London
False Alarm
Index.
Every automatic fire alarm false alarm London Fire Brigade attended from 2019 to 31 July 2026, by borough, by type of building and by time of day. Compiled by Gemini AMPM from the Brigade’s own open incident records.
2025, all of London
47,038
automatic fire alarm false alarms attended by London Fire Brigade
- 128.9 a day, on average
- 11 min between attendances, on average
- 34.1% of everything the Brigade attended
What is counted. An AFA false alarm here is an incident London Fire Brigade attended and recorded as a false alarm from an automatic fire alarm. Alarms that sounded and were not attended are not in the data. How the Index is built.
A fall in attendance,
not in alarms.
On 29 October 2024 London Fire Brigade changed how it responds to automatic fire alarms. Between 07:00 and 20:30 it no longer attends one at most non-residential buildings unless a person also calls 999 to report a fire. Homes, hospitals, schools, care homes, hotels and the other premises on the Brigade’s exemption list still get a response at any hour.
In the year before the change the Brigade attended 13,713 AFA false alarms at non-residential buildings in those hours. In the year after it attended 6,396. That is a change of -53.4%.
Non-residential AFA false alarms attended, by half hour of the day
The year before the change against the year after it. The shaded band is 07:00 to 20:30, the hours in which the Brigade no longer attends most non-residential automatic fire alarms.
- Year before: 29 October 2023 to 28 October 2024
- Year after: 29 October 2024 to 28 October 2025
- LFB policy hours
Show the numbers as a table
| Half hour starting | Year before | Year after |
|---|---|---|
| 00:00 | 255 | 237 |
| 00:30 | 236 | 218 |
| 01:00 | 214 | 230 |
| 01:30 | 205 | 180 |
| 02:00 | 226 | 200 |
| 02:30 | 193 | 171 |
| 03:00 | 185 | 182 |
| 03:30 | 207 | 170 |
| 04:00 | 175 | 181 |
| 04:30 | 198 | 176 |
| 05:00 | 207 | 182 |
| 05:30 | 245 | 210 |
| 06:00 | 305 | 299 |
| 06:30 | 303 | 301 |
| 07:00 | 394 | 185 |
| 07:30 | 458 | 179 |
| 08:00 | 535 | 217 |
| 08:30 | 608 | 280 |
| 09:00 | 687 | 290 |
| 09:30 | 653 | 313 |
| 10:00 | 753 | 333 |
| 10:30 | 642 | 329 |
| 11:00 | 730 | 319 |
| 11:30 | 644 | 275 |
| 12:00 | 574 | 295 |
| 12:30 | 575 | 256 |
| 13:00 | 580 | 245 |
| 13:30 | 531 | 237 |
| 14:00 | 559 | 286 |
| 14:30 | 538 | 280 |
| 15:00 | 483 | 243 |
| 15:30 | 472 | 184 |
| 16:00 | 419 | 192 |
| 16:30 | 357 | 176 |
| 17:00 | 373 | 156 |
| 17:30 | 360 | 161 |
| 18:00 | 398 | 174 |
| 18:30 | 346 | 143 |
| 19:00 | 404 | 155 |
| 19:30 | 334 | 160 |
| 20:00 | 306 | 333 |
| 20:30 | 337 | 284 |
| 21:00 | 313 | 314 |
| 21:30 | 327 | 289 |
| 22:00 | 301 | 292 |
| 22:30 | 291 | 243 |
| 23:00 | 294 | 273 |
| 23:30 | 287 | 242 |
Before and after, in full
| Group | Year before | Year after | % |
|---|---|---|---|
| Non-residential, 07:00 to 20:30 | 13,713 | 6,396 | -53.4% |
| Property types not named as exempt | 10,748 | 3,670 | -65.9% |
| Hospitals, schools, nurseries, prisons | 2,965 | 2,726 | -8.1% |
| Non-residential, outside those hours | 5,304 | 4,874 | -8.1% |
| Residential, all hours | 33,729 | 35,232 | +4.5% |
| All AFA false alarms | 52,793 | 46,536 | -11.9% |
Within that non-residential 07:00 to 20:30 group, attendances at hospitals, schools, nurseries, prisons and young offenders' units (property types that LFB lists as exempt from the change) went from 2,965 to 2,726 (-8.1%). Attendances at all other non-residential property types went from 10,748 to 3,670 (-65.9%). Listed buildings and premises the Brigade has exempted one by one cannot be picked out in the data, so the “not named as exempt” group still contains some exempt buildings.
