Fire management in the boreal forest has relied on the idea that changes in fuel availability and structure caused by fire and forest succession will influence subsequent fire behavior, leading to self-regulation of fire activity at the landscape scale. However, fire managers and researchers in Alaska are noticing higher than expected incidence of burning in early successional and broadleaf vegetation, raising questions about the strength of these fuel and fire interactions. Here we synthesize the current state of knowledge about fire self-regulation in the North American boreal forest, carried out as a research collaboration between fire managers and the Bonanza Creek LTER program in Alaska. We examine the current weight of evidence for different aspects of fire self-regulation in the boreal forest, namely the effects of stand age and alternative vegetation types on probability of burning. We discuss the implications of fire self-regulation for landscape responses to increasing fire in boreal Alaska in the context of ongoing climate change. Finally, we explore the implications of this understanding for current management practices such as fuel treatments and use of prescribed fire to manage wildfire risk in the boreal forest.

In the boreal biome wildfire has increased in size, severity, and frequency under climate warming, prompting the need for fuel-reduction treatments to protect human communities and resources. However, the optimal ecosystem structure and timescale at which these treatments are effective operationally and in mitigating fire behavior is unknown. To address these knowledge gaps, we re-measured tree seedling density, soils, permafrost, vegetation, and ground cover across a network of thinned or shearbladed areas in interior Alaska and analyzed how the structure and flammability of these fuel breaks changed up to 17 years post-treatment. We also categorized each site into a fuel type and compared the rate of fire spread (ROS) at different moisture conditions and over time. In shearbladed sites, conifer tree recruitment was minimal (0.2 stems yr-1) and there were 12 times more deciduous than conifer seedlings, indicating that shearbladed treatments will likely become deciduous-dominated stands. In thinned sites, both conifer and deciduous tree seedling density were stable over time but highest in areas with thin, disturbed, soil organic layers. Permafrost thaw increased in shearbladed and thinned treatments (2.8 and 2.1 cm yr-1, respectively), yet was ~1.5 times greater in shearbladed than thinned areas. Thinned and unmanaged sites were consistently dominated by evergreen shrubs whereas grass abundance increased in shearbladed sites over time. Thinning was more effective than shearblading in reducing ROS during dry moisture conditions. In shearbladed areas, ROS spread increased with years after treatment, with some sites returning to pre-treatment levels within a decade. We conclude that both treatments increase permafrost thaw and while thinning decreases flammability and ROS within the first few decades after treatment, mature shearbladed treatments will likely be more effective at reducing flammability and ROS due to high deciduous tree establishment and survival.

The boreal forest zone is the second largest contiguous forest zone accounting for ~30% of all of the world's forested area and ~20% of the world's terrestrial carbon sink with most models predicting that the boreal will remain a sink for the rest of the century. These regions also have some of the largest areas of intact and unmanaged forest in the world. Increases in temperature and changes in precipitation have been linked to an increase in the frequency and severity of boreal forest fires, especially in southern Siberia. A growing number of field studies have observed the failure of tree species to regenerate after stand replacing fires, leading to the permanent loss of forest and a transition to a steppe ecosystem. We use a combination of remotely sensed data, field observations and statistical models to determine the current and future risk of widespread forest loss in Siberia. Our study suggests that climate change has already pushed parts of Siberia past the point where boreal forests can be sustained and raises the possibility that large areas of the southern boreal forests are approaching a tipping point which could result in widespread forest collapse and a permanent shift to steppe biome.

Climate warming leads to increase of fires number and burned area within Siberian Arctic. In the second decade of the 21st century, the number of fires increased 2.3 times compared to the first decade. The northern burning boundary is migrating northward and fires already reached the Arctic Ocean shore in the Eastern Siberia. The extremes of the boundary migration synchronized with air temperature and drought index SPEI anomalies (“heatwaves” and “drought waves”).

