyone who receives an invitation to tour the oil sands is to take it.