Where attendance changed most
| Property type (as recorded by LFB) | Before | After | % |
|---|---|---|---|
| Purpose built office | 2,287 | 676 | -70.4% |
| Single shop | 1,519 | 438 | -71.2% |
| Infant/Primary schoolExempt | 1,395 | 1,342 | -3.8% |
| Secondary schoolExempt | 812 | 724 | -10.8% |
| HospitalExempt | 482 | 446 | -7.5% |
| Warehouse | 445 | 36 | -91.9% |
| Large supermarket | 439 | 61 | -86.1% |
| Restaurant/cafe | 404 | 173 | -57.2% |
| Converted office | 399 | 159 | -60.2% |
| College/University | 391 | 350 | -10.5% |
One thing the chart shows that the policy does not say. The published policy hours end at 20:30. In the records, attendance is back at its earlier level from 20:00, 30 minutes sooner. In the year after the change the Brigade attended 160 non-residential AFA false alarms in the half hour from 19:30 and 333 in the half hour from 20:00. The year before, the figures were 334 and 306. The start of the window matches the published 07:00. The data does not say why the end differs, and the published hours are the ones to plan around.
Why this matters to a building manager. The records cannot show how many alarms sounded and went unattended. What they show is that, in a non-exempt building in the policy hours, a crew is no longer sent on the alarm signal alone. Finding out why the alarm went off, and calling 999 if there is a fire, is now the building’s job. What that means in practice.
The same attendances, month by month
Non-residential AFA false alarms attended between 07:00 and 20:30, Jan 2023 to Jul 2026. The fall arrived with the policy and has held since.
- Months before the change
- Months after the change
Show the numbers as a table
| Month | Attended | Period |
|---|---|---|
| Jan 2023 | 1,168 | Before |
| Feb 2023 | 1,018 | Before |
| Mar 2023 | 1,154 | Before |
| Apr 2023 | 944 | Before |
| May 2023 | 1,073 | Before |
| Jun 2023 | 1,085 | Before |
| Jul 2023 | 1,166 | Before |
| Aug 2023 | 1,155 | Before |
| Sep 2023 | 1,181 | Before |
| Oct 2023 | 1,250 | Before |
| Nov 2023 | 1,337 | Before |
| Dec 2023 | 1,059 | Before |
| Jan 2024 | 1,157 | Before |
| Feb 2024 | 1,206 | Before |
| Mar 2024 | 1,134 | Before |
| Apr 2024 | 1,083 | Before |
| May 2024 | 1,119 | Before |
| Jun 2024 | 1,037 | Before |
| Jul 2024 | 1,115 | Before |
| Aug 2024 | 1,052 | Before |
| Sep 2024 | 1,148 | Before |
| Oct 2024 | 1,225 | Before (includes the first 3 days after) |
| Nov 2024 | 596 | After |
| Dec 2024 | 524 | After |
| Jan 2025 | 569 | After |
| Feb 2025 | 450 | After |
| Mar 2025 | 469 | After |
| Apr 2025 | 454 | After |
| May 2025 | 512 | After |
| Jun 2025 | 515 | After |
| Jul 2025 | 561 | After |
| Aug 2025 | 565 | After |
| Sep 2025 | 594 | After |
| Oct 2025 | 595 | After |
| Nov 2025 | 564 | After |
| Dec 2025 | 514 | After |
| Jan 2026 | 593 | After |
| Feb 2026 | 537 | After |
| Mar 2026 | 538 | After |
| Apr 2026 | 521 | After |
| May 2026 | 587 | After |
| Jun 2026 | 591 | After |
| Jul 2026 | 563 | After |
Seven full years,
and 2026 so far.
