Fire ecology

Fire is the master disturbance of the North American boreal forest. Lightning has ignited boreal fires for as long as there have been boreal forests, and most boreal plants and animals are adapted to a world in which much of the landscape burns every century or so. Yet fire is also a hazard to communities, a large source of carbon emissions and smoke, and a process that is changing fast as the climate warms. This lesson describes fire regimes, how fire behaves in boreal fuels, how fire weather is rated, what fire does to forests and soils, and how fire is managed, including the revival of Indigenous cultural burning.

Fire regimes

A fire regime is the pattern of fire characteristic of an area over long periods: how often fire occurs, how large fires are, when in the year they burn, how intensely they burn and how severely they affect vegetation and soil. Fire frequency is often expressed as the fire cycle, the time required to burn an area equal to the whole study area, or its reciprocal, the annual area burned as a fraction of the landscape. A landscape where 1 per cent burns each year has a fire cycle of 100 years. The fire return interval at a point is the time between successive fires there; if fire strikes at random with respect to stand age, the mean return interval equals the fire cycle.

North American boreal forests are dominated by stand-replacing crown fires. Black spruce and jack pine have low branches, ladder fuels of lichens and dead twigs, and resinous needles, so fire climbs easily into the crowns and runs through the canopy, killing most trees. Most of the area burned each year comes from a small number of very large fires, often tens or hundreds of thousands of hectares, many in remote areas. In parts of Eurasia, by contrast, Scots pine and larch forests often burn in surface fires that leave thick-barked trees alive. Typical boreal fire cycles in North America range from about 50 to 150 years in the dry continental west, such as interior Alaska, the Northwest Territories and northern Saskatchewan, to several centuries in the wetter east, such as Labrador and eastern Quebec.

Time since fire

If fire strikes stands regardless of their age, the proportion of a landscape older than age \(t\) follows a negative exponential distribution, \(A(t) = e^{-t/b}\), where \(b\) is the fire cycle. Fire history studies reconstruct a time-since-fire map from tree ages, fire scars and old maps, then fit this curve, or the more flexible Weibull distribution, \(A(t) = \exp[-(t/b)^c]\), in which a shape parameter \(c\) greater than 1 means that older stands are more likely to burn. In many boreal landscapes young stands burn less readily because they lack fuel, and very old stands may burn more readily, so \(c\) is often somewhat above 1.

# Time-since-fire distribution for the Sable watershed (synthetic stand ages).
import numpy as np
from scipy import stats
rng = np.random.default_rng(404)
cycle_true = 90.0
ages = rng.weibull(1.3, 600) * cycle_true            # stand ages in years, Weibull shape 1.3
print(f"stands: {len(ages)}, mean age {ages.mean():.0f} years, oldest {ages.max():.0f} years")
b_exp = ages.mean()                                  # MLE of the negative exponential fire cycle
c, loc, b_wei = stats.weibull_min.fit(ages, floc=0)
print(f"negative exponential fit: fire cycle {b_exp:.0f} years, annual area burned {100 / b_exp:.2f}%")
print(f"Weibull fit: shape c = {c:.2f}, scale b = {b_wei:.0f} years (c > 1 means older stands burn more readily)")
for t in (50, 100, 200):
    print(f"share of landscape older than {t:>3} years: exponential {np.exp(-t / b_exp):.2f}, Weibull {np.exp(-(t / b_wei) ** c):.2f}, observed {np.mean(ages > t):.2f}")

stands: 600, mean age 84 years, oldest 408 years
negative exponential fit: fire cycle 84 years, annual area burned 1.19%
Weibull fit: shape c = 1.28, scale b = 91 years (c > 1 means older stands burn more readily)
share of landscape older than  50 years: exponential 0.55, Weibull 0.63, observed 0.63
share of landscape older than 100 years: exponential 0.30, Weibull 0.32, observed 0.32
share of landscape older than 200 years: exponential 0.09, Weibull 0.06, observed 0.06

The fitted shape parameter above 1 says that the chance of burning rises with stand age here, and the Weibull curve tracks the observed share of old forest better than the exponential. Both fits put the fire cycle at roughly 85 to 90 years, meaning a little over 1 per cent of the watershed burns in an average year. Such estimates underpin forest management planning: if harvest and fire together remove forest faster than it regrows to maturity, the share of old forest declines.

How boreal fires behave

Fire behaviour depends on the fire triangle of fuel, weather and topography. Boreal fuels are layered: the forest floor of mosses, lichens and duff, surface fuels of shrubs and dead wood, ladder fuels of low branches, and the canopy. Fine, dry surface fuels ignite and carry fire; deep organic duff smoulders for days or weeks and determines how much carbon is released and how severely the soil is heated. Crown fires can spread several kilometres per hour, throw embers ahead to start spot fires, and release energy at rates far beyond what crews on the ground can fight. Some fires overwinter as smouldering combustion deep in peat and duff, so-called zombie fires, then re-emerge in spring; these have become more common in recent hot, dry years in Alaska and northern Canada.

Weather drives the daily and seasonal course of fire. Long summer days, low humidity, wind and drought dry fuels. Lightning from summer thunderstorms causes most of the area burned in remote boreal regions even though people start many more fires near roads and communities. Large fire years are associated with persistent blocking high-pressure ridges that bring heat and drought for weeks.

