When the electrical grid went down in Manhattan during Superstorm Sandy, the effects on the city’s infrastructure were tremendous. Cellular communication towers failed to transmit, drinking water did not reach high-rise apartments, toilets did not flush, streetlights and traffic signals went out, and gasoline pumps stopped running. The Sandy-affected area extended well beyond the FEMA 100-year floodplain maps, and almost everyone living in the region realized that they were, in some way, vulnerable to the impacts of a flood. Electricity turned out to be a lynchpin to urban resiliency, which, when disrupted, sent shockwaves upland and downstream.

Not all neighborhoods in New York felt the power-outage shockwave equally. The service boundary between the haves and have-nots ran not across racial, social, or economic lines, but those of electrical grid service territories. For once Brooklyn, Queens, and the Bronx were spared while Manhattan suffered. The famously explosive outage of the East 13th Street electrical substation left Lower Manhattan without power for four days. Stories of able-bodied residents trekking flight after flight of stairs abound. Hospitals had to be evacuated. Wall Street was closed for over a week. Refrigerators and grocery-store aisles stored millions of pounds of rotten food.

On the surface, the story of this catastrophic substation explosion is straightforward: the station flooded during Sandy, causing a short circuit that took four days to repair. But digging a bit deeper into the tales of two substations offers a more nuanced story that illuminates a more effective path to resiliency, energy planning, and decision making during disaster situations.

The Farragut Substation

The Farragut substation near the Brooklyn Navy Yard provides over a million people with electricity and is well within the FEMA 100-year flood plain. Because of its proximity to the East River and the obvious risk of flooding, the Farragut Substation was preventatively shut down many hours before the peak storm tide. Thus when the substation eventually flooded, it was not electrically charged. Some equipment was inundated during the event, but when the water receded, it dried out to reveal minimal direct flood damage. By noon the day after the storm the substation was back online, as was most of Brooklyn along with it.

The East 13th Street Substation

But the story is quite different across the East River in Manhattan. The East 13th Street substation lies just within the FEMA 100-year floodplain boundary, so the risk of flooding was technically a Rumsfeldian “known known.” The substation serves Lower Manhattan from the Battery up to 39th Street, supporting over 600,000 residents, several major hospitals, thousands of businesses, and the heart of Wall Street. So with a major storm approaching this substation, which serves such an important service area and is located within the floodplain, why was nothing done to protect it?

First, despite officially being in the floodplain, floodwaters had never actually reached this substation before. Many major storms, such as Irene just one year earlier, had not even come close. Though Sandy’s wind speeds did not reach hurricane levels, its strong winds coincided with high tide and moved in such a direction as the storm passed over the New York Bight that they caused a storm surge over 3 feet higher than any previously recorded at the National Oceanic and Atmospheric Administration’s observation station near the Battery. Sandy was generally considered to be a 1-in-70-year event today, though some projections of 2050 climate and sea level rise predict it will be a 1-in-a-40-year event instead.

Second, there were few options for protection at the moment of the storm: sandbags to attempt to hold back the surge, or a preventative shutdown similar to the ones Brooklyn and Queens. Unlike the Farragut substation service area, which contains predominantly low-rise and brownstone-scale buildings, many residences in this territory are in high-rises. Buildings over 4 stories high require elevators to access homes and pumps to deliver water. Because of ConEd’s historical reliability in Manhattan (it is one of the most reliable electric service areas in the country), most buildings in the East 13th Street service territory lack backup generators beyond those required for fire codes, which last for only a few hours and serve only small emergency demands. Thus a preventative power outage here would have left many more residents without elevator service, drinking water, flushing toilets, or other life-safety essentials.

But possibly the most important factor in the decision not to shut down the substation was that it services not only residents and businesses, but also a number of critical infrastructures such as the Hospital Row complex just north of the station, as well as major telecommunications hubs and other facilities deemed essential. ConEd considers very seriously before proactively shutting down primary substations. The assessment criteria in Manhattan were different from those in Brooklyn and Queens for good reasons. The irony of this particular event, in hindsight, is that being without power for twelve hours would have saved four days of power outage for all those living and working below 39th Street. But had Sandy hit a few hours earlier or later, or taken a slightly different route and resulted in less severe flooding, ConEd may have been criticized for being overly cautious and blamed for the consequences of an unnecessary shutdown.

Hurricane Sandy’s Wider Impact

Although a lot of attention was given to power outages and disruptions during and after Sandy, it is a testament to New York City’s power grid that most functionality was normalized by noon the day after the storm. The discrete outages that remained, such as those at the 13th Street substation and in the flooded subway tunnels, were significant but relatively rare. ConEd and the Metropolitan Transportation Authority are continuing to address these weak points.

The real failure of the power grid was on Long Island and New Jersey, where trees fell on above-ground power lines as they usually do during a storm—but this time on a scale never seen before. In order to bring a section of the grid back online in an area dominated by above-ground wires, the utility provider must ensure that each and every line in that section is safe before electrifying it. Lines with trees still leaning on them—or worse, lines that are broken and hanging or lying on the ground—put the public in serious danger. It is no wonder it took over a month to repair and inspect every overhead line throughout the entirety of Long Island.

This delay in Long Island and New Jersey’s power restoration disrupted gasoline supplies and communications throughout the region. Fiber-optic cables are inherently waterproof and stored underground, so communication networks suffered only discrete damage. For instance, a major telecommunications hub on Water Street in Manhattan was flooded and out of service for months after the storm. The communication apparatus itself was largely unharmed, but the building was put out of service while technicians replaced mechanical and electrical equipment in the basement. Once power was restored to Lower Manhattan, most fiber-optic and cellular communications networks were up and running.