AFA false alarms attended in London numbered 41,477 in 2019 and 47,038 in 2025. The highest year from 2019 to 2025 was 2024 with 51,625. The 2025 total was 4,587 lower than 2024 (-8.9%).
Non-residential attendances went from 17,762 in 2024 to 11,211 in 2025. Residential attendances went from 33,815 to 35,796.
2026 is a part year. From 1 January to 31 July 2026 the Brigade attended 28,591 AFA false alarms, against 26,001 in the same dates of 2025 (+10.0%). It should not be set against a full year.
AFA false alarms attended each year, 2019 to 2025
London, by calendar year of the 999 call. 2026 covers 1 January to 31 July 2026 only and is drawn paler.
- Residential (dwellings and other residential)
- Non-residential
- Part year
Show the numbers as a table
| Period | AFA false alarms | Residential | Non-residential | Per day | % of all incidents | Change |
|---|---|---|---|---|---|---|
| 2019 | 41,477 | 23,556 | 17,899 | 113.6 | 39.5% | +4.0% |
| 2020 | 37,908 | 23,813 | 14,076 | 103.6 | 38.5% | -8.6% |
| 2021 | 41,762 | 25,760 | 15,965 | 114.4 | 38.1% | +10.2% |
| 2022 | 46,564 | 28,594 | 17,946 | 127.6 | 37.1% | +11.5% |
| 2023 | 49,810 | 31,225 | 18,542 | 136.8 | 39.5% | +7.0% |
| 2024 | 51,625 | 33,815 | 17,762 | 141.1 | 38.5% | +3.6% |
| 2025 | 47,038 | 35,796 | 11,211 | 128.9 | 34.1% | -8.9% |
| 2026, 1 January to 31 July 2026 | 28,591 | 21,739 | 6,834 | 134.9 | 33.2% | +10.0% |
Most are at homes,
not workplaces.
By property category, 30,932 of the 2025 AFA false alarms (65.8%) were at dwellings, 11,211 (23.8%) at non-residential properties and 4,864 (10.3%) at other residential properties such as care homes, hostels, hotels and student halls. Dwellings and other residential property together account for 76.1% of the 2025 total.
Residential buildings are outside the daytime policy, so the Brigade still attends them at every hour. Between the year before the change and the year after, residential AFA false alarms went from 33,729 to 35,232 (+4.5%). The data does not say why.
The twelve property types with the most AFA false alarms, 2025
Property type as recorded by London Fire Brigade crews. The percentage is the share of all AFA false alarms in London.
- Residential
- Non-residential
Show the numbers as a table
| Property type | Category | AFA false alarms | % of London total |
|---|---|---|---|
| Purpose Built Flats/Maisonettes - 4 to 9 storeys | Dwelling | 8,426 | 17.9% |
| Purpose Built Flats/Maisonettes - Up to 3 storeys | Dwelling | 6,043 | 12.8% |
| Self contained Sheltered Housing | Dwelling | 4,498 | 9.6% |
| House - single occupancy | Dwelling | 4,446 | 9.5% |
| Converted Flat/Maisonettes - 3 or more storeys | Dwelling | 2,468 | 5.2% |
| Purpose Built Flats/Maisonettes - 10 or more storeys | Dwelling | 2,315 | 4.9% |
| Retirement/Old Persons Home | Other Residential | 2,078 | 4.4% |
| Infant/Primary school | Non Residential | 1,619 | 3.4% |
| Converted Flat/Maisonette - Up to 2 storeys | Dwelling | 1,387 | 2.9% |
| Purpose built office | Non Residential | 1,260 | 2.7% |
| Single shop | Non Residential | 1,071 | 2.3% |
| Secondary school | Non Residential | 943 | 2.0% |
| Nursing/Care Home/Hospice | Other Residential | 654 | 1.4% |
| Hospital | Non Residential | 651 | 1.4% |
| Hotel/motel | Other Residential | 628 | 1.3% |
| Licensed House in Multiple Occupation - Up to 2 storeys | Dwelling | 584 | 1.2% |
| Student Hall of Residence | Other Residential | 532 | 1.1% |
| Hostel (e.g. for homeless people) | Other Residential | 508 | 1.1% |
| College/University | Non Residential | 474 | 1.0% |
| Restaurant/cafe | Non Residential | 381 | 0.8% |
The property type with the most AFA false alarms in 2025 was "Purpose Built Flats/Maisonettes - 4 to 9 storeys" (8,426), followed by "Purpose Built Flats/Maisonettes - Up to 3 storeys" (6,043) and "Self contained Sheltered Housing" (4,498). Among non-residential property types the largest were "Infant/Primary school" (1,619), "Purpose built office" (1,260) and "Single shop" (1,071). A count by property type says where crews were sent. It does not say what set the alarm off or whether the system was at fault.
