M1 mapped corridor injury collisions rose 26% in 2025 — but the 13-year record tells a fuller story

Police-recorded injury collisions near the mapped M1 corridor rose from 104 in 2024 to 131 in 2025, a 26% increase, according to PureCar's analysis of Northern Ireland's historical collision files. It is a significant change in the number of reported incidents selected by the road map. It is not evidence, on its own, that driving on the motorway became 26% more dangerous.

The distinction matters because the M1 comparison is unusually easy to misread. PureCar identifies incidents occurring within 100 metres of 94 mapped M1 road sections. The dataset can include collisions at motorway junctions, slip roads and nearby roads, not exclusively vehicles travelling on the motorway's main carriageway. It counts police-reported incidents resulting in injury rather than every breakdown, near miss or damage-only crash.

The increase follows a relatively low year in 2024. Across the thirteen years from 2013 to 2025 the mapped area recorded 1,611 injury collisions, and the latest figure was below several earlier years, including the 153 recorded in 2017. The data establish a renewed rise after a dip, not an all-time high. That more precise description is the starting point for any serious examination of M1 safety.

The 2025 rise in its thirteen-year setting

The M1's mapped annual sequence begins at 114 injury collisions in 2013 and 130 in 2014. It then records 150 in 2015, 140 in 2016, 153 in 2017 and 141 in 2018. The 2019 figure was 120, followed by a marked fall to 87 in 2020. Counts recovered to 103 in 2021, 111 in 2022 and 127 in 2023 before dropping to 104 in 2024 and climbing to 131 in 2025.

A reader who saw only the 2024–25 comparison might assume the latest year was exceptional for the whole period. It was not. The count reached at least 130 in five earlier years: 2014 (130), 2015 (150), 2016 (140), 2017 (153) and 2018 (141). The 2017 peak exceeded the 2025 count by 22. A 26% rise from one year therefore brings the count back within the range observed repeatedly in the 2010s.

The 2023 count of 127 provides another useful comparison. The 131 incidents in 2025 were only four more than two years earlier, roughly 3.1% higher. That does not cancel the 27-incident rise from 2024; it changes its interpretation. Looking at a three-year sequence exposes both the intervening drop and the return towards the earlier level.

In 2020 the corridor recorded only 87 nearby injury collisions, its lowest annual total in the displayed period. Comparing 2025 directly with that year would generate an increase of 44, or roughly 50.6%. Such a claim could be numerically correct while being misleading about normal travel conditions. The pandemic altered mobility, and a researcher should not treat that disruption as an ordinary reference year.

PureCar's published three-year rolling figures also show why the short-term movement deserves care. The average ending in 2022 was approximately 100.3 collisions per year. By 2024 it was 114.0, and by 2025 approximately 120.7. This smoother measure has been rising from the early-2020s low, although it remained below the 147.7 three-year average recorded for the period ending in 2017. The long curve provides a more credible account than a slogan about an unprecedented motorway crisis.

Six-year totals show a different scale of change

For 2013–2018, the mapped M1 area recorded 828 injury collisions, averaging 138 a year. From 2019–2024, it recorded 652, an average of about 108.7. Across equal-length windows, the second period contained 176 fewer collisions, a decline of about 21.3%. That comparison is not a risk assessment, but it is an objective measure of how recorded incident frequency differed between two halves of the published history.

The latest year, 2025, sits above that more recent six-year average but below the earlier one. That makes the central finding more nuanced than “collisions up” or “collisions down”: the corridor saw lower counts overall in the later years, interrupted by a rise in 2023, a fall in 2024 and another rise in 2025. Both trends are visible without choosing a flattering start date.

Even the 2013–2018 and 2019–2024 comparison has limitations. The second window includes 2020, when travel was disrupted by Covid-19 restrictions and changes to commuting. It also does not account for shifts in traffic volumes, the distribution of journeys across the day, changes in adjacent infrastructure or any adjustments to how incidents are geocoded. A lower total in six calendar years is not proof of an equivalent reduction in exposure-adjusted collision risk.

The immediate question is which part of the 27 extra recorded incidents in 2025 occurred on motorway carriageways and which near junctions or adjoining routes. The answer determines what the headline means for engineers and road users. A cluster on a slip road could have a very different explanation and solution from a rise in collisions between motorway junctions.

