On the night of July 13, 1977, lightning strikes on Con Edison’s transmission system in Westchester County started a sequence that left almost all of New York City without electricity for about a day. Looting and arson followed in several neighborhoods. The 1965 blackout had produced nothing on that scale, and the 2003 blackout did not repeat it. For an emergency manager, 1977 shows what happens when a city that is already short of police and fire staff loses its power supply and public order at the same time. It also shows how much of a response depends on electricity that nobody had planned around.
- July 13, 1977: lightning in Westchester and an hour to total collapse
- Twenty-five hours dark: why restoring a collapsed system is slow
- The city it hit: a fiscal crisis, thinned departments and a heat wave
- The looting and the fires: where it happened and what was asked of the departments
- Arrests, fires and damage: the repeated numbers and their limits
- Recall in the dark: calling people in when the trains and phones fail
- What Con Edison and the regulators examined, and what the city left
- 1965, 1977 and 2003: keeping three New York blackouts apart
- What to do at your agency
- Takeaways
July 13, 1977: lightning in Westchester and an hour to total collapse
Con Edison served New York City and most of Westchester County in 1977. On a hot summer evening the city drew a heavy load, and part of that load was met by power brought in over transmission lines from generating plants north of the city and from neighboring utilities. The standard account of the event, as summarized from the investigations, begins with a lightning strike at a substation near Buchanan on the Hudson River sometime after 8:30 p.m. The protective equipment that operated after the strike disconnected the output of the Indian Point generating station from the system. Further strikes in the following minutes took more transmission lines in Westchester out of service.
Each lost line shifted its flow onto the lines that were still in service, and those lines and the ties to neighboring systems began carrying more than they were rated for. Over the next hour the remaining connections opened, either by protective action or under overload. That left the Con Edison system isolated, and the generation inside the city could not carry the load still connected to it. When a system in that condition cannot shed load fast enough to match its remaining generation, frequency falls and the generators trip off to protect themselves. Most accounts put the final collapse at shortly after 9:30 p.m. By then almost the whole of Con Edison’s service area was dark. Parts of Queens served by the Long Island Lighting Company rather than Con Edison kept their power.
The detail that matters for anyone planning around utility failure is that this was not instantaneous. Roughly an hour passed between the first strike and the total loss of power. During that hour the system was visibly degrading, and operators were making decisions about it. Whether those decisions were adequate was the central question for the investigations, and I take it up below. For the response side, the hour matters for a simpler reason. Nothing in the record I have seen suggests that the city’s public safety agencies were warned during that interval that a citywide outage was coming. In 1977 a utility control room and a police or fire dispatch floor had no routine channel for that kind of warning, and many jurisdictions still do not have a formal one today.
Twenty-five hours dark: why restoring a collapsed system is slow
Power came back in stages through July 14, borough by borough and network by network. The last areas were restored late that evening, and the total time is commonly given as about 25 hours. Some neighborhoods got their power back many hours earlier than others, so “25 hours” describes the end of the event rather than what any one household experienced. The investigations reviewed the restoration as well as the collapse. Anyone who needs the specific sequence and the specific delays should read those reports rather than rely on newspaper summaries, which compressed the restoration into a few lines.
The general mechanics explain why a collapse of this kind takes far longer to repair than a single failed line. Once a system has gone completely dark, the generators need a source of power to restart. That source is either units able to start on their own or power fed in from a neighboring system. The transmission network then has to be re-energized in sections, and load picked up a piece at a time so that the generation that is running is not overwhelmed. Engineers call the last problem cold load pickup. A circuit that has been off for hours on a hot night comes back with every air conditioner, refrigerator and motor on it trying to start at once, and the demand is well above normal. If too much is reconnected too quickly, the protective equipment trips again and the work has to be repeated.
New York added a complication that most cities lack. Much of the Con Edison distribution system in Manhattan and other dense areas was underground, in networks built for reliability under normal conditions, and those networks had to be brought back carefully. The broader point holds for any utility. Restoration after a total collapse is a sequence of controlled steps, each of which can fail, and the time a utility gives at the start of an outage is a projection that will change as the steps succeed or fail.
An EOC that plans to the first restoration time a utility gives will be caught short whenever restoration runs long, and after a full system collapse it often does. The safer practice is to ask the utility liaison for the estimate and what it assumes. Plan generator fuel, staff relief and shelter decisions to the long end of the range, then scale back as power returns. That costs some fuel and overtime when restoration goes well. Planning to the short end costs much more when it does not.
