Disaster electricity statistics show how weather and other major events affect customers, reliability, finances, and health. In the United States, 2024 interruption data, NOAA disaster records, DOE cost modeling, and a peer-reviewed blackout study measure different parts of that impact.
Key Disaster Electricity Statistics
The most useful headline figures span national reliability, major events, costs, and outcomes:
- In 2024, U.S. electricity customers averaged 11 hours of interruption across the United States.
- In 2024, major events accounted for 80% of U.S. interruption hours.
- During 2014–2023, major-event interruptions averaged nearly 4 hours per customer per year in the United States.
- In 2024, the national average interruption frequency was 1.5 instances per customer.
- In July 2024, Hurricane Beryl left 2.6 million Texas customers without power.
- In September 2024, Hurricane Helene left 5.9 million customers without power across 10 states.
- From 1980–2024, the United States recorded 403 billion-dollar weather and climate disasters.
- From 1980–2024, those disasters produced more than $2.915 trillion in total damage and costs in 2024 dollars.
- In 2024, 27 U.S. billion-dollar disasters caused $182.7 billion in CPI-adjusted losses.
- In 2024, South Carolina customers experienced nearly 53 hours of interruption.
- In 2024, Hawaii customers experienced 4.4 interruptions per customer.
- In 2024, Michigan’s all-event SAIDI was 470.4 minutes per customer.
- In January 2024, a Detroit Edison weather event affected 172,481 Michigan customers for 112 hours 23 minutes.
- In January 2024, a Portland General Electric event affected 165,000 Oregon customers for 55 hours 10 minutes and recorded 4,031 MW lost.
- DOE/LBNL’s base-case model estimated annual U.S. power-interruption costs at $79 billion.
- Commercial customers represented more than 70% of the $79 billion base-case interruption cost, or $57 billion.
- The August 2003 New York blackout was associated with a 122% increase in accidental mortality risk.
- The same blackout study estimated approximately 90 excess deaths in New York City.
Contents
- Key Disaster Electricity Statistics
- Disaster Electricity Outage Scale and Reliability Trends
- Weather Disasters and the Cost of Electricity Disruption
- State Electricity Interruption Data in 2024
- Major Disaster Electricity Events and Restoration Durations
- Economic Cost of Disaster-Related Electricity Interruptions
- Health and Critical-Service Outcomes During Power Outages
Disaster Electricity Outage Scale and Reliability Trends
The U.S. Energy Information Administration (EIA) reported an average of 11 interruption hours per customer in 2024. Major-event interruptions averaged nearly 9 hours per customer, compared with a routine non-major-event average of about 2 hours per year in the EIA comparison.
The 2024 major-event share was unusually concentrated. Hurricanes Beryl, Helene, and Milton together accounted for 80% of interruption hours, while major events averaged nearly 4 hours per customer annually during 2014–2023.
| EIA measure | Period | Result |
|---|---|---|
| Average interruption duration | 2024 | 11 hours per customer |
| Major-event share of interruption hours | 2024 | 80% |
| Major-event average | 2024 | Nearly 9 hours per customer |
| Major-event average | 2014–2023 | Nearly 4 hours per customer per year |
| Average interruption frequency | 2024 | 1.5 instances per customer |
Source: EIA, “Hurricanes in 2024 led to the most hours without power in the United States in 10 years”. EIA’s SAIDI-style average includes interruptions reported by utilities, and major-event classifications follow utility reporting conventions.
The event counts also show why customer totals and interruption hours should not be treated as interchangeable. Beryl left 2.6 million Texas customers without power in July 2024, while Helene affected 5.9 million customers across 10 states in September; event totals can include customers affected more than once during restoration.
Weather Disasters and the Cost of Electricity Disruption
NOAA’s billion-dollar disaster series provides an all-hazard context rather than an electricity-only total. It counted 403 U.S. billion-dollar weather and climate disasters from 1980 through 2024, producing more than $2.915 trillion in total damage and costs in 2024 dollars.
In 2024, NOAA counted 27 billion-dollar disasters, with $182.1 billion in nominal losses or $182.7 billion after CPI adjustment. The events caused 568 deaths, but those are event-attributed deaths, not deaths attributed specifically to power outages.
| NOAA hazard category | Billion-dollar events, 1980–2024 |
|---|---|
| Severe storms | 203 |
| Tropical cyclones | 67 |
| Flooding | 45 |
| Drought | 32 |
| Winter storms | 24 |
| Wildfires | 23 |
Source: NOAA NCEI, “U.S. Billion-Dollar Weather and Climate Disasters 1980–2024”. The totals describe overall disaster damage and deaths, not electricity-specific damage.
Severe storms were the most frequent category in the 1980–2024 record, followed by tropical cyclones, flooding, drought, winter storms, and wildfires. These categories describe hazard events and do not establish how many outages each event caused.
