At rush hour on August 1, 2007, the 1,907-foot-long I-35W Bridge near downtown Minneapolis fell into the Mississippi River, killing 13 people, injuring 145 more, severing a key link in the Interstate system, and costing over $300 million in damages and for construction of a new bridge. 1 After more than a year of investigation, the National Transportation Safety Board concluded that the engineers who designed the bridge in the early 1960s had undersized the gusset plates that connected its steel segments. That error, compounded by the weight of extra lanes added over time and repaving equipment and materials on the fatal day, caused 456 feet of roadway, and 111 vehicles, to collapse in seconds and drop 108 feet into the river. 2
Experts estimate that some 465 U.S. bridges are similar in design to the I-35W span. Inspection and reinforcement of these structures are vitally important as the nation prepares to upgrade its infrastructure. 3 Just as crucial, we need to see the I-35W and similar spans not as isolated cases but instead as harbingers of a problem that plagues much infrastructure and development of the last 60 years. The problem is what engineers call “fracture-critical” design.
A fracture-critical design has four key characteristics. The first is lack of redundancy, which makes a structure susceptible to collapse should any individual component fail. The I-35W’s undersized gusset plates might not have brought down the span if it had had additional members to carry the structural load. At the time of the bridge’s design, such redundancy no doubt seemed expensive and wasteful. But given the extraordinary costs — financial and human — of collapse, the incremental expense of redundancy would have been cost-effective and wise. Engineers understand this and in recent years have increased the redundancy of bridge designs — but the pressure to reduce initial costs continues to threaten the durability of our infrastructure.
Other weaknesses of fracture-critical design are interconnectedness and efficiency. The I-35W had both. When the gusset plates cracked near the bridge’s southern end, this overstressed other structural members — all interconnected so efficiently that nothing could interrupt serial collapse. The 10th Avenue Bridge adjacent to the I-35W shows the advantage of less interconnection, less efficiency. Completed in 1929, that bridge consists of independent concrete arches separated by concrete pylons that divide the structure into discrete parts. The concrete columns supporting the road deck seem oversized, making the entire ensemble less than efficient, but more than sufficient to compensate for the failure of any one element. 4 Even if several columns or one of the arches failed, the bridge wouldn’t collapse.
The final characteristic of fracture-critical systems is sensitivity to stress. Had inspectors attached strain gauges to the I-35W gusset plates, they would have detected a gradual increase in stress, with a rapid rise in strain, just before the plates fractured and the bridge fell. Sudden, exponential increase in strain prior to failure is a well-known phenomenon, and a fracture-critical design magnifies its effect. What seems a localized, controllable problem can quickly become catastrophic, because of the nature of exponential growth, doubling with each increment of time.
To understand how lack of redundancy, connectedness, efficiency, and exponential stress relate to each other, consider the concept of “panarchy,” explored by ecologists Lance Gunderson and C.S. Holling. 5 Panarchy explains that human and natural systems move in continuous adaptive cycles, and that exponential growth in connectedness and efficiency actually makes systems less and less resilient, inevitably leading to collapse and then return to a state of greater resilience, with fewer connections and less efficiency. The collapse of fracture-critical designs like the I-35W — which we would be wise to see as part of an adaptive cycle — warns us that we need to replace such structures with designs that are less connected, less efficient, more resilient.
But we’ll need to change more than the design of our bridges. It’s clear in retrospect that the fracture-critical structures of the 1950s and ’60s reflected the larger culture — this was when John Kenneth Galbraith famously critiqued the United States as a nation of private affluence and public squalor. In an era when America could have afforded the best infrastructure in the world, we began instead to channel wealth into private hands and to impoverish the public realm. 6 This was also when a deeper, though subtler shift began to be felt in American culture. The United States had emerged from World War II as the dominant global power and, as many commentators have noted, dominance easily led to hubris, to the pride of pax americana in the ’50s and more recently to theories of American exceptionalism. 7 We now know that our wartime enemies as well as allies have proven to be formidable competitors, and that we can no longer take dominance for granted. In this sense the I-35W stood — and fell — not just as a physical bridge across the Mississippi but also as a symbol of postwar overconfidence. Fracture-critical design epitomizes all the postwar systems vulnerable to sudden failure. The bridge’s collapse warns us that future catastrophic events will surely occur. The I-35W Bridge is both metaphor and omen.
For much of the past year we’ve witnessed the collapse of what turned out to be a fracture-critical global financial system. It’s not immediately apparent that global finance is a designed system. Yet just as the failure of one set of structural components caused the I-35W to collapse, so too the failure of key investment banks — Bear Stearns in March 2008, Lehman Brothers a few months later, in September — tripped a chain-reaction collapse of other banks and their insurers, and then of credit and stock markets around the world. Just as highway repair crews had piled on extra weight while resurfacing the I-35W roadway before it failed, so too did the markets pile huge amounts of debt onto the financial system, overloading banks to the point of collapse. And just as government inspectors and engineering consultants failed to detect the bridge’s weakening plates or to understand the risk of inaction, so too government regulators and independent auditors failed to provide adequate oversight or public explanation about how mortgage-backed securities might endanger the global financial system.
Once we see the collapse of our fracture-critical financial system as an adaptive cycle, we can predict what will follow and how to prevent future catastrophes. Our global banking system will likely emerge — or should emerge — from the current crisis less connected, less efficient, and thus more resilient. As in a resilient bridge, a transformed financial system will have more discrete, disconnected parts, with strong internal divisions so that even if one part fails, others will be insulated. It will have more redundant parts, with checks and balances to ensure that inspectors and auditors catch calculation errors or outright fraud before they do systemic damage. And it will have — or should have — less speed and efficiency; transactions might have built-in delays, allowing for extra time and added review. Indeed, the very idea of a globally integrated financial system might disappear, as nations hurt by the current collapse (over which they had little control) set up review procedures and regulatory policies to prevent worldwide meltdowns so adversely affecting them again.