Meanwhile, burning rate response to warming was uneven in the different sectors of Siberian Arctic. Fire frequency, burned area and burning period increased in the Western and Eastern Siberia, whereas in the zone of Pacific Ocean influence (the Far East sector) a fire rate was not changed. Burning rate was increasing with air temperature and drought increase and it was decreasing with precipitation and ground cover moisture content increase.

Lightning strokes frequency increase via warming led to fire number and burned area increase in the Arctic. Within burns in Arctic tundra and forest tundra, the EVI (Enhanced Vegetation Index) and GPP (Gross Primary Productivity) quickly (within ~ 10 - 15 yr.) recovered and regularly surpassed the pre-fire level. GPP and EVI indicated positive trends on the major part of the Siberian Arctic. These GPP and EVI increases in combination with rapid vegetation recovery on burns may support Arctic ecosystems as a carbon sink in spite of increasing pyrogenic carbon losses.

Very large boreal forest fires usually take place in Russia and Canada, but significant fire events have occurred in Fennoscandia in recent years. The objective of our study is to understand the wildfire vulnerability of forest areas in Sweden and Finland by analyzing spatially explicit historical burned areas and their connection to weather, landscape and fire-related socioeconomic factors. Specifically, we question: what factors make the forest in Finland less vulnerable to fire as compared to Sweden, despite the similar climatic conditions?

In this study, we developed an extensive database for both countries at 1x1 km2 resolution (787124 grid cells), re-projected to the EPSG:3035 geographic coordinate system. Monthly burned areas in forests were processed from MODIS CCI (250 x 250 m2) and compared with burned areas reported by national statistics, EFFIS, and GFED 4.1 for 2001-2018. Spatial factors, including camping sites, lakes, forest management intensity, forest stand volume, roads, topographic features (aspect, slope, mean elevation), and changes in population density were collected from various sources. We used daily and sub-daily weather data from the HARMONIE forcing (0.125° x 0.125°) to calculate monthly Fine Fuel Moisture Code (FFMC) and Angström index values over 2001-2018.

The analysis of fire danger indices helped to identify a fire season spreading from April to September, with equivalent fire weather conditions found in both countries. The challenge in analyzing the collected data was the low number of historical fires, leading to the issue of sampling the very large number of cells without burned areas. In the talk, we will present the database and results of our modeling efforts. The emphasis will be on the different models’ predictive capability of the 2018 fire events in Sweden, and an outlook on possible future wildfire dynamics under climate change scenarios.

Fire is the dominant ecological disturbance in interior Alaska boreal forests and a strong control on landscape characteristics that affect freshwater processes and stream fish habitats. Fire frequency, size, and severity are changing across Alaska as a result of climate and land use change. Evidence from other ecoregions suggests fire negatively impacts fishes and aquatic habitats through removal of hillslope and riparian vegetation resulting in increased water temperatures and turbidity, and facilitation of further disturbance effects such as flooding and erosion. However, such disturbances also contribute to the creation and maintenance of stream habitats that provide a mosaic of dynamic habitats that support resilient populations. The overall goal of the five-year Boreal Fish and Fire Project is to investigate the effects of fire on boreal stream fish and their habitats though a series of field, lab, and modeling studies focused on elucidating relationships among climate, fire, the physical environment, and biological responses at multiple spatial scales. Our study area encompasses a 20,000 km2 region in interior Alaska that includes four river basins: the Chatanika, Chena, Salcha, and Goodpaster Rivers. These basins are important spawning and rearing habitats for fishes including Chinook Salmon and Arctic Grayling, and nearly one-quarter of this area has burned since the early 1980s. Here we highlight initial results from a suite of integrated, spatially-explicit models to identify where and when aquatic populations may be vulnerable to fire across this broad landscape. For the contemporary landscape, we explore potential interactions among observed fires, stream network topology, geomorphic conditions, and fish habitat suitability with consideration of the ability of riparian forest and valley bottoms to buffer streams from fire effects. Additionally, we will use output from dynamic ecosystem models to forecast vulnerability of boreal stream habitats to changes in flammability and active layer depth under future climate scenarios.