State Electricity Interruption Data in 2024
State results vary sharply depending on whether the measure is duration, frequency, event inclusion, and utility reporting coverage. EIA reported nearly 53 hours of interruption for South Carolina customers in 2024, the highest state total cited in its narrative, while Arizona, South Dakota, North Dakota, and Massachusetts each averaged less than 2 hours.
| Geography | 2024 all-event SAIDI or frequency | Unit |
|---|---|---|
| South Carolina | Nearly 53 | hours per customer |
| New England | 419.6 | minutes per customer |
| Connecticut | 153.6 | minutes per customer |
| Michigan | 470.4 | minutes per customer |
| Ohio | 510.5 | minutes per customer |
| Wisconsin | 286.1 | minutes per customer |
| Hawaii | 4.4 | interruptions per customer |
Sources: EIA narrative and Electric Power Annual with final 2024 data. The state table uses EIA any-method metrics; reported coverage includes 96.0% for New England, 101.2% for Connecticut, 97.5% for Michigan, 96.6% for Ohio, and 92.3% for Wisconsin.
Hawaii’s 4.4 interruptions were almost three times the national 1.5-instance average, a simple comparison of rounded EIA figures rather than a causal estimate. EIA notes that Hawaii’s interruptions can involve weather, volcanic activity, plant outages, and interconnection issues.
Year-to-year changes can also be substantial. Michigan’s all-event SAIDI fell from 1,093.6 minutes in 2023 to 470.4 minutes in 2024; Indiana fell from 455.4 to 312.4 minutes, and Illinois fell from 208.2 to 156.9 minutes.
Major Disaster Electricity Events and Restoration Durations
EIA’s DOE-417 incident records provide event-level observations, including utility, geography, affected customers, duration, and sometimes megawatts lost. They are not interchangeable with annual SAIDI or SAIFI, and the table can contain overlapping records or blank restoration times.
| Event and geography | Period | Customers and duration |
|---|---|---|
| Puget Sound Energy, Washington | Jan. 9–10, 2024 | 75,180; 39h 56m |
| Detroit Edison, Michigan | Jan. 9–14, 2024 | 172,481; 112h 23m |
| National Grid, New York | Jan. 9–11, 2024 | 96,000; 46h 20m |
| Portland General Electric, Oregon | Jan. 13–15, 2024 | 165,000; 55h 10m; 4,031 MW lost |
| Portland General Electric, Oregon | Jan. 18–19, 2024 | 116,998; 27h 39m; 3,744 MW lost |
| Sacramento Municipal Utility District, California | Feb. 4–5, 2024 | 200,000; 13h; 1,230 MW lost |
| ISO New England | Mar. 23–26, 2024 | 110,000; 54h 20m |
Source: EIA Electric Power Monthly, Table B.1, “Major Disturbances and Unusual Occurrences”. MW is a reported loss measure, not MWh of energy lost.
Additional records show a separate Sacramento event affecting 167,000 customers for 1 hour 59 minutes on February 4, an ISO New England event affecting 50,000 customers for 9 hours 15 minutes on February 28–29, and another affecting 60,755 customers for 17 hours 5 minutes on March 10–11.
A March 22 CenterPoint Energy event affected 125,077 customers in Harris County, Texas, for 15 hours 37 minutes. A January 12 WEC Energy record listed 250,000 Wisconsin customers, while a January 12 ComEd record listed 186,778 Illinois customers; neither record had restoration duration reported at access.
Economic Cost of Disaster-Related Electricity Interruptions
DOE and Lawrence Berkeley National Laboratory modeled the annual cost of U.S. power interruptions at $79 billion in the base case. The estimate excludes power-quality costs and depends on assumptions about interruption frequency, duration, and customer costs; it is not limited to weather disasters.
The study’s sensitivity range was $22 billion to $135 billion annually. A narrower scenario produced $23 billion, while an intentionally unrealistic assumption placing all reliability events on summer weekday afternoons produced a modeled cost as high as $119 billion.
Commercial customers accounted for more than 70% of the base-case cost, or $57 billion. The model included 114.3 million residential customers, 14.9 million commercial customers, and 1.6 million industrial customers.
An unadjusted survey-cost sensitivity produced $378 billion, more than five times the base case. DOE describes that result as biased upward because the underlying surveys overrepresented large industrial customers.
Source: DOE/LBNL, “Understanding the Cost of Power Interruptions to U.S. Electricity Consumers”. These are modeled estimates, not a measured annual total of disaster-only losses.
Health and Critical-Service Outcomes During Power Outages
A peer-reviewed observational study of the August 14–15, 2003 New York City blackout estimated a 122% increase in accidental mortality risk, with a 95% confidence interval of 28%–287%. The same study estimated a 25% increase in non-accidental mortality risk, with a 95% confidence interval of 12%–41%.
The study estimated approximately 90 excess deaths in New York City. This is a counterfactual model estimate rather than a death-certificate count of outage-attributed deaths, and the association is not randomized causation.
Four of New York City’s 75 hospitals lost power for part of the blackout. That hospital count is study-specific and does not measure every critical facility or the outage duration at each hospital.
Source: “Lights out: Impact of the August 2003 power outage on mortality in New York, NY”. The study controlled for weather and seasonal and long-term trends.