Canada rates fire danger with the Canadian Forest Fire Danger Rating System, developed by the Canadian Forest Service and used in many countries. Its Fire Weather Index (FWI) System uses daily noon temperature, relative humidity, wind speed and 24-hour rain to track three fuel moisture codes: the Fine Fuel Moisture Code for litter and fine fuels, which responds within a day; the Duff Moisture Code for loosely compacted organic layers, which responds over about two weeks; and the Drought Code for deep, compact organic layers, which responds over about seven weeks. These combine into an Initial Spread Index, a Buildup Index and the overall Fire Weather Index. The companion Fire Behaviour Prediction System predicts spread rate and intensity for fuel types such as C-2 boreal spruce and C-3 mature jack pine.

# A simplified fire weather screen for the station (illustrative thresholds, not the official FWI equations).
import numpy as np
rng = np.random.default_rng(44)
days = 92                                            # June to August
temp = 18 + 6 * rng.standard_normal(days).cumsum() / np.sqrt(np.arange(1, days + 1)) + 4 * np.sin(np.linspace(0, np.pi, days))
rh = np.clip(55 - 1.5 * (temp - 18) + rng.normal(0, 10, days), 15, 100)
rain = np.where(rng.uniform(size=days) < 0.22, rng.exponential(6, days), 0.0)
wind = np.clip(rng.normal(14, 6, days), 0, None)
dryness = 0.0
risk = []
for t, h, r, w in zip(temp, rh, rain, wind):
    dryness = max(0.0, dryness + max(0.0, t - 5) * 0.12 - r * 1.5)   # deep-fuel drying, reset by rain
    fine = (100 - h) / 100 * (1 if r < 1 else 0.3)                   # fine-fuel dryness today
    index = fine * (1 + w / 20) * (1 + dryness / 20) * 10
    risk.append('extreme' if index > 34 else 'high' if index > 24 else 'moderate' if index > 12 else 'low')
for level in ('low', 'moderate', 'high', 'extreme'):
    print(f"{level:>8}: {risk.count(level):2d} days")
print(f"longest run of high or extreme days: {max(len(s) for s in ''.join('x' if x in ('high', 'extreme') else '.' for x in risk).split('.'))}")

     low: 28 days
moderate: 27 days
    high: 22 days
 extreme: 15 days
longest run of high or extreme days: 7

This screen is deliberately simplified and should not be used for operational decisions, but it captures the logic of fire danger rating: fine fuels respond to today's humidity and rain, deep fuels integrate weeks of drying, and wind multiplies spread. Long runs of high and extreme days, not single hot afternoons, set up the largest fire years.

Recent fire seasons

Canada's 2023 fire season was the largest on record by a wide margin, with about 15 million hectares burned, more than twice the previous record, including more than 5 million hectares in Quebec alone. Some 230,000 people were evacuated at various times, including the entire city of Yellowknife, and smoke reached the eastern United States and Europe. In July 2024 a wildfire entered the town of Jasper, Alberta, damaging or destroying about a third of its structures. The 2025 season burned more than 7 million hectares, the second-largest area on record, driven by fires in Manitoba and Saskatchewan. Researchers attribute the trend toward larger seasons to hotter, drier conditions and longer fire seasons, consistent with climate change.

Fire effects

Fire severity, the degree of change to vegetation and soil, varies widely within a single fire. Ecologists measure it in the field with composite burn indices and from satellites using the differenced Normalized Burn Ratio, which compares near-infrared and shortwave-infrared reflectance before and after fire. Severity of burning of the organic layer is especially important in the boreal forest. A deep burn exposes mineral soil, which favours spruce and pine seedlings and also aspen and birch, while a shallow burn leaves a moss and duff seedbed on which conifer seedlings struggle. Deep burns also release more carbon and, where permafrost lies beneath, can trigger thaw by removing insulation, as lesson 2 showed.

Fire releases nutrients, raises soil temperature and opens the canopy, producing a flush of growth in shrubs, fireweed and young trees. It creates snags used by woodpeckers and cavity nesters; black-backed woodpeckers specialize on recently burned forests, feeding on wood-boring beetle larvae. Moose often increase on burns 10 to 30 years after fire as browse becomes abundant. Caribou, in contrast, depend on terrestrial lichens that take 50 to 100 years or more to recover after fire, so large fire years reduce winter range for decades, a point taken up in lesson 7. At the landscape scale, fire maintains the mosaic of stand ages that supports this diversity.

Carbon emissions from boreal fires are large. A severe boreal fire may release 2 to 4 kilograms of carbon per square metre, most of it from the organic soil rather than the trees. In 2023 Canadian fires released an estimated amount of carbon comparable to or larger than the country's annual fossil fuel emissions, according to several analyses, although regrowth gradually recaptures carbon over the following decades if the forest returns.

Managing fire

Fire management in the boreal forest is a matter of priorities. Agencies protect communities, infrastructure and commercial forests, and in remote areas often monitor rather than suppress fires that pose no threat, recognizing fire's ecological role and the impossibility of suppressing every fire. Alaska divides its land into full, critical, modified and limited protection zones; Canadian provinces and territories use comparable zones. FireSmart programs in Canada and Firewise and similar programs in the United States help communities reduce hazard by thinning conifers, planting fire-resistant aspen, cleaning roofs and gutters and creating fuel breaks around towns.

Indigenous peoples across the boreal forest used fire deliberately for millennia: spring burning of meadows, trails and wetland margins to improve grazing for wildlife, encourage berries and reduce the risk of summer fires near camps. Colonial fire suppression policies in the twentieth century largely ended these practices. Today Indigenous fire stewardship and cultural burning are reviving through community programs, partnerships with agencies and training of Indigenous fire crews, with research suggesting that such low-intensity spring burning can reduce hazard around communities while restoring cultural connections to the land.