The two most abundant conifer species in the North America’s fire-prone boreal forest are Picea mariana (black spruce) and Pinus banksiana (jack pine). Both species exhibit cone serotiny, a reproductive adaptation common among woody plants living in habitats with high fire frequency. Serotinous cones provide an aerial seed bank that allows the release of seeds after a fire, facilitating post-fire recruitment. Here, we assessed the drivers of the reproductive maturity in black spruce and jack pine individuals and examined its relationship with post-fire recruitment. We assessed reproductive maturity by recording the presence or absence of cones in > 16,000 black spruce and jack pine individuals from recently burned (pre-fire values) and unburned stands, within 562 plots in two ecozones within the southern Northwest Territories (NWT) – the Taiga Plains and the Taiga Shield. We found that this simple metric corresponded well with seed rain. Individual trees reached reproductive maturity at smaller diameters in the Shield than the Plains, and jack pine became reproductive at smaller diameters than black spruce. The proportion of reproductive individuals per plot was also higher in the Shield than the Plains regardless of the age of the stands, for both species. We found that post-fire recruitment in recently burned plots was constrained by the proportion of reproductive trees prior to the fire, for both species in both ecozones. Nevertheless, both species showed lower post-fire recruitment in the Shield than the Plains. This is most likely the result of lower cone production in the Shield, a less productive ecozone with shallower soils than the Plains. Our results will help us understand how changes in the fire regime can alter the reproductive status and ultimately the post-fire recruitment of both tree species.

Boreal forests in northwestern Canada hold large reservoirs of carbon. Wildfires play an integral role in these ecosystems by driving plant succession and cycling nutrients. As wildfire regimes change, there is a need to understand the effect of fire on soil microbial ecology and its impact on post-fire ecosystem recovery. In this study, we investigated fire effects on soil physical and chemical properties and soil bacterial communities through laboratory wildfire simulations and incubations. We collected soil cores from 19 sites within Wood Buffalo National Park, located in Alberta and the Northwest Territories, Canada. We subjected paired cores to low and high moisture conditions before applying a heat flux in a cone calorimeter, in addition to a control, unburned core. We tracked soil temperature during and after the burn treatment, and measured soil pH, C:N, and moisture. We measured soil microbial respiration during a 5-week incubation and characterized bacterial communities using high-throughput amplicon sequencing. We hypothesized that fire would cause a larger increase in soil pH and higher rates of bacterial mortality in dry soil than in wet soil and expect bacteria with the ability to form spores, short replication times, and higher temperature tolerance thresholds to survive and thrive following the burn treatment. During the burn treatments, dry core temperatures ranged from 26 to 588 ºC, whereas wet core temperatures did not exceed 44 ºC. Higher severity fires under dry conditions caused an increase in soil pH and decreased rates of soil respiration at 18 of the 19 sites, which could be driven by microbial mortality at high temperatures and a reduction in easily-mineralizable C by combustion. Next, we will use high-throughput amplicon sequencing to investigate changes in post-burn bacterial community composition and to identify specific bacteria well-suited to the post-burn environment.

