Late on the evening of 2 December 1984, water entered a storage tank holding roughly 42 tonnes of methyl isocyanate at the Union Carbide India Limited pesticide plant in Bhopal, Madhya Pradesh, and in the early hours of 3 December that tank vented to the atmosphere over dense neighborhoods that began at the plant fence. Nobody told the people in those neighborhoods what the gas was, whether to run or stay inside, or what would help. This piece covers the release, the absent warning, the clinical improvisation in the hospitals, the contested death toll, and the American community right-to-know law that came out of it.

The plant, the tank, and the condition of its safety systems

The Bhopal plant was operated by Union Carbide India Limited, a company in which the American parent, Union Carbide Corporation, held a majority stake of roughly 51 percent, with the balance held by Indian investors and Indian state-owned insurers. The facility formulated and, from the early 1980s, manufactured the carbamate insecticide carbaryl, sold under the Sevin trade name, and the process route ran through methyl isocyanate, a low-boiling, highly reactive liquid that reacts violently with water and whose vapor is about twice as dense as air. Methyl isocyanate was stored on site in three partly buried tanks designated 610, 611 and 619, and tank 610 held on the order of 42 tonnes on the night of the release.

By late 1984 the plant was running well below capacity, carbaryl demand had fallen, staffing and maintenance budgets had been cut, and several of the engineered protections around the methyl isocyanate inventory were either shut down or not in a condition to function. The refrigeration system intended to keep the stored liquid near zero degrees Celsius had been out of service for months, with the refrigerant reportedly removed for use elsewhere. The vent gas scrubber, which was meant to neutralize escaping isocyanate with caustic, was not in a state of operational readiness and was in any case sized for a small process leak rather than a runaway tank. The flare tower downstream was out of service because a section of connecting pipe had corroded and been removed, and the water spray curtain could not reach the height at which the gas actually left the stack.

None of this was a secret to everyone. A safety audit conducted in 1982 by a team from the American parent identified a substantial list of hazards at the plant, including concerns about the potential for a large release. A worker, Mohammed Ashraf, had died after a phosgene exposure at the plant in 1981. A local journalist, Rajkumar Keswani, published a series of articles in Bhopal papers between 1982 and 1984 warning that the plant could destroy the city, and those pieces reached readers without reaching anyone with the authority to act on them.

The night of 2 and 3 December 1984

Late on 2 December, workers in the methyl isocyanate area noticed eye irritation and a smell consistent with a leak, which was not by itself unusual at that plant, and the shift accounts describe a search for the source and a decision to deal with it after the tea break. Water was by then already inside tank 610, and the reaction between water and methyl isocyanate is strongly exothermic and generates gas, so temperature and pressure in the tank climbed together. Contaminants in the tank, including iron from corroding vessel walls, are generally understood to have catalyzed further reactions once the temperature rose, which accelerated the process.

Shortly after midnight the tank pressure exceeded the relief setting and the contents discharged through the relief valve, the scrubber and the stack into the open air. The venting continued for something on the order of one to two hours before the tank emptied. Published estimates of the mass released commonly fall between roughly 30 and 40 tonnes of methyl isocyanate together with reaction and decomposition products, and the exact composition of what came out of the stack has never been fully characterized, which matters clinically and I will come back to it.

The meteorology could not have been worse. Winds were light, the night air was stable, and the gas, being heavier than air, stayed low and drifted slowly over the settlements immediately south and east of the plant, including JP Nagar, Kazi Camp and Chola Kenchi, and on toward the Bhopal railway station and the old city. People woke up unable to breathe, with burning eyes and vomiting, in the dark, in an area with narrow lanes and no street lighting in many blocks. Railway staff at the station are widely reported to have stopped incoming trains from discharging passengers into the cloud, at the cost of their own exposure, and the station master, Harish Dhurve, died.

No warning, and no guidance on what to do

There was no community warning system in any meaningful sense. Accounts of the night describe the plant siren being sounded and then silenced, with a public siren following later, and the plant reportedly used the same siren for routine purposes such as shift changes and drills, so residents had no basis for distinguishing an emergency signal from ordinary noise. I would treat the precise siren sequence as contested in the literature rather than settled, and the operationally important point does not depend on resolving it, because even a correctly sounded siren would have conveyed nothing actionable to a population that had never been told what the plant made, what it would do to them, or what to do about it.

