How Do Wildfires Start? From Natural to Human Causes

Humans start the vast majority of wildfires. Across the contiguous United States, people are responsible for roughly 84% of all wildfires and about 44% of the total area burned, adding an average of 40,000 extra fires per year beyond what lightning alone would produce.¹ Lightning accounts for most of the remainder, and a handful of rarer natural mechanisms fill in the gaps. But the question of how a wildfire starts is really two questions: what provides the initial spark, and what allows that spark to catch and spread?

Lightning as the Dominant Natural Cause

Lightning is by far the most common natural ignition source for wildfires worldwide. Not every bolt starts a fire, though. The critical ingredient is something called a continuing current, a sustained flow of electrical charge that lasts well after the bright flash fades. Researchers have found that a lightning flash carrying a continuing current lasting more than about 10 milliseconds has a higher probability of igniting vegetation than an ordinary cloud-to-ground strike.² One video-recorded fire-igniting strike showed a continuing current lasting roughly 400 milliseconds and produced a visible glow on the ground lasting two to four seconds, long enough to heat fuel to combustion. The mechanism behind these continuing currents is still an active area of research.³

Most lightning-ignited fires happen during the warm, dry months when thunderstorms roll over parched landscapes. In the United States, the lightning fire season is concentrated in summer, whereas human-caused fires stretch across a season three times longer.¹ That seasonal concentration means lightning fires, while fewer in number, tend to burn under the hottest, driest conditions and can account for a disproportionate share of total burned area.

Rarer Natural Ignition Sources

Beyond lightning, nature has a few other ways of starting fires, though they are far less common. Volcanic eruptions can ignite surrounding vegetation when lava flows encounter dry organic material. During the 2023 Litli-Hrútur eruption in Iceland, lava flows triggered the largest moss wildfires in the country’s modern record-keeping history.⁴ These volcanically ignited wildfires create a cascading hazard: the fire generates its own plume of particulate matter that mixes with volcanic emissions, compounding the air-quality threat. Researchers expect such events to become more frequent as climate change extends wildfire-prone conditions globally.

Even more obscure is self-heating ignition, where organic soil catches fire with no external spark at all. Peat and other carbon-rich soils undergo slow exothermic reactions when oxygen is present. Under the right thermal conditions, that slow heat buildup can eventually reach a tipping point and ignite smoldering combustion. Laboratory work has shown that self-heating can occur even in soils with very high mineral content, though the required ambient temperature climbs steeply as mineral content rises.⁵ Simulations suggest that a peat layer about a meter and a half deep could self-ignite at average ambient temperatures above 40°C.⁶ That threshold is already being reached in some tropical and subtropical peatlands, making spontaneous peat ignition a real if uncommon contributor to global fire.

The Human-Caused Fire Landscape

The sheer variety of ways people start wildfires is striking. Power line failures are one of the best-known culprits: when a component breaks, or when high-voltage conductors contact other wires or vegetation, the resulting electrical arc or spray of hot material can land in dry brush and ignite a fire.⁷ Hot metal fragments and sparks also come from grinding, welding, overheated vehicle brakes, exhaust systems, and railroad operations. Together, power lines, equipment use, and railroads account for an estimated 28,000 wildland fires per year in the United States alone.⁸

Debris burning, escaped campfires, and discarded cigarettes round out the list of common accidental causes. Prevention education campaigns have been shown to meaningfully reduce fires from all three of those categories, along with fires started by children playing with fire.⁹ Arson is a persistent problem too, particularly around forestry plantations and areas where wildland and urban development meet.¹⁰ Even fireworks and firearms discharge contribute a small share; in one Swedish analysis, fireworks and related pyrotechnics accounted for about 1.5% of wildfire-building incidents.¹¹

What makes human ignitions so consequential is not just their frequency but their timing and location. People start fires in places lightning rarely strikes and during seasons when natural fires would be unlikely. That effectively tripled the length of the fire season in the US and pushed fire activity into landscapes, such as the southeastern coastal plain and the wildland-urban interface, that would otherwise see far less burning.¹

Can a Broken Bottle Really Start a Wildfire?

This is one of the most enduring pieces of fire-safety lore: a shard of glass focuses sunlight like a lens and torches the forest floor. The reality is far less dramatic. In controlled experiments, researchers tested whether glass bottle fragments could ignite natural fuels. The best-performing bottle bottom managed to raise temperatures to 327°C and lightly charred spruce needles, beech leaves, and grass blades, but no flame ever formed.¹² In a more realistic setup, with a bottle bottom simply lying on leaf litter as it would in the wild, the temperature beneath the glass peaked at just 82°C. That is nowhere near enough to ignite anything. So while the physics of light concentration is real, the geometry of a random glass fragment sitting on uneven ground is almost never favorable enough to start a fire. Litter on roadsides deserves cleanup for other reasons, but the “glass magnifying lens” scenario is essentially a myth.

