From the scorched heathlands of the New Forest to the smouldering valleys of south Wales and burned-out homes in Stourbridge, an extraordinary summer of wildfires has left its mark across Britain.
To most of us, these charred landscapes may look like pure devastation, but to a trained wildfire investigator they are full of clues. Heat-scorched leaves can become frozen in position, preserving evidence of the wind direction as the fire passed through. Soot or staining on rocks, the pattern of fallen grass and bracken, and even how those stems have burned, provide further information about how the fire itself moved.
By combining such clues, investigators can work backwards through a fire, progressively narrowing their search to a likely area of origin, then to a smaller patch within it. “At this point you’re on your hands and knees, looking in great detail,” said Andy Elliott, the UK’s only serving fire officer trained in specialist wildfire origin. “If, for example, somebody has used a match, or if a firework was the cause, the chances are you’ll find it.”
Yet, surprisingly few people are trained to read a burnt landscape in this way. While fire crews routinely record a probable cause, a 2023 Forestry Commission report warned that these judgments were generally made by personnel with limited or no specialist training, leaving their accuracy “open to scrutiny”.
More serious or complex fires may be referred to specialist fire investigators, but their expertise is generally in structural fires, rather than burning vegetation. “Wildfire investigation is a different discipline, almost a completely different world,” said Rob Gazzard, the Forestry Commission’s wildfire contingency planning adviser, who co-authored the report.
Elliot said: “At the moment there is no capability for wildfire fire investigation in the UK, full stop.” He spoke to the Guardian in his capacity as founder of the wildfire training and consultancy business WildfireTaC.
That knowledge gap is growing more problematic as the UK faces hotter, more fire-prone summers. Researchers and wildfire specialists are racing to develop the tools and skills to identify what starts wildfires and predict what will happen once they do.
Without specialist investigation, the Forestry Commission warned, preconceived ideas about what starts wildfires can influence investigators’ conclusions, potentially misdirecting prevention efforts. Take discarded cigarettes: despite UK campaigns often focusing on them, “research undertaken by wildfire scientists has determined that the cigarette as a cause is extremely unlikely and has been debunked by most eminent wildfire investigators”, the report said.
Similar uncertainty surrounds disposable barbecues, which various organisations have called on to be banned. While Elliott does not dispute that they can start fires, the existing data cannot reliably establish how often this happens – or what other causes may be going undetected.
“If we knew that 40% of our wildfires are caused by portable barbecues, you could ban them – though people may light another kind of fire – or you could educate people in how and where to use them safely,” said Elliott. “We’ve got three common causes of wildfires in the UK: men, women and children. Beyond that, we don’t know.”
The Forestry Commission is seeking to address this, with plans for two courses teaching people how to preserve evidence at wildfire scenes, and how to investigate wildfire origins and causes. Both are awaiting funding.
For Gazzard, understanding the cause of wildfires is fundamental to prevention. England alone experiences roughly 30,000 vegetation fire ignitions each year, and while most quickly fizzle out, understanding where, when and why they occur could reveal patterns, and prevent more from developing into major incidents. “Preparedness and prevention is where you win the wildfire,” said Gazzard. “It’s not about response.”
The need for such prevention is only likely to grow, as hotter, drier conditions make it easier for such ignitions to catch hold and spread. This year’s summer was the UK’s hottest on record, coupled with extreme and persistent drought. “The UK basically has been a tinderbox,” said Prof Stefan Doerr, a wildfire researcher based at Swansea University.
Although last year set a record for land burned by wildfires, by mid-August this year firefighters in England and Wales had already attended more wildfires than in the whole of 2025. “Fires are starting to become much more difficult to control. They’re also starting to destroy people’s homes, businesses and vehicles,” Elliott said.
Hotter, drier conditions could also mean that previously minor ignition sources start more fires. Lightning is one example, birds striking overhead power cables another. “The bird will typically receive a very considerable voltage, catch fire and drop to the ground,” said Elliott. Usually, wet or green vegetation prevents this going any further. “Now we’re seeing conditions where the vegetation is very dry, bird strikes could be contributing to more fires.”
The timing of Britain’s wildfires is changing too. Traditionally, they have predominantly occurred in late winter and spring, when dead grasses, bracken and other vegetation left over from winter provide plentiful fuel. By summer, however, most of the landscape has greened up, creating a natural barrier to fire. Prolonged drought can disrupt this, creating the kind of parched, brown summer landscape more familiar in the Mediterranean.