No emergency plan existed that covered the neighborhoods outside the fence. There was no off-site notification protocol to the district administration, no pre-scripted public message, no designated shelter or evacuation route, and no arrangement with the police or the municipality to move people crosswind. District officials, police and hospital staff learned what was happening from the casualties arriving in front of them. The people living closest to the plant were, in many cases, those with the least ability to leave, in unauthorized settlements that had grown against an industrial boundary because that was where the land was available.

The absence I want to underline is informational rather than mechanical. Residents did not know the name of the chemical, did not know it was heavier than air, did not know that a wet cloth over the face offered partial protection, and did not know which direction was away from it. Every one of those items is cheap to distribute and expensive to withhold, and the Bhopal case is the reason that public information about hazardous materials stored in a community stopped being treated as a courtesy in American law two years later.

Hospitals treating a gas nobody would name

Hamidia Hospital and Jawaharlal Nehru Hospital took the bulk of the load, and the volume of patients in the first hours exceeded anything either institution had planned for, with figures commonly cited in the range of tens of thousands seen within the first day and over a hundred thousand in the first several days. The presenting picture was severe conjunctival and corneal injury, laryngeal and tracheal irritation, bronchospasm, and pulmonary edema, with deaths occurring from respiratory failure. Clinicians did what the picture indicated, which was oxygen, bronchodilators, steroids in many cases, and eye irrigation, all of it supportive because there is no antidote to methyl isocyanate.

What they could not get was the identity and composition of the exposure. Accounts of the night report that hospital staff telephoning the plant were told the gas was not poisonous and that washing the eyes with water would suffice, and I present that as what is reported in the published accounts rather than as a verified transcript of any single conversation. Union Carbide did not have, and India did not require, a mechanism for pushing toxicological information to a treating physician at three in the morning. Physicians in Bhopal were reduced to inferring the toxicology from the patients in front of them, which is a defensible clinical approach and a terrible substitute for knowing what is in the cloud.

The sodium thiosulfate dispute grew directly out of that gap. Some clinicians, including the German toxicologist Max Daunderer who traveled to Bhopal, and subsequently the Indian Council of Medical Research, advocated intravenous sodium thiosulfate on the theory that cyanide-like metabolites contributed to the systemic toxicity, and reported clinical improvement in treated patients. Union Carbide’s position and that of other authorities was that the cyanide mechanism was not established, the recommendation was withdrawn and reinstated at different points, and the treatment became entangled with the litigation. The efficacy question has never been settled to everybody’s satisfaction, so I would send anyone who needs to resolve it to the ICMR’s published research and the peer-reviewed toxicology literature rather than to any secondary summary, including this one.

Counting the dead where the numbers are contested

The Bhopal death toll is not merely uncertain, it is politically loaded, and the numbers published at each end of the range come from institutions with different interests and different counting rules. Bodies were buried and cremated in large numbers in the first days without registration, many of the exposed were migrants and unauthorized residents who appeared in no municipal roll, and deaths from the exposure continued for years, so the question of where to close the count determines the answer.

At the low end, an immediate official toll of roughly 2,259 deaths was recorded in the first days from hospital and municipal counts. The Madhya Pradesh state government has long carried an official figure of 3,787 deaths attributed to the gas release. An affidavit filed by the Government of India in the Supreme Court in 2006 stated the leak had caused 3,787 deaths and some 558,125 injuries, of which roughly 3,900 were classified as severely and permanently disabling. Separately, the claims machinery established under Indian law awarded compensation in a larger number of death cases, with the figure commonly reported at around 5,295, and the difference between that number and 3,787 is a difference in what was being counted rather than a simple contradiction.

At the high end, Amnesty International’s 2004 report on the twentieth anniversary estimated that roughly 7,000 people died within days of the release and about 15,000 more in the years that followed, with over 100,000 people suffering chronic illness. Survivor and campaign organizations, including the International Campaign for Justice in Bhopal, publish totals in the region of 25,000. Those figures and the Indian government’s figures disagree by a factor of five or more, and averaging them would be dishonest. Anyone who has to cite a number needs to publish the range with the publisher named at each end of it and the counting rule stated alongside, and the underlying records sit with the Madhya Pradesh Directorate of Claims, the ICMR’s long-term cohort studies, and the Supreme Court of India’s docket in the settlement litigation.