The city it hit: a fiscal crisis, thinned departments and a heat wave
The New York of July 1977 was two years out of the fiscal crisis of 1975, when the city came close to default and had to accept state oversight of its finances and deep cuts in spending. Those cuts reached the uniformed services. The police department lost officers through layoffs and attrition, and the fire department closed companies and reduced staffing. The size of those reductions is reported differently depending on the source and on the date each source measured against. I am not giving a single staffing figure here, and anyone who needs one should go to the city’s budget documents and the departments’ own annual reports for 1975 through 1977. The general picture is not in dispute. Both departments entered the summer of 1977 smaller than they had been a few years earlier, and the fire department was already carrying a heavy load of structure fires in parts of the Bronx and Brooklyn.
The weather added to the strain. The blackout came during a stretch of hot weather, so it struck on a night when people were outdoors and electrical demand was high. The heat also mattered for the next day. Without power, the electric pumps that carried water up to many apartment buildings stopped, along with fans, refrigeration and air conditioning, for people who were already uncomfortable or medically fragile. The subway stopped wherever it was. Accounts describe passengers being led out of stalled trains, and people trapped in elevators across the city had to be found and freed.
The city’s government was also in the middle of a mayoral primary. Mayor Abraham Beame was seeking renomination, and the blackout and its aftermath became part of that campaign, which Beame lost that September. There was also no dedicated city emergency management office of the modern kind. New York City’s Office of Emergency Management, which today runs the city’s emergency operations center and coordinates multi-agency response, was created in 1996. In 1977, coordination ran through the mayor’s office, the police and fire departments, and whatever civil defense structure remained from earlier decades. Each department managed its own part of the problem largely through its own chain of command.
The looting and the fires: where it happened and what was asked of the departments
Looting began within a short time of the lights going out in a number of neighborhoods, and it continued through the night and into the next day in some places. The areas most often named in accounts of the night include Bushwick, Bedford-Stuyvesant and Crown Heights in Brooklyn, along with Harlem and parts of the South Bronx. Retail strips were hit hardest. Stores were broken into, emptied and in many cases set on fire, and the commercial stretch of Broadway in Bushwick is the place most often described as devastated. Some merchants stood guard over their own stores, and accounts describe confrontations between owners and looters.
For the fire department, the night meant a very large number of fires spread across several boroughs, many of them in occupied or mixed-use buildings, arriving while the department was smaller than it had been and while the water supply, traffic control and communications it normally relied on were degraded. Fires in looted stores could spread to the apartments above them and to adjoining buildings before companies arrived, because companies were committed elsewhere and travel through dark streets was slow. Firefighting in a neighborhood where looting is active also raises the question of crew safety and police protection, and the police department was simultaneously trying to make arrests, protect property and manage traffic without signals.
The reasons the looting happened in 1977 and not in 1965 have been argued about ever since, and I am not going to rank them. Commentators at the time and researchers since have pointed to poverty and unemployment in the affected neighborhoods, the cuts that followed the fiscal crisis, the heat, the speed with which looting spread once it started, and the visible absence of police in some areas in the first hours. Robert Curvin and Bruce Porter’s 1979 study, Blackout Looting!, is the most frequently cited extended examination, and anyone teaching the event should read it directly rather than through secondary summaries. For an emergency manager, the useful lesson does not depend on settling the cause. A utility outage in a stressed city can become a public order event within an hour, and the plan for the outage has to assume that this may happen.
Arrests, fires and damage: the repeated numbers and their limits
The figures that appear most often in later news retrospectives and reference works are 3,776 arrests, 1,037 fires and 1,616 stores damaged or looted. A damage figure of a little over $300 million is also widely repeated and is usually attributed to a later congressional study. These are the numbers a reader will meet first, and they are reasonable to use with attribution. I have not been able to trace each of them to a single primary tally, though, such as a police department arrest log, a fire department incident count or the text of the congressional study. That is the honest basis on which I am giving them here.
Several things make the figures less settled than they look. An arrest total depends on when it was taken. Arrests continued after the first night, and a count published on July 14 or 15 will differ from one compiled weeks later. A fire count depends on what the department counted as a fire and over what period, and whether it includes rubbish and vehicle fires along with structure fires. A store count depends on whether damaged stores, looted stores and burned stores are counted together or separately. Damage estimates are the least stable of all, because they mix insured losses, uninsured losses, public costs such as police and fire overtime, and economic losses from closed businesses, and different studies have made different choices about what to include. Some accounts give arrest totals above the figure I quote, without stating their basis. Where a source gives a number different from the common one, cite that source by name and say that it differs.