A recent study published in the journal Environmental Research Letters found that the UK’s wildfire season lengthened from between one and four months in 2011-16 to between six and nine months in 2017-21. This year, many of the UK’s most significant fires occurred during summer. “Summer fires are much more difficult to put out, and they’re also much more dangerous in many ways,” Doerr said.
One reason is that summer droughts can dry out grasses and crops extending right up to towns and homes. Fires can then become too intense to tackle directly and rapidly escalate. “They’ll cross roads quite happily, and they move into people’s gardens and into their homes,” said Elliott.
The environmental consequences can also be more severe. Sarah Baker, the assistant director of the Wildfire Lab at the University of Exeter, who led the study on the UK’s lengthening fire season, has been examining the aftermath of recent summer fires on British moorlands. “We’re seeing a lot more fires that are capable of burning down into the underlying peat layers,” she said. “We’re getting measurements of up to 20 centimetres in some areas, which is quite a chunk of soil gone, consumed by the fire.”
This could hinder the regeneration of burnt moorland and release additional carbon as peat burns.
If preventing ignitions is one part of the challenge, predicting how and where they will spread is another. Until recently, efforts to predict wildfire risk have largely focused on the weather, but the fuel also matters. This is particularly true of the UK, where the condition and flammability of vegetation changes dramatically through the seasons.
To capture this complexity, Baker and other researchers have been burning samples of British wildfire fuels inside a laboratory cone calorimeter, measuring how readily they ignite and how much heat and energy they release at different times of year.
These experiments have contributed to the development of a UK wildfire-behaviour prediction system known as FireInSite, which firefighters are increasingly using to predict how rapidly a fire may spread through the vegetation in front of them and how large its flames could become, based on the type and quantity of that vegetation, the season and current weather conditions.
By looking up to five days ahead, FireInSite can help firefighters and land managers identify “windows of opportunity” for tackling a fire, as well as devising an appropriate strategy. “If, for example, it tells you you’re going to get flame lengths of six to eight metres, then attacking it with a fire beater is just not going to work, because you’re looking at helicopters,” said Elliott. “It also tells you when to step back and let the fire burn.”
Useful as FireInSite is, its predictive power is limited. “Our model tells you how a fire will burn. It won’t tell you where the fire will go,” said Doerr. This requires a different kind of modelling, combining information about how individual fuels burn, with terrain, weather and the precise patchwork of vegetation across a landscape.
Fire spread models exist in other countries such as Canada, but they are built around local fuel types, rather than British vegetation. The goal is not necessarily to predict the precise path of an individual wildfire in real time, which remains extremely tricky. “There’s a certain element of randomness with fire that no model will ever be able to recreate,” said Prof Tom Smith, of the London School of Economics, who is trying to adapt fire-spread models to UK landscapes.
Instead, researchers could simulate thousands of fires starting at different points and calculate the probability of a fire reaching particular places. This could identify high-risk communities or “pinch points” through which fires repeatedly spread, helping land managers target and test interventions such as removing or changing vegetation in strategic locations.
Modelling could also help high-risk communities develop evacuation plans, potentially avoiding tragedies such as the 2017 Pedrógão Grande fire in Portugal, where dozens died trying to escape the flames. “The beauty of modelling is you can run hundreds of [simulations], and you don’t destroy any landscape. You don’t kill anybody,” Elliott said. “You just keep running the models, and you learn from them.”
A big obstacle is that Britain currently lacks a sufficiently detailed map of its wildfire fuels. FireInSite relies on users selecting the vegetation they see in front of them, but a spread model needs to know what vegetation a fire will encounter as it travels across the landscape, as well as the terrain and weather. Fires generally spread faster uphill, while the direction a slope faces can affect how dry it is.
The patchiness of the British landscape makes this fine detail particularly important. Smith said: “In the UK, a lot can happen in 30 metres.”
For now, researchers can construct detailed fuel maps for relatively small areas using aerial imagery, but doing this for the whole country is another matter. Even the difference between a 1-metre and 2-metre-wide path could determine whether flames cross from one patch of vegetation to another, said Smith.
Yet it is worth persevering. Severe wildfires were once unusual features of Britain’s green, frequently damp summers. Increasingly they look like something we will have to learn to live with. Understanding how they start and where they are likely to go may be our best shot at stopping more of them from becoming disasters.