What is agreed about the mechanism and what is still disputed

The physical mechanism is not in dispute. Water reached the methyl isocyanate in tank 610, the exothermic reaction produced heat and gas faster than the tank could contain it, and the engineered mitigation downstream was incapable of handling the resulting flow. How the water got into the tank is genuinely disputed, and the dispute has never been resolved by any tribunal in a way that both sides accept.

The Indian government’s position, supported by most independent analyses and by the accounts of workers on shift, is that water entered through the process piping during or after a water-washing operation on lines connected to the tank system, made possible by missing or ineffective isolation, in a plant whose maintenance and instrumentation had been allowed to degrade. Union Carbide Corporation maintained, on the basis of an investigation it commissioned and presented publicly in 1988 through Arthur D. Little, that a disgruntled employee deliberately connected a water hose to the tank and that the release was sabotage rather than an operating failure. That is a company’s contested contention rather than a finding by an independent investigative body, and it should be read as one party’s account of the origin.

The legal record is separate from the causal one. India enacted the Bhopal Gas Leak Disaster (Processing of Claims) Act in 1985, making the Union of India the sole legal representative of the victims, and in 1989 the Supreme Court of India approved a settlement of 470 million United States dollars between the government and Union Carbide, which extinguished the civil claims and which survivor organizations have contested ever since. Criminal proceedings continued, and in June 2010 an Indian trial court convicted seven former UCIL officials of causing death by negligence, with two-year sentences and fines, verdicts that went on appeal. Warren Anderson, the corporation’s chairman at the time, was never tried in India and died in 2014, and a curative petition by the Union of India seeking additional compensation was dismissed by the Supreme Court of India in March 2023.

The American answer: EPCRA and SARA Title III

Bhopal landed in the United States as a question about the plants next door. Union Carbide operated a methyl isocyanate unit at Institute, West Virginia, and in August 1985 a release of aldicarb oxime and methylene chloride from that plant sent a reported figure in excess of a hundred people to area hospitals, which removed any remaining political argument that this was a problem confined to another country. The Environmental Protection Agency had already stood up a voluntary Chemical Emergency Preparedness Program in 1985 with a list of acutely toxic chemicals, and several states moved first, New Jersey with its Toxic Catastrophe Prevention Act and California with a risk management and prevention program.

Congress then enacted the Emergency Planning and Community Right-to-Know Act of 1986, passed as Title III of the Superfund Amendments and Reauthorization Act, and the structure it created is the one American responders still work inside. It required each state to establish a State Emergency Response Commission, which in turn designated emergency planning districts and appointed a Local Emergency Planning Committee for each, with statutory membership drawn from elected officials, fire, police, EMS, public health, hospitals, broadcast and print media, community groups, and the regulated facilities themselves. Each committee had to write and annually review an off-site emergency response plan covering facilities holding extremely hazardous substances above threshold planning quantities, and each covered facility had to name an emergency coordinator to work with that committee.

The information provisions are the part that carries the Bhopal fingerprint most directly. Facilities must give immediate notice of a reportable release to the LEPC and SERC and, for certain substances, to the National Response Center, with a written follow-up. They must provide safety data sheets or chemical lists and annual chemical inventory reports, generally filed as Tier II, to the LEPC, the SERC and the local fire department, and those inventories are available to the public on request. A separate section created the Toxics Release Inventory, an annual, publicly searchable database of routine releases and transfers by facility, which is the first time an American resident could look up what a neighboring plant was putting into the air without asking anyone’s permission. The Clean Air Act Amendments of 1990 added the section 112(r) general duty clause and the Risk Management Program rule in 40 CFR part 68, and authorized the Chemical Safety and Hazard Investigation Board, while OSHA issued the Process Safety Management standard at 29 CFR 1910.119 in 1992. Reporting thresholds, forms and deadlines have changed repeatedly since, so verify the current requirements with EPA and your state emergency response commission rather than with any summary of the original statute.

Where community right to know actually stands now

The framework works when somebody uses it, and the failure mode I have seen most often is a Local Emergency Planning Committee that exists on the state roster, has a chair who was appointed years ago, and has not met, reviewed its plan, or run a commodity flow study in longer than anyone can document. Tier II reports arrive annually, get filed, and are never read against a map, so the officer who arrives first at a facility fire has no idea what is inside until somebody finds the plate on the tank. Correcting that costs staff time rather than money, and it is the single highest-yield hazmat preparedness task available to a small county.