How severe have the recorded incidents been?

Across the entire 2013–2025 corridor dataset, PureCar records ten fatal-severity injury collisions, 70 serious-severity collisions and 1,531 slight-severity collisions. The three classes total 1,611 incidents. Approximately 95% were classified slight, 4.3% serious and 0.6% fatal. These proportions describe the published mapped area over thirteen years, not the severity of the 131 incidents in 2025 specifically.

The corresponding casualty files contain 2,676 people killed or injured: ten killed, 88 seriously injured and 2,578 slightly injured. The linked records also identify 3,432 vehicles. Those are separate counts of events, people and vehicle involvement. A multi-vehicle collision is still one collision, and one event can produce several casualties.

This distinction is particularly important on a motorway network where headlines often describe several vehicles involved in one incident. An incident with three vehicles cannot be reported as three separate injury collisions simply because three vehicle records are attached. The 3,432 vehicles instead help establish the scale of recorded involvement across the incidents selected for the corridor.

Ten people killed in ten fatal-severity collisions in this dataset means the totals happen to be equal. That does not mean fatal collisions and fatal casualties are interchangeable measures as a rule. On other roads, one fatal event can claim multiple lives; on the A5 mapped dataset, for example, the two counts are different. The M1 equality is a feature of this dataset, not a definition.

The published corridor summary does not give a 2025-specific serious-injury or fatal-collision tally. Without analysing the annual subset at casualty level, it is not possible to state whether the rise from 104 to 131 came mainly from slight, serious or fatal incidents, or whether it involved more injured people. Saying the rise made the motorway “deadlier” would therefore exceed the available evidence.

A motorway is not one uniform environment

The M1 is a strategic transport route connecting Belfast with areas to the south and west. Its traffic environment includes mainline running, junction approaches and connections to local roads. Those different settings are why the boundary of a mapped corridor matters so much. A collision near a junction can fall inside the map's search area whether it happened on a motorway lane, on an exit slip, or on a closely adjoining road.

The relevant PureCar investigation uses a 100-metre proximity calculation around 94 selected mapped sections. The advantage is consistency: the same method can be applied across the historical PSNI series and compared between years. The limitation is that the road name attached to the study is not a verified police finding that each collision occurred on that named carriageway.

It would be irresponsible to infer from a map-density display that a particular junction is defective without validating coordinates against road layout, date, manoeuvre and traffic movement. Closely spaced roads can place several different collision types in the same apparent cluster. Equally, a mapping method that is useful for broad screening may be too coarse for identifying the exact site of a road engineering problem.

The corridor contains 94 mapped sections; the number is an important part of the study definition. It should not be compared directly with the A1's 76 sections or the A5's 60 as though all sections were the same length or carried the same traffic. A raw count is partly influenced by the geometry and amount of transport network included. A road league table assembled without matching exposure and coverage would not be an authoritative ranking of driver safety.

Speed enforcement figures cannot explain these collisions

National speeding statistics can sound persuasive when placed next to a motorway collision chart, but they answer different questions. The Department for Infrastructure's Road Safety Strategy 2025 statistical release reports that 26% of vehicles exceeded the speed limit on motorways under free-running conditions measured during the hours of 11pm to 7am. That is an observation about free-running vehicle speeds, not a statement that 26% of M1 crashes involved speed or that the M1's 26% rise resulted from speeding.

The numerical coincidence of two 26% figures is especially important to resist. One measures the annual rise in collision counts within a selected mapped corridor; the other describes the proportion of motorway vehicles exceeding limits under a defined monitoring method. They have different denominators, geographic scopes and periods of observation. Conflating them would manufacture a causal explanation.

PSNI's 2025 Key Statistics Report identifies inattention or diverted attention, driving too close, and emerging from minor roads without care among major principal causation factors nationally. It records 826 injury collisions with inattention or attention diverted as the principal factor and 548 with driving too close. These national findings may be relevant to the kinds of behaviours road safety programmes address, but they cannot be attributed to the M1 without a road-specific analysis of collision causation fields.

Nor can a general catalogue of motorway hazards—tailgating, lane changes, breakdowns, weather or congestion—serve as evidence of what happened in 2025. The collision dataset contains individual records capable of more detailed scrutiny, but no verified finding in the published aggregate has singled out one behaviour as responsible for the year-on-year change.