We recognize fracture-critical designs after their failures, whether caused by inadequate steel or sub-prime mortgages. What about systems yet to fail? What are they and how can we prevent or at least mitigate their collapse? Most fracture-critical systems send warning signs before they fail. (The I-35W gusset plates bent long before they broke; prominent investors warned about the dangers of credit default swaps.) The challenge now is to see the signs.
The U.S. electrical grid has already sent an unmistakable signal. On August 15, 2003, an outage near Cleveland cascaded into the largest power failure in North American history, leaving 50 million people across the U.S. and Canada without power for days and causing an estimated $10 billion in damages. 8 Years after the blackout, industry experts are worried that the situation has worsened, with excess capacity declining and demand for electricity by 2030 expected to increase 29 percent from 2006 levels. This is a fracture-critical system needing immediate attention. A singe failure can cause damages far costlier than the expense of adding capacity and building in firewalls. (A fracture-critical electrical grid is especially vulnerable to sabotage, making the cost of added resiliency still more crucial.)
Still, the electrical grid is comparatively easy to comprehend and plan for because it constitutes a nation-wide system. Harder to spot are failing systems that don’t seem connected or even related: for instance, the tens of thousands of U.S. suburban housing developments constructed in the postwar era. For most of American history, we built communities over time, deploying diverse building types and accommodating different kinds of households. Such communities are resilient precisely because they are socially and economically diverse. Yet lately we’ve constructed a different America: subdivisions consisting entirely of single-family houses alike in design, size, price, et al., usually built all at once by a single contractor. Developers like uniformity because it’s easier to finance, build, market and sell, with the promise to prospective homeowners that the neighbors will be much like them and they needn’t worry about the frictions that crop up in more compact, mixed-use, mixed-income communities. 9 But postwar suburbia is proving fracture-critical. When homeowners default on mortgages and get foreclosed upon, banks typically lower the price of the house to sell quickly and recoup losses. But if enough foreclosures happen in a development where the houses are more or less interchangeable, then the value of all the properties drops, often to the point where many owe more on their mortgages than the houses are worth. 10 This in turn pushes more homeowners to walk away and more banks to foreclose, intensifying a bleak spiral that can destroy the value and morale of the neighborhood.
Just as fracture-critical as suburbia is the oil-dependent transportation system that’s made suburbia possible, and whose failure would devastate our economy. We know we are vulnerable to decreases or cut-offs of the U.S. petroleum supply, more than half of which comes from foreign sources. Yet we continue to rely on oil as the major energy source for transportation, panicking when prices rise and relaxing when they fall back to “normal.” 11 Developing alternative fuels is necessary — yet the more fundamental problem is rooted in our dependence on any single source. (We’ve already seen how overreliance on ethanol can unexpectedly pollute water supplies and raise food prices.) True energy resilience means a range of sources — not just oil and biofuel, but electricity and hydrogen as well as solar, wind, and even human pedal or pedestrian power. Visitors to India are often amazed by the transit diversity of the streets, from cars and trucks to mopeds and rickshaws to bikes and cows. Once (in our hubristic postwar decades), we might have dismissed all this as vestiges of earlier, less modern centuries, but this kind of transportation mix is precisely what we all need right now.
Unlike its fracture-critical predecessor, the bridge that replaced the I-35W represents the kind of constructive change that might guide us. The new bridge, designed by Linda Figg of Figg Engineering, has redundancy to spare (so to speak). 12 The sheer size and depth of its post-tensioned, concrete box beams not only compensate for the weakness of the previous bridge but also recognize that it’s better to build well than to need to rebuild. It is resilient — essentially two side-by-side, unconnected bridges — so if one side failed, the other would function. The bridge accommodates multiple transit modes, with some lanes strengthened to support future light-rail and a pedestrian suspension bridge planned for beneath the highway. The more alternatives a system offers, the more likely it is to last. And the new bridge arose from local conditions. These ranged from site specifics (the closed-off highway on one end became a construction yard for the new structure) to community concerns (local decision-makers offered input to the engineers) to job opportunities (the rebuilding employed local construction workers and material suppliers). The new I-35W Bridge exemplifies the deep advantages of fracture-resistant infrastructure. It also suggests that a resilient future will look like our more distant past than like the sleek future-utopian fantasies of popular culture.
Fracture-critical infrastructure, finance, housing, energy — all need attention, commitment, investment. But of course we face challenges still more basic. Jared Diamond, in Collapse, estimates that we have about 50 years before we experience the irreversible effects of exponential declines in natural habitats, fish populations, biological diversity and farmable soil; of serious shortages of fossil fuels, fresh water and plant growth per acre; of toxic chemicals in the air and water, invasive plant species devastating ecosystems, ozone-depleting atmospheric gases, impoverished human populations and unsustainable consumption. 13 Climate scientists like James Hansen warn we have even less time than once predicted to counter global warming and ensure that the planet remains habitable. 14 Is it an exaggeration to say our species itself is now fracture-critical? But of course we have the capacity to envision and create a better future. For a long time humans lived in resilient and sustainable ways, husbanding finite resources for future generations, cultivating renewable resources to maintain quantity and diversity, and encouraging pleasure in unquantifiable resources like community, creativity and empathy.
But the first challenge is to mind the warnings. The systems we’ve created to support civilization seem strong, even invincible. We never really expect the bridge to fall into the river.