Public access has also narrowed in one specific area. The offsite consequence analysis portion of a facility’s risk management plan, which is the part describing worst-case release scenarios and the population that could be affected, was restricted from broad public dissemination by federal legislation in 1999 out of concern that it functioned as a target list, and access tightened further after 2001. The material still exists and can be reviewed under controlled conditions, so check with EPA on the current procedure for obtaining it, but the practical result is that the part of the record closest to the Bhopal question of who lives inside the plume is the part a resident cannot simply download.

On the warning side, the tools now exist that Bhopal lacked entirely, including the Integrated Public Alert and Warning System, Wireless Emergency Alerts geotargeted to a polygon, outdoor sirens where communities maintain them, and telephone notification systems. What has not automatically followed is the content. An alert that says a chemical release has occurred and to shelter in place is close to useless if nobody in the neighborhood has ever been told what shelter in place means, that it involves closing windows and shutting off the HVAC system, that it is different from evacuation, and that in some releases going up is better than going out. That public education work belongs to the LEPC and the local emergency management office, it is dull, and it has to be done before the pager goes off.

What to do at your agency

• Have your emergency manager or fire marshal request the current Tier II filings from the Local Emergency Planning Committee or the state emergency response commission, and produce a single map plotting every reporting facility in the jurisdiction with its extremely hazardous substances and quantities.
• Find out from your county emergency manager when your Local Emergency Planning Committee last met and last reviewed its off-site plan, and if the answer is more than a year, put the plan review on the agenda of the next meeting of whatever multi-agency body already convenes in your county.
• Call the emergency coordinator named for the largest chemical facility you protect, confirm the 24 hour number is answered, ask who can provide chemical identity and quantity to a hospital physician at night, and record that name and number in the CAD contact file with the date verified.
• Ask the emergency department medical director at your receiving hospital how the department would obtain toxicological guidance during a chemical mass casualty event, confirm the current poison control center number is posted, and note whether the hospital has your jurisdiction’s facility list.
• Have the communications supervisor run one test message through your alerting system using shelter-in-place language, and check whether the text actually tells a resident to close windows, shut off heating and air conditioning, and stay off the phone, rather than only naming the hazard.
• Write one paragraph into your existing emergency operations plan hazardous materials annex specifying who is authorized to order shelter in place versus evacuation for an off-site chemical plume, and how that decision reaches the public within ten minutes.
• Deliver one public information item this quarter, at a school, a housing complex or a community meeting inside a plume footprint you have already mapped, explaining what is stored nearby and what to do, and document the date and attendance.

Takeaways

• Water entered a tank holding roughly 42 tonnes of methyl isocyanate at the Union Carbide India Limited plant in Bhopal late on 2 December 1984, and the resulting runaway reaction vented an estimated 30 to 40 tonnes of gas and reaction products over adjacent neighborhoods in the early hours of 3 December.
• The refrigeration system was out of service, the vent gas scrubber was neither ready nor sized for the event, and the flare tower was disconnected, so the release reached the community essentially unmitigated.
• There was no community warning capability worth the name, no off-site emergency plan, no designated shelter or evacuation route, and no public education, so residents ran into a vapor that was heavier than air instead of sheltering indoors and covering their faces.
• Bhopal’s physicians treated tens of thousands of patients without a confirmed chemical identity or composition, with no antidote available, and the sodium thiosulfate controversy that followed remains unresolved in the literature.
• The death toll is contested across a wide range, with an immediate official count of roughly 2,259, a Madhya Pradesh government figure of 3,787, compensation awarded in about 5,295 death cases, and Amnesty International’s 2004 estimate of roughly 7,000 deaths within days and about 15,000 more in later years, while survivor organizations publish figures near 25,000.
• The physical mechanism of water contacting methyl isocyanate is agreed, while the route by which the water entered the tank is disputed, with the Indian government and most independent analyses pointing to maintenance and isolation failures and Union Carbide contending it was deliberate sabotage.
• The United States responded with the Emergency Planning and Community Right-to-Know Act of 1986, enacted as Title III of SARA, which created state commissions, Local Emergency Planning Committees, off-site plans, release notification, chemical inventory reporting and the Toxics Release Inventory, followed by Clean Air Act section 112(r), the Risk Management Program and OSHA’s Process Safety Management standard.
• That framework only functions where the committee actually meets and the inventory data reaches the first-arriving officer, so verify current reporting requirements with EPA and your state emergency response commission and check when your own LEPC last reviewed its plan.