All 33 areas,
ranked.
Westminster had the most AFA false alarms of London's 33 local authority areas in 2025: 4,248, or 11.6 a day and 9.0% of the London total. Camden (2,655) and Kensington and Chelsea (2,369) were next. Kingston upon Thames had the fewest, 560.
Non-residential AFA false alarms attended between 07:00 and 20:30 were lower in the year after the change than in the year before in 33 of the 33 areas (32 boroughs and the City of London). The largest fall in number was in Westminster, from 1,978 to 915 (-53.7%). The largest percentage fall was in Kingston upon Thames, from 223 to 66 (-70.4%), and the smallest percentage change was in Greenwich, from 332 to 242 (-27.1%).
The counts are not adjusted for population or for the number of buildings and alarm systems in each area. A borough with more buildings will tend to have more incidents. Select a borough for its own page.
AFA false alarms attended in 2025, by borough
One square for each of the 32 boroughs and the City of London, in rough geographic position. Deeper orange means more attendances. Each square links to that borough’s page.
- WST , Westminster: 4,248 AFA false alarms in 2025, rank 1 of 33
- CMD , Camden: 2,655 AFA false alarms in 2025, rank 2 of 33
- KNS , Kensington and Chelsea: 2,369 AFA false alarms in 2025, rank 3 of 33
- LAM , Lambeth: 2,146 AFA false alarms in 2025, rank 4 of 33
- SWR , Southwark: 2,005 AFA false alarms in 2025, rank 5 of 33
- TWH , Tower Hamlets: 1,807 AFA false alarms in 2025, rank 6 of 33
- BRN , Barnet: 1,750 AFA false alarms in 2025, rank 7 of 33
- HMS , Hammersmith and Fulham: 1,718 AFA false alarms in 2025, rank 8 of 33
- CRD , Croydon: 1,692 AFA false alarms in 2025, rank 9 of 33
- HCK , Hackney: 1,636 AFA false alarms in 2025, rank 10 of 33
- LSH , Lewisham: 1,599 AFA false alarms in 2025, rank 11 of 33
- WNS , Wandsworth: 1,593 AFA false alarms in 2025, rank 12 of 33
- ISL , Islington: 1,569 AFA false alarms in 2025, rank 13 of 33
- EAL , Ealing: 1,466 AFA false alarms in 2025, rank 14 of 33
- HDN , Hillingdon: 1,394 AFA false alarms in 2025, rank 15 of 33
- BRT , Brent: 1,355 AFA false alarms in 2025, rank 16 of 33
- WTH , Waltham Forest: 1,323 AFA false alarms in 2025, rank 17 of 33
- GRN , Greenwich: 1,270 AFA false alarms in 2025, rank 18 of 33
- NWM , Newham: 1,267 AFA false alarms in 2025, rank 19 of 33
- BRM , Bromley: 1,241 AFA false alarms in 2025, rank 20 of 33
- STN , Sutton: 1,231 AFA false alarms in 2025, rank 21 of 33
- HGY , Haringey: 1,191 AFA false alarms in 2025, rank 22 of 33
- HNS , Hounslow: 1,146 AFA false alarms in 2025, rank 23 of 33
- ENF , Enfield: 854 AFA false alarms in 2025, rank 24 of 33
- RDB , Redbridge: 851 AFA false alarms in 2025, rank 25 of 33
- HVG , Havering: 841 AFA false alarms in 2025, rank 26 of 33
- HRW , Harrow: 833 AFA false alarms in 2025, rank 27 of 33
- MRT , Merton: 767 AFA false alarms in 2025, rank 28 of 33
- BXL , Bexley: 716 AFA false alarms in 2025, rank 29 of 33
- BAR , Barking and Dagenham: 687 AFA false alarms in 2025, rank 30 of 33
- RCH , Richmond upon Thames: 684 AFA false alarms in 2025, rank 31 of 33
- CTY , City of London: 574 AFA false alarms in 2025, rank 32 of 33
- KNG , Kingston upon Thames: 560 AFA false alarms in 2025, rank 33 of 33
Key: AFA false alarms in 2025
- Under 1,000 (10 areas)
- 1,000 to 1,499 (10 areas)
- 1,500 to 1,999 (8 areas)
- 2,000 to 2,999 (4 areas)
- 3,000 or more (1 area)
Most: Westminster, 4,248. Fewest: Kingston upon Thames, 560. The three-letter codes are listed beside each name in the table below, which holds the same figures.