What a meaningful route-level assessment would compare

A useful comparison would first split the 131 recorded incidents in 2025 by mapped segment and by whether their coordinates were confirmed on the motorway itself, on slip roads or on adjoining streets. It would then examine collision severity and casualties in those same locations for 2024 and 2023. This is not mere technical refinement. A count spread across many locations implies something different from one concentrated around a single piece of infrastructure.

Timing would be another important dimension. A concentration during weekday commuting would require analysis of congestion and traffic exposure, whereas a pattern at night would call for a different review. Year-on-year comparisons should be made using equivalent time bands and traffic volumes rather than assigning any one journey type as the cause in advance.

For evaluating policy, the ideal denominator is not simply the number of kilometres of road included. It is how far vehicles travelled on the relevant sections, preferably with comparable information about vehicle types and time periods. Twenty additional incidents in a year with substantially more traffic mean something different for per-mile exposure than twenty additional incidents alongside lower traffic. The present mapped summary does not supply that calculation.

The available public data can establish which events occurred near the motorway, their classifications and their linked records. It cannot on its own document the effectiveness of a camera deployment, a junction design change or a police operation. Claims that an intervention worked, failed or caused the increase require dated evidence about the intervention and a comparison that accounts for other influences.

Northern Ireland's wider road-safety position

The M1's 2025 rise occurred during a year when Northern Ireland as a whole recorded an increase in injury collisions from 4,753 to 5,015. The national rise was approximately 5.5%, much smaller than the M1 corridor's 26% change. That comparison shows the mapped M1 figure moved more sharply between these particular years than the overall Northern Ireland count. It does not establish that its underlying risk rose more sharply, because one is a selected vicinity and the other the entire police-reported injury network.

The national serious-injury figures highlight why a simple incident comparison is inadequate. PSNI recorded 1,066 people seriously injured in 2025, 127 more than in 2024. Fatalities fell from 69 to 56. These opposite movements make it especially important to check who was hurt in the additional M1 corridor incidents instead of assuming an increase in one severity category from the overall total.

Across the full national thirteen-year database, PureCar displays 70,652 injury collisions and 109,562 casualties. That scale allows a broad historical perspective, but the central task of a motorway investigation is still local: to identify what actually occurred on the relevant alignment, rather than letting a national statistic stand in for route evidence.

A further numerical comparison makes clear what this story does and does not establish. The rise of 27 incidents in 2025 amounts to about 1.7% of all 1,611 nearby injury collisions recorded across thirteen years. That fraction is not a measure of importance or the share of journeys affected, because those 1,611 events accrued over many different traffic conditions. It is simply a useful scale check: the unusually large year-on-year percentage is produced by comparing two annual totals close to one hundred, not by evidence of an unprecedented overall collision volume.

The finding that survives scrutiny

The defensible headline is that police-recorded injury collisions near the selected mapped M1 corridor increased by 27 between 2024 and 2025, from 104 to 131. The figure was the highest since 2018, when 141 incidents were recorded, although it remained below the series peak of 153 in 2017. It also sits below the series peak of 153 in 2017.

The wider record points to an extended period with lower counts than in the middle of the 2010s, followed by a fluctuating recovery from the 2020 low. The thirteen-year casualty records show that the mapped area has also been the setting of fatal and serious injuries, but they do not allocate those injuries to 2025 without a separate annual breakdown.

This is enough to justify sustained public scrutiny, especially of junction-level concentrations and severity. It is not enough to pronounce the M1 Northern Ireland's most dangerous road, to assign blame to a type of driver or to claim a specific enforcement policy failed. The role of the data is to narrow the question to something verifiable. In this case the next answer needs to come from the exact locations and circumstances of the additional injury collisions, not from speculation attached to the size of the percentage increase.

Record and source details

The year-by-year data, section count and linked records are published in PureCar's M1 Collision History and Road Safety investigation. Its collision selection uses historical PSNI records and DfI geometry at a 100-metre radius. National comparisons come from the PSNI's 2025 Key Statistics Report, published on 27 March 2026. The free-running speed monitoring figure comes from the Northern Ireland Road Safety Strategy annual statistical report. Injury-only police reports are not a comprehensive record of all incidents or near misses.