The figures for the 1965 and 2003 blackouts are occasionally mixed in with these, usually by accident. They belong to different events with different durations and different footprints, and they should never be compared to 1977 without saying what each count measured. If a training slide puts a 1977 arrest total beside a 2003 arrest total to make a point about crime, it should say who counted each one and over what period. Without that, the comparison may be measuring the difference between two record-keeping systems rather than two nights.
For every historical disaster figure in a presentation, put three things beside the number: who published it, when, and what it counted. For 1977, that means writing “3,776 arrests, the figure most often repeated in later accounts” rather than “3,776 arrests”. It adds a few words to the slide. It also prevents the argument that starts when someone in the room has read a different figure in a source that counted the second day.
Recall in the dark: calling people in when the trains and phones fail
When a city’s uniformed services are short-staffed, the first tool they reach for in a major event is the recall of off-duty personnel. New York did that in 1977. The police department ordered off-duty officers to report, and the fire department brought in additional personnel. Accounts of the night describe many officers arriving late or not at all, and the recall was criticized afterward. I have not seen an official figure for how many officers were recalled and how many reported within a given time, and the various reconstructions disagree. I am leaving the numbers out. The department’s own records and the later reviews are where they would be found.
The reasons a recall underperforms in a blackout are structural, so they apply well beyond New York in 1977. Notification depends on telephones, and today on cellular networks, both of which degrade during a large outage as backup power runs down and traffic climbs. Travel depends on mass transit, traffic signals and the roads being clear. In 1977 the subway was stopped, and personnel who lived outside the city or relied on the trains had no easy way in. Responders also have families in the same dark city, and some will reasonably check on them before reporting. A recall plan that assumes everyone hears the call and can reach their assigned station in normal time will overstate the people actually available in the first hours, which is when a disturbance either grows or is contained.
The second staffing problem in a long outage is relief. A 25-hour event means the people on duty when the power failed either work through the night and the next day, or are relieved by people who could get in. In 1977 both departments were already working with fewer people, so every hour of overtime on the first night was an hour of fatigue on the second day. Restoration also does not end the event for public safety, because arrests have to be processed, damaged buildings secured and fire watch maintained after the lights come back. Planning the second and third operational periods is as important as the first, and it is the part most often left out of short outage plans.
Most agencies know their authorized strength and have a recall roster, but few know what share of that roster actually reports within one hour, two hours and four hours when commercial power and transit are degraded. That share can be estimated from a callback drill and from home addresses, and it should be written into the outage annex as a planning figure. An annex that assumes full recall is describing an agency that will not exist on the night it is needed.
What Con Edison and the regulators examined, and what the city left
The 1977 blackout was investigated by Con Edison itself, by the New York State Public Service Commission, and at the federal level. The federal inquiry began under the Federal Power Commission, and its responsibilities passed to the new Department of Energy and the Federal Energy Regulatory Commission when the commission was abolished later in 1977. I am describing these reports rather than citing report numbers or section references, because I have not verified the document identifiers. Anyone who needs the text should look for the Public Service Commission’s investigation of the July 1977 outage and the federal staff report through a library with government documents holdings, or through the agencies themselves.
In broad terms, the reviews accepted that lightning started the sequence. They then examined why a set of lightning strikes on part of the system became a total collapse over the following hour. The reviewers looked at how protective equipment behaved and whether some of it was in poor condition. The operation of the ties to neighboring systems came under scrutiny too, as did the timing and amount of load shedding by Con Edison’s system operators. Load shedding, deliberately disconnecting some customers so that the rest of the system can survive, is the operator’s main tool once generation and transmission no longer match demand. The reviews were critical of how it was used that night. Readers who want the specific findings and the weight each review gave them should read the reports themselves, because summaries of these findings tend to flatten disagreements among the reviewers and between the reviewers and the utility.
The public argument ran alongside the investigations. Con Edison’s chairman, Charles Luce, initially described the lightning as an act of God, and Mayor Beame publicly accused the utility of gross negligence, according to contemporary news reporting. Businesses that suffered losses sued the utility, and the litigation went on for years. Anyone who needs the outcome of those cases should read the New York appellate decisions on blackout claims rather than a summary. On the city side, I have not been able to identify a single comprehensive after-action report from city government comparable to the utility and regulatory reviews. The city’s material appears in public statements, departmental records, later hearings and the police and fire departments’ own internal reviews, and the NYC Municipal Archives is the place to start looking for it.