Borough league table
| Rank | Borough | AFA false alarms | Daytime non-residential |
|---|---|---|---|
| 1 | Westminster WST | 4,248 | -53.7% |
| 2 | Camden CMD | 2,655 | -39.8% |
| 3 | Kensington and Chelsea KNS | 2,369 | -52.5% |
| 4 | Lambeth LAM | 2,146 | -63.6% |
| 5 | Southwark SWR | 2,005 | -43.3% |
| 6 | Tower Hamlets TWH | 1,807 | -56.9% |
| 7 | Barnet BRN | 1,750 | -38.4% |
| 8 | Hammersmith and Fulham HMS | 1,718 | -51.9% |
| 9 | Croydon CRD | 1,692 | -49.9% |
| 10 | Hackney HCK | 1,636 | -48.8% |
| 11 | Lewisham LSH | 1,599 | -58.9% |
| 12 | Wandsworth WNS | 1,593 | -62.0% |
| 13 | Islington ISL | 1,569 | -57.8% |
| 14 | Ealing EAL | 1,466 | -61.4% |
| 15 | Hillingdon HDN | 1,394 | -58.8% |
| 16 | Brent BRT | 1,355 | -69.9% |
| 17 | Waltham Forest WTH | 1,323 | -46.3% |
| 18 | Greenwich GRN | 1,270 | -27.1% |
| 19 | Newham NWM | 1,267 | -58.0% |
| 20 | Bromley BRM | 1,241 | -47.1% |
| 21 | Sutton STN | 1,231 | -50.0% |
| 22 | Haringey HGY | 1,191 | -44.9% |
| 23 | Hounslow HNS | 1,146 | -61.5% |
| 24 | Enfield ENF | 854 | -63.6% |
| 25 | Redbridge RDB | 851 | -46.9% |
| 26 | Havering HVG | 841 | -41.4% |
| 27 | Harrow HRW | 833 | -52.6% |
| 28 | Merton MRT | 767 | -55.4% |
| 29 | Bexley BXL | 716 | -55.6% |
| 30 | Barking and Dagenham BAR | 687 | -68.8% |
| 31 | Richmond upon Thames RCH | 684 | -54.4% |
| 32 | City of London CTY | 574 | -60.4% |
| 33 | Kingston upon Thames KNG | 560 | -70.4% |
| London | 47,038 | -53.4% |
Daytime non-residential attendance was lower in the year after the change in 33 of the 33 areas. Small boroughs have small counts, so their percentages move more. Full figures for every year are in the CSV.
Busiest from 20:00.
Quietest from 04:00.
In 2025 the hour with the most AFA false alarms was 20:00 to 20:59 (2,670) and the hour with the fewest was 04:00 to 04:59 (1,028). Friday had the highest daily average (134.3) and Sunday the lowest (121.3).
The figures for non-residential buildings show why the evening stands out. In 2025 the Brigade attended 309 non-residential AFA false alarms in the hour from 19:00 and 613 in the hour from 20:00, when daytime attendance resumes. The half-hour chart shows the step.