The story for gasoline shortages parallels that of communications. Underground, waterproof distribution networks were left unscathed by the storm, but storage facilities and pumping stations located in the ports on Long Island and New Jersey were flooded and damaged. And again, even after storage facilities were repaired, networks remained down until the power grid came back online. On Long Island it took weeks to bring power back to the primary gasoline storage area, and over a month to bring it back to all of the gas stations throughout the island.

Many Costly Solutions

A common gut reaction after Sandy’s devastating events has been to recommend building protective measures against the next storm. Proposals for building levees and seawalls, elevating mechanical equipment, and even lifting entire buildings fill news coverage on how to prepare for next time. Singular assets such as electrical substations demand highly reliable and costly protective measures, because their protection provides enormous value. But assets such as  building lobbies and ground-floor retail stores, even with their expensive finishes and valuable inventories, justify protective investment that are relatively much lower.

It is important to keep in mind the relationship between the probability of extreme natural disasters occurring and the level of devastation that they inflict. The larger the storm, the less likely it will happen in any given year. Thus costly protective measures must be weighed against the value that they protect multiplied by the likelihood of storm occurrence. Insurance companies call this value the “annual expected loss” of a given property or urban area. Of course losses do not occur each year, but this averaged value provides an amortized number for which insurance companies can set premiums, or a figure that property owners can weigh against a single, upfront payment for elevating mechanical equipment or building a levee wall. Sometimes leaving an asset unprotected but insuring it against known risks provides the least-cost solution, particularly for businesses and individuals with limited capital.

A number of lessons can be taken from this set of experiences. The entire grid does not need to function for all customers all the time, particularly during extreme weather events. What we do need to ensure is that critical infrastructure, like hospitals and emergency services, can be served with electrical power from diverse, redundant sources in addition to the ubiquitous power grid. To be sure, NYU Langone Medical Center and the other essential infrastructures that experienced distinct failures during Hurricane Sandy will have fully functioning backup generators during the next storm event. But those critical facilities that narrowly escaped failure during Sandy should take note, as the next storm will likely have a different trajectory and could affect their performance.

The Permanent Temporary Solution

Sometimes the best long-term solution is a temporary solution. Temporary protections include operational changes, such as the Farragut station’s preventative shutdown (although that response did not address direct property damage from flooding). Another example is the sort of temporary wall used in flood-prone, historic Europe towns. These flood walls can be installed quickly with just 24 or 48 hours’ warning, and range in height from 3 feet up to 10 feet in some cases. They require diligent maintenance and skilled installation, but provide a much better level of protection than sand-filled or even water-filled temporary barriers.

SIRR Leads the Way

Just months after Hurricane Sandy, Mayor Bloomberg created the Special Initiative for Rebuilding and Resiliency (SIRR). This task force worked tirelessly for nearly six months developing a comprehensive disaster-response plan for the city. The recommendations in its report ranged from small-scale operational changes to revisions in the New York City Building Code, to a major multi-purpose levee wall around Lower Manhattan. The SIRR report was the most thorough and comprehensive planning document of any municipality in the Sandy-affected area. Based on rigorous analytical research and thorough engagement with a wide range of communities and stakeholders, the 438-page report is a model of the long-range, integrated planning typical of the Bloomberg administration.

Although the task force intentionally did not prioritize among the report’s 195 initiatives and site-specific designs, some do deserve close attention, particularly ones in the chapters on power and data. Initiative 13 in the “Utilities” chapter recommends speeding up the implementation of smart-grid technology. This technology provides real-time information to operators so that the grid can be more efficiently and effectively controlled. In a disaster scenario, substations can be powered down more quickly, or power can be re-routed from more reliable sources to maintain a secure network. The other recommendations that require attention are Initiatives 1 and 2, which call for integrating consideration of climate change into the regulatory decision-making process. Utility providers must justify their investments and upgrades to regulatory bodies—they cannot make unilateral spending decisions to strengthen network resiliency. Often regulators limit investments or do not fully account for the effects of climate change—such as increased storm events—during their approvals processes. This, more than any other initiative, would change the shape of today’s electrical grid.

Proper Preventative Preparation

Many infrastructure networks, businesses, and residents sufficiently prepared themselves in the days before Sandy as forecasts predicted its path and intensity. Subway service stopped 24 hours before the event and trains were parked in upland rail yards. Many low-lying businesses boarded up their windows and moved inventories to higher ground. New Yorkers emptied bodegas of batteries, water, and nonperishable staples. As discussed, the Farrugut substation in Brooklyn was taken offline nearly 8 hours before the height of the storm. But in the case of the East 13th Street substation in Manhattan, the continuance of full electrical power was deemed essential and the decision was made to forego a preventative shutdown.

Many lessons can be drawn from this tale of two substations. First, a resilient network never relies on one central component for its many critical parts to function. Dependencies should be shared, interconnected, and backed up. Critical components of urban social and physical infrastructure networks, such as hospitals, water pumps, emergency lighting, and communications, must be provided sources of electricity that follow this principle of resiliency. Second, resilient networks should have components that can safely fail. It should be possible to disconnect, shut down, or permanently jettison these components if necessary.

Sandy was the worst natural disaster to ever hit the New York region. As designers and policy makers, we should be careful to recommend investments and policies based on such an extreme event. But as extreme weather events become more and more common, we may see many subsequent events take that mantle of “worst natural disaster to ever hit the region.”