1965, 1977 and 2003: keeping three New York blackouts apart
New York City has lost power on a large scale three times in living memory, and training material often confuses them. The Northeast blackout of November 9, 1965, began at about 5:16 p.m. when a protective relay operated at a generating station near Niagara Falls in Ontario. The resulting cascade spread across much of the northeastern United States and parts of Ontario. In New York City, power was restored overnight in most areas, and the night is remembered for its general calm. That event led to the formation in 1968 of the National Electric Reliability Council, the predecessor of today’s North American Electric Reliability Corporation.
A myth attaches to 1965, and it should be labeled as a myth whenever it comes up. The story that births in New York rose nine months after the blackout began with newspaper reporting in 1966. A demographer, J. Richard Udry, later examined birth records and found no rise attributable to the blackout. The story still circulates, and it has sometimes migrated to 1977 as well. It has no support for either event.
The blackout of August 14, 2003, began in Ohio. The joint U.S. and Canadian task force that investigated it found that transmission lines contacting trees, together with a failed alarm system in FirstEnergy’s control room, let a local problem spread across much of the Northeast and Ontario. New York City lost power in the late afternoon, and service came back over the following day. The city saw little of the looting and arson that marked 1977. That outcome is often credited to changed conditions in the city, a different policing posture and a far more developed emergency management structure, though that is an interpretation, and people who have studied both events weigh the factors differently. The 2003 event also led to mandatory reliability standards for the bulk power system under the Energy Policy Act of 2005. Taken together, the three events differ in where they started, how long they lasted, what time of day and year they struck, and what followed. The 1977 event is the only one of the three that was confined largely to one utility’s territory, and the only one remembered chiefly for disorder.
What to do at your agency
- The radio system administrator pulls the generator run logs and current fuel levels for the PSAP and every radio site, and the task is finished when each site has a written figure for hours of autonomy at current fuel and the shortest one is sent to the emergency manager.
- Each shift commander runs one unannounced callback test of off-duty personnel during an evening shift this month, and the task is finished when the number who confirmed and the number who could report within two hours are recorded and given to the operations chief as the planning figure for the outage annex.
- The emergency manager asks the electric utility’s emergency liaison for the written section of its emergency restoration plan that covers a total system loss, and the task is finished when the county’s critical facility list has been checked against the utility’s priority list and any facility missing from it has been reported to the utility in writing.
- The fire operations chief and the police patrol commander add one item to their next regular coordination meeting covering fire company protection and staging during an active civil disturbance, and the task is finished when one written paragraph naming the staging locations and the request procedure is placed in the existing civil disturbance annex.
- The PSAP director confirms that the UPS and generator transfer for the CAD servers and dispatch consoles were tested under real load within the last twelve months, and the task is finished when the date and result of that test are on file or a test date is scheduled with facilities.
- The training officer reviews any blackout material in the department’s courses, and the task is finished when the 1965, 1977 and 2003 events are clearly separated, every arrest and damage figure carries its source, and the 1965 birth-rate story is either removed or labeled as a myth.
Takeaways
- The July 13, 1977, blackout began with lightning strikes on Con Edison’s system in Westchester County and became a total collapse of the city’s supply roughly an hour later.
- Power was restored in stages through July 14, and the commonly given total of about 25 hours describes when the last areas came back rather than what every neighborhood experienced.
- Restoring a fully collapsed system is a sequence of controlled steps that can each fail, so an EOC should plan fuel, staffing and sheltering to the long end of any restoration estimate.
- New York entered the blackout with police and fire departments reduced by the 1975 fiscal crisis, during a heat wave, and without the city emergency management office that was created in 1996.
- The most often repeated figures of 3,776 arrests, 1,037 fires and 1,616 stores damaged should be cited with their source and basis, because counts vary with when and how they were taken.
- Recall of off-duty personnel underperformed in 1977 for reasons that still apply, because notification and travel both depend on power, telephones and transit that fail in a large outage.
- The utility and regulatory investigations examined the hour between the first strike and the collapse, including load shedding and equipment performance, and their reports should be read directly rather than through summaries.
- The 1965, 1977 and 2003 blackouts differ in origin, duration and aftermath, and the story of a 1965 birth surge is a myth that should be labeled as one.
Reach me through the contact page. I read every message.