AFA false alarms attended by day and hour, 2025
All property types, all of London. Each cell is the number attended in that hour on that day of the week across the year.
Show the numbers as a table
| Hour | Mon | Tue | Wed | Thu | Fri | Sat | Sun | All days |
|---|---|---|---|---|---|---|---|---|
| 00:00 | 217 | 205 | 220 | 215 | 213 | 211 | 222 | 1,503 |
| 01:00 | 171 | 186 | 180 | 157 | 208 | 238 | 196 | 1,336 |
| 02:00 | 142 | 162 | 159 | 151 | 185 | 181 | 189 | 1,169 |
| 03:00 | 152 | 157 | 135 | 153 | 149 | 180 | 137 | 1,063 |
| 04:00 | 145 | 129 | 146 | 136 | 144 | 171 | 157 | 1,028 |
| 05:00 | 142 | 139 | 139 | 139 | 157 | 159 | 154 | 1,029 |
| 06:00 | 213 | 178 | 179 | 191 | 218 | 167 | 141 | 1,287 |
| 07:00 | 192 | 189 | 187 | 184 | 210 | 159 | 129 | 1,250 |
| 08:00 | 257 | 259 | 252 | 288 | 266 | 213 | 167 | 1,702 |
| 09:00 | 322 | 355 | 366 | 345 | 333 | 280 | 235 | 2,236 |
| 10:00 | 351 | 392 | 408 | 428 | 374 | 300 | 286 | 2,539 |
| 11:00 | 349 | 407 | 438 | 392 | 414 | 306 | 295 | 2,601 |
| 12:00 | 360 | 401 | 425 | 380 | 358 | 330 | 329 | 2,583 |
| 13:00 | 324 | 373 | 383 | 360 | 383 | 334 | 356 | 2,513 |
| 14:00 | 345 | 347 | 410 | 378 | 324 | 330 | 344 | 2,478 |
| 15:00 | 315 | 323 | 346 | 310 | 312 | 330 | 332 | 2,268 |
| 16:00 | 324 | 290 | 335 | 310 | 330 | 321 | 366 | 2,276 |
| 17:00 | 334 | 325 | 337 | 337 | 324 | 337 | 323 | 2,317 |
| 18:00 | 328 | 351 | 387 | 346 | 365 | 334 | 358 | 2,469 |
| 19:00 | 320 | 380 | 341 | 342 | 349 | 320 | 340 | 2,392 |
| 20:00 | 365 | 367 | 406 | 370 | 398 | 361 | 403 | 2,670 |
| 21:00 | 340 | 337 | 316 | 360 | 374 | 358 | 325 | 2,410 |
| 22:00 | 273 | 298 | 296 | 307 | 318 | 303 | 285 | 2,080 |
| 23:00 | 253 | 277 | 257 | 266 | 280 | 265 | 241 | 1,839 |
| All hours | 6,534 | 6,827 | 7,048 | 6,845 | 6,986 | 6,488 | 6,310 | 47,038 |
If you manage
a building.
The Index describes what the Brigade attended. What you do about your own alarms is a matter for your fire risk assessment and the people who maintain the system. The points below are general, and follow London Fire Brigade’s own advice in Guidance Note 54: Managing and reducing false alarms and its guide to fire alarm systems.
Offices, shops, warehouses and other non-exempt buildings
Between 07:00 and 20:30, the first response is yours
An alarm signal alone no longer brings a crew in these hours. A 999 call that confirms a fire does, at any time.
- Have a safe way to investigate. Name and train the people who check the panel and the area it points to, and only where it is safe. The Brigade’s guidance note lists what that training should cover.
- Call 999 on any sign of fire. Smoke, flames, heat or the smell of burning count as confirmation. Do not wait to be certain, and do not rely on the alarm signal to make the call for you.
- Bring the fire risk assessment and emergency plan up to date. The Brigade asks non-exempt premises to review them. Any investigation period or staff alarm is a change to your fire procedures and has to be justified there, with advice from a competent assessor. We can introduce independent fire risk assessors.