Why Some Sparks Catch and Others Don’t

An ignition source is only half the equation. Whether a spark, ember, or lightning strike actually starts a spreading fire depends heavily on how dry the fuel is. Dead plant material has a moisture threshold below which it will ignite and above which it simply will not catch. The specific threshold varies dramatically by fuel type. In experiments with northeast Chinese forest fuels, some materials refused to ignite at 15% moisture content, while others caught fire even at 40% moisture when exposed to a match flame.¹³ The ignition source itself matters too: a cigarette butt, being a weaker heat source than a match, could only ignite those same fuels at a lower moisture threshold of about 27%.

Forest structure also plays a role. Dense, wet eucalypt forests need drier fuel and drier air to ignite than sparser, more open forests in the same region, because the canopy traps humidity and slows fuel drying.¹⁴ The strongest predictors of whether a particular patch of forest will ignite are the moisture content of dead fine fuel on the ground and the vapor pressure deficit within the forest itself. This means that two identical sparks, one landing in a shaded, humid gully and one landing on a sunlit, wind-exposed ridge, can have completely opposite outcomes.

Environmental Conditions That Amplify Fire Risk

Vapor pressure deficit, a measure of how aggressively the atmosphere sucks moisture out of fuels and vegetation, has become one of the most important fire-weather indicators in the western United States. It is linked to increases in burned area across the region and has been trending upward in recent decades.¹⁵ The relationship between atmospheric dryness and fire is not always straightforward, though. When dry spells are short, they increase fuel flammability and fire severity. When they persist for longer periods, they can actually slow the growth of vegetation and reduce the accumulation of combustible material, which in some boreal settings paradoxically lowers fire severity over time.¹⁶

Invasive grasses have fundamentally reshaped fire regimes in the arid western United States. Cheatgrass, originally from Eurasia, fills in the gaps between native shrubs with a continuous carpet of fine, fast-drying fuel. Satellite data show that cheatgrass-dominated land burned at more than double the rate of all other vegetation types combined over recent decades, and during the 1990s the disparity was nearly four to one.¹⁷ Cheatgrass was also disproportionately represented in the largest fires and was more likely to be the point of origin in fires that spread across multiple vegetation types. The grass-fire cycle, where fire kills native shrubs, cheatgrass colonizes the bare ground, and the next fire comes sooner, has transformed millions of acres across the Great Basin and beyond.¹⁸

Climate Change and the Future of Fire Ignition

Climate change affects wildfire ignition in two reinforcing ways: it dries out fuels, making existing ignition sources more effective, and it increases lightning activity, adding more natural ignition sources to the landscape. Machine-learning models trained on recent data show that even over less than a decade, climate shifts have steadily raised the global risk of lightning-ignited wildfires.¹⁹

Projections for the western United States are especially concerning. Earth-system modeling points to widespread increases in days with cloud-to-ground lightning by mid-century, with some northern areas seeing 4 to 12 additional lightning days per year compared to the recent past.²⁰ Globally, simulations under a moderate warming scenario project that total lightning flash rates could increase by roughly 40% by the 2090s, driven by the relationship between rising surface temperatures and atmospheric instability.²¹ More lightning over drier fuel means more fires, even in landscapes with no increase in human activity.

Detecting Ignition Before It Becomes a Disaster

One reason early detection matters so much is that a fire in its first minutes is vastly easier to contain than one that has had an hour to grow. Satellite-based systems that detect active fires and measure their radiated energy can deliver data in near-real time, making them useful for early warning and tracking how quickly a fire is spreading.²² Unlike burned-area mapping, which looks at post-fire imagery after the fact, active-fire detection is designed for speed. Networks of ground-based cameras, often using machine-learning image recognition, supplement satellite coverage in regions where rapid detection is a priority. Utility companies in fire-prone areas have also begun installing weather stations and de-energization protocols to shut off power lines preemptively when fire weather conditions exceed safe thresholds, directly targeting one of the most destructive human ignition pathways.

Indigenous Cultural Burning and the Longer View

For thousands of years before European colonization, Indigenous communities across North America, Australia, India, and elsewhere used deliberate, small-scale burning to manage their landscapes.²³ These were not wildfires. They were carefully timed burns tied to the ecology of specific plants, seasonal movement patterns, and spiritual practices. In one study of the Karuk Aboriginal Territory in California, researchers estimated that prior to colonization, cultural burning practices produced roughly 7,000 annual ignitions across the landscape, averaging about six to seven ignitions per fire steward per year.²⁴ Those frequent, low-intensity burns kept fuel loads from building up, maintained habitat diversity, and supported food and fiber resources.

The suppression of cultural burning during the colonial era removed a key pressure valve from fire-adapted ecosystems. Decades of fuel accumulation followed, setting the stage for the severe, landscape-scale wildfires that now dominate the headlines. Indigenous practitioners in the Pacific West and elsewhere are increasingly advocating for the restoration of cultural burning, and their goals generally align with broader ecological restoration objectives, though their philosophy tends to be holistic rather than narrowly focused on fuel reduction.²⁵ Many tribal practitioners would like to burn more widely and frequently, but legal, regulatory, and jurisdictional constraints limit how much burning they can do. The irony is hard to miss: a landscape shaped by thousands of years of intentional fire is now treated as one that should be protected from fire at all costs, and the resulting fuel buildup makes every unplanned ignition more dangerous than it otherwise would be.