Where did that oil spill go? We've got millions of barrels of oil sloshing around off the most sensitive coastline on the continent, and for lack of oiled birds, the Deepwater Horizon Spill disappeared from the front pages today. The iconography of ecodisasters (oiled birds) has a match in the narrative motifs of technological failure -- normally we find a "bad part" (as with the faulty O-rings on the Challenger disaster),  or a villain (as in the supposedly drunk Captain Hazelwood of the Exxon Valdez Spill, who took it on the nose for a series of other unfortunate decisions made elsewhere). These familiar motifs reassure us moderns that there are practical solutions to technological disasters -- what's wrong is the part (or the person) and not the whole damn undertaking.

Such a reassuring motif is failing to materialize in the Deepwater Horizon disaster. In a rash of fairly detailed "first hand" stories this weekend, the problem appears not to be exclusively the Blow Out Protectors (the star suspicious part of the first week of the disaster), but harder-to-understand bit players like the foaming nitrogen-enriched "mousse" of cement, and a "packoff outside the casing." (The most detailed account is this AP story, which bears a strong resemblance to this anonymous narrative and analysis that's been circulating through the oil industry -- scroll down the page to May 6. Go ahead, be skeptical. I am.)  There are early signs that the problem might be the whole thing: stories came out this weekend about the failed regulator MMS, and the troubled safety record of Transocean, the rig operator -- but still no smoking gun, or smoking guy.

In fact, doggedly searching for the "one thing" that could have prevented the spill may lead us in entirely the wrong direction. Over the past week I've been lurking in the forums of gcaptain.com, a virtual hangout for offshore drilling boat types. Gcaptain may have had the first news of the Deepwater Horizon fire when a nearby boat captain reported it ten minutes after it happened, kicking off a long thread of discussion about possible causes and meanings.
The conclusion I've drawn from my eavesdropping, and interviews with some posters on the boards is this: IT ALL FAILED. Deepwater Dynamic Positioning rigs have three lines of defense against blowouts: first is control of the pressure in the hole through "mud" and cementing which keep gas and oil from coming out of the well; the second is the Blow Out Protectors, which pinch off the top of the pipe coming up from the well, thus keeping the oil and gas inside; and the third and most dire is the Planned Drive Off, in which the captain fires up the engines of the rig (which is really a ship) and drives off very fast, in the hopes of smashing the riser, depriving the fire of gas, and getting everyone safely out of there. But in the case of the Deepwater Horizon, the mud and cement failed to control the well; the BOP's failed to work (possibly because they were jammed by well casing or disabled by sand from the big "kick" of methane); and then the rig engines caught on fire which is NOT SUPPOSED TO HAPPEN -- preventing the possibility of a planned drive off. 

So how does everything fail, when much of the time most things more or less work? One interesting poster in the gcaptain forums is Bob Couttie, a consultant on ship safety who keeps a website called maritimeaccident.org where he creates podcasts of accident case studies that mariners listen to in their free time. Bob pointed to two aspects of the disaster that have troubled people who work on offshore rigs. First, the accident occurred on the last night of the crew's two-week "tour." Secondly, there was a party on board the rig to celebrate 7 years of safe operation -- which included executives from BP (as many as 7, according to published accounts) and probably Transocean executives as well.

Speaking by text message from the Philippines, Bob said:

When a kick starts the difference in good or bad outcomes can be a fraction of a second. The crew needed to be fully focused on what they were doing. Even a second's distraction, in these circumstances, would be enough. I'd say we have to look at 'human elements' and whether procedures were followed and whether the procedures were adequate -- not just those on the rig but the decisions made onshore. I'd be looking at the effect of what I gather was a 'head office' visit at a time of intense activity. 

People on similar rigs told me that head office visits include days of extra work leaving everyone tired and stressed out. Bob continued by email:

Another thing I'd look at is the 'end of job effect'. It's interesting how many seafarers die doing something wrong the day before they leave the ship. I think what happens is that towards the end of the job time is short, the work is intense, and people become more and more focused on what they are doing rather than what's going on around them -- they want to get their part of the job done. So they loose situation awareness -- almost a sixth sense of what's going on around them.

There is a vast literature on accidents and causes and concepts. The classic is James Reason's "Swiss Cheese Model." My own working concept is the "tower block:" you're probably familiar with the tower game which consists of layers of wooden blocks at right angles to each other and each player pulls out one block, adjusting other blocks if necessary. The loser is the one who tries to pull out a block and the tower falls over.  What happens is you start with a stable structure and as each block is removed the structure becomes more and more unstable until it becomes so unstable that it fall over. It's a good way of looking at accidents. Each block is something that contributes to safety, might be procedures, situational awareness, rules, technical safety equipment. As each one of these blocks is removed the structure -- "safety" -- becomes more and more unstable until it doesn't actually matter which "block" goes net, the structure will fall. An example is the Bourbon Dolphin.

If you follow the tower block analogy, it's clear that while we may want a simple explanation for the Deepwater Horizon accident that allows us to go back to business as usual with a few modifications, what we're going to get is a long, detailed, thoroughly modern flow chart about the limits of technology, humans, geology, and regulation. As people who like simple narratives, the public and policy makers will be tempted to try to find one locus for blame -- whether it's BP or BOP's (blow out protectors) -- but that may prevent us from figuring out the deeper system of problems that lead to this accident. And we may determine that business as usual doesn't work for offshore drilling -- which leaves us unable to count on the 40 percent of domestic oil production we were expecting to get from the offshore industry in the next ten years. Rereading what's been written about offshore oil drilling over the last few years, it's obvious it was thought to be the methadone for our overseas oil addiction. Now what?  

(An aside: given all the help they're getting from the Coast Guard, the military, and state, federal, and local agencies, it's a bummer that Transocean, which was a Houston-located company until the brass moved to Switzerland in 2008, has been so acutely focused on NOT paying U.S. taxes, according to this report from Portfolio.)

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.