- Agree the daytime routine with your alarm receiving centre. If the system is monitored, settle who the centre calls first and what it tells the Brigade. Our fire alarm monitoring page explains how a verification call works.
- Remove the causes. Match the detector to the space: a multi-sensor or heat detector can often stop alarms from steam, cooking or dust, assessed case by case. Fit covers to manual call points that get knocked. Control contractors: isolate or cover detectors during dusty or hot works, put interim measures in place, and remove the covers afterwards.
- Maintain it, and log every false alarm. Servicing to BS 5839-1 keeps the system working (the AMPM Group Compliance Calendar sets out how often a fire alarm should be tested and serviced), and a logbook entry for each activation, with its cause, is how patterns get found. Our fire alarm maintenance visits review the activation history each time.
- If you think you are exempt, do not assume it. Schools, listed buildings and others on the list still get a response, but the Brigade asks that whoever calls 999 knows the building is exempt and says so.
Blocks of flats, sheltered housing, student halls and other residential buildings
The Brigade still comes, so the aim is fewer alarms
Residential buildings get a response at every hour, and they are where most AFA false alarms are attended: 76.1% of the 2025 total.
- Know what you have. List the systems in the building that can raise a fire signal: detection in common parts, smoke ventilation, warden call or telecare links, any communal alarm. Record which of them call out, and to whom.
- Log every activation and look for repeats. The same detector, the same floor or the same time of day usually points to a fixable cause. Keep the record with the fire logbook.
- Get competent advice before changing detection. Moving a detector or changing its type can cut alarms from cooking and steam, but only where it suits the building’s fire strategy and risk assessment. Never cure a nuisance alarm by disabling the device.
- Control works in the building. Dust and hot works set detectors off. Agree isolations and temporary covers with contractors, with interim measures while they are in place, and make sure they come off.
- Tell the receiving centre before you test. Tests and servicing that are not announced are a common cause of a call to the Brigade.
- Keep the system maintained. Faults and dirty detectors cause false alarms. For housing providers and managing agents, our residential and social housing page sets out how we work in occupied blocks, and fire alarm remedial works covers device and panel changes.
Gemini AMPM designs, installs, maintains and monitors fire alarm systems across London and the South East. If false alarms are a problem in your building, send us the panel’s event log and we will start there.
Questions the
data can answer.
How many false alarms does London Fire Brigade attend?
London Fire Brigade attended 47,038 automatic fire alarm (AFA) false alarms in 2025, an average of 128.9 a day. That was 34.1% of the 137,981 incidents it attended and 76.2% of its 61,751 false alarms. So far in 2026 (1 January to 31 July 2026) it has attended 28,591, against 26,001 in the same dates of 2025.
Does London Fire Brigade still attend automatic fire alarms?
Yes, but not everywhere at every hour. Since 29 October 2024, between 07:00 and 20:30, the Brigade does not attend an automatic fire alarm at most non-residential buildings unless a person also calls 999 to report a fire. It still attends automatic fire alarms at any building between 20:30 and 07:00, and at homes and other exempt premises at all times. The full policy is on London Fire Brigade’s AFA policy page.
Which buildings are exempt from the London Fire Brigade AFA policy?
The Brigade’s policy page lists these as exempt, which means crews attend an automatic fire alarm there without confirmation of a fire: private dwellings and houses; flats, including high-rise; houses in Multiple Occupation; mobile and park homes; houseboats; hospitals; residential care homes, nursing homes and hospices; children's homes; specialised housing premises (for example sheltered housing, extra care sheltered housing, supported living); student accommodation and halls of residence; residential boarding schools; hotels, motels, B&Bs and other guest accommodation; hostels (for example homeless, rehabilitation) and youth hostels; prisons, young offenders' institutions and other secure establishments; Grade 1, Grade 2 and Grade 2S listed heritage buildings; schools; nurseries; buildings of substantial public significance identified as exempt by London Fire Brigade. The Brigade asks exempt premises to make sure whoever calls 999 knows the building is exempt and says so, and not to assume an exemption. Check the current list before you rely on it.
Has the policy cut the number of false alarms?
It has cut the number the Brigade attends. After LFB changed its response on 29 October 2024, AFA false alarms attended at non-residential properties between 07:00 and 20:30 fell from 13,713 in the year before (29 October 2023 to 28 October 2024) to 6,396 in the year after (29 October 2024 to 28 October 2025), a change of -53.4%. Outside those hours non-residential attendances went from 5,304 to 4,874 (-8.1%). Residential attendances went from 33,729 to 35,232 (+4.5%), and all AFA false alarms from 52,793 to 46,536 (-11.9%). The records count attendances, so they cannot show how many alarms sounded and were not attended.
Which London borough has the most fire alarm false alarms?
Westminster had the most AFA false alarms of London's 33 local authority areas in 2025: 4,248, or 11.6 a day and 9.0% of the London total. Camden (2,655) and Kensington and Chelsea (2,369) were next. Kingston upon Thames had the fewest, 560. The counts are not adjusted for population or for the number of buildings, so a borough with more buildings will tend to have more. The league table ranks all 33 areas.
What should I do if the fire alarm sounds in a commercial building during the day?
Follow your emergency plan. For a building that is not exempt, the Brigade’s guidance is to have a safe way for trained staff to investigate the alarm, and to call 999 as soon as there is any sign of fire, such as smoke, flames, heat or the smell of burning. Do not wait for a crew to arrive first: between 07:00 and 20:30 one will not be sent on the alarm signal alone. LFB Guidance Note 54 covers safe investigation and the policy FAQs cover what the control room will ask.
What is the “notional cost” of a false alarm in the data?
The notional cost recorded against AFA false alarms in 2025 was £22,472,940, and the pump time recorded was 2,941,997 minutes (49,033.3 hours). 43,118 of the 47,038 incidents carry the 60-minute minimum that LFB applies to any incident of under an hour, so the recorded time is not time on scene. LFB's metadata defines notional cost as time spent multiplied by the notional annual cost of a pump. It is a notional figure and not a bill to any building owner. The rate behind it rises each April, so totals in different years are not like for like. More detail is in the methodology.
Take the numbers.
Check the working.
Download
One row for each area and period: 33 areas plus a London total, calendar years 2019 to 2025 and 2026 to 31 July. Column definitions are in the README.
Methodology
Definitions, filters, the checks run on every figure and the reconciliation with totals published by London Fire Brigade and by government are on the methodology page.
Updates
London Fire Brigade publishes its incident records on the London Datastore. This Index is refreshed when new data is published. This edition was generated on 5 October 2026 from files dated 25 August 2026, and covers incidents to 31 July 2026.
Source and licence
Source: London Fire Brigade, London Fire Brigade Incident Records, published on the London Datastore (data.london.gov.uk). Contains public sector information licensed under the Open Government Licence v2.0. The Greater London Authority cannot warrant the quality or accuracy of the data. This analysis is by Gemini AMPM. It is not endorsed by, affiliated with or approved by London Fire Brigade or the Greater London Authority.
How to cite
Gemini AMPM (2026). London False Alarm Index: automatic fire alarm false alarms attended by London Fire Brigade, 2019 to 31 July 2026. https://geminiampm.co.uk/false-alarm-index/ (data: London Fire Brigade Incident Records, London Datastore, Open Government Licence v2.0).
Journalists and researchers: you may quote the figures and reuse the charts with that credit. For anything not on the page, email [email protected].
Related insights.
Fire alarm maintenance contracts: what BS 5839-1 requires and what a good contract includes
What BS 5839-1:2025 requires of a fire alarm maintenance contract, what a proper service visit covers, and the red flags in a cheap quote.
Fire alarm monitoring options compared: ARC signalling, keyholder response and app-based alerts
The Redcare era is over. Keyholder response, app alerts and ARC monitoring compared: who watches, who calls the fire service, what insurers accept.
Nine fire alarm faults, no risk assessment, £33,700 to pay
A Somerset guest house was ordered to pay £33,700 after inspectors found nine fire alarm faults and no risk assessment. What to check in your building.