Climate change is altering the boreal wildfire regime through increases in the extent and severity of burning and reductions in fire return intervals. These changes have been shown to alter the regeneration trajectory of canopy species and ground vegetation, with implications for boreal wildlife habitat selection. One aspect of selection is the timing of use of burn habitats by wildlife species as their preferred forage taxa recover following fire, and how such recovery is mediated by environmental factors. Here, we aim to address these knowledge gaps through the following questions: 1) How does time after fire affect ground vegetation community assembly, in interaction with environmental factors? and 2) How is this likely to influence the recovery of focal wildlife species in burned areas? Vegetation data collected from 581 plots in the Northwest Territories, Canada, ranging in time from 1 to 100+ years post-fire, was used to model trends in the abundance of key forage taxa for several important wildlife species (e.g., caribou, moose), and to test the influence of time after fire on plant community composition. Time after fire was a significant driver of boreal plant community assembly, but accounted for only a small proportion of total variation in plant abundance. Patterns of post-fire plant recovery varied greatly among species, and were often strongly mediated by site drainage, sometimes altering the direction of the relationship. This suggests that, although time after fire influences wildlife forage over the long-term, site specific environmental conditions are also important and should be considered when assessing the implications of an intensified fire regime. The results of this research are intended for use by local communities, in order to anticipate and adapt to the consequences of increased burning, and by northern land-use managers charged with the effective conservation and management of wildlife habitat.

To evaluate functioning of a mesotrophic larch dwarf shrub-Sphagnum fuscum bog, suffered from the catastrophic fire in 2008, we compared phytomass and production (NPP) on its undisturbed and burnt part.

According to our estimation, about 18 t ha-1 of phytomass was burned owing to the fire, which is more than 8 t C ha-1. The undisturbed community produce roughly the same amount of biomass in 3 years.

The NPP of the major producer of the community, the dwarf shrub-grass layer, increased 2 times on the burnt area after the fire. The NPP of the community ranged from 480 to 730 kg ha-1 on the undisturbed site, and 1020-1380 kg ha-1 on the burnt one. The succession of Polytrichum strictum Brid. J. developed in both areas. In the 10th year of observations, its production reached 201.7 g m-2 on the undisturbed part of the bog, and its phytomass was 1850 kg ha-1 on the undisturbed site and 640 kg ha-1 on the burnt-out one in the acrotelm (0-30 cm) of the bog.

On the 6th, 10th, and 11th years of the monitoring, the regenerating bog phytocenosis was influenced by repeated fires, which caused damage to plants of the dwarf shrub-grass layer and the tops of Polytrichum cushions. At the undisturbed area, the living phytomass of dwarf shrubs and grasses decreased by 18-43% as well as the phytomass of P. strictum sharply increased. In the 11 year after the fire in 2008, the rate of biological cycle defined as the ratio of phytocenosis mortmass to NPP was increased by 32% in the bog part recovering from the fires and burns in comparison with the unburned one (8.4-5.7). This indicates that fires significantly change the structure of the phytomass and increase the productivity of the bog phytocenosis.

Soil surface CO₂ efflux (R_s) includes both plant (R_a) and microbial (R_h) respiration and is a key part of postfire C fluxes. It can account for most of total ecosystem respiration in boreal forests regenerating after high-intensity crown fires where most of the vegetation is killed (i.e. stand-replacing fires). Although surface fires leave most crowns intact, they may also substantially alter R_s, through ground vegetation mortality, delayed tree mortality and combustion of the soil organic layer. To better understand the effects of the type of fire on soil CO₂ fluxes, we measured R_a and R_h in burned and unburned (i.e. control) jack pine and black spruce forested areas in western Canada. The study sites had been burned at different times (0-3 years before this study) by both experimental surface and crown fires. In pine forests, R_a was not reduced immediately after a surface fire, but a decrease was detected 1-3 years afterwards. Conversely, R_h measured immediately after a surface fire was higher than in control, but no difference was detected in R_h measured 1-3 years after fire. R_a measured 3 years after crown fires did not significantly change, whereas R_h was lower than in control. In spruce forests, R_a measured 1-2 years after crown fires was lower than in control, whereas R_h did not significantly change. Given that R_a represented most of the R_s in all measured areas, our results suggest that although crown fires may immediately affect R_s through vegetation mortality, surface fires may also reduce R_s if they cause severe vegetation mortality, especially of trees. This result might not be the case in boreal forests composed of fire resister tree species (e.g. scots pine, larch) which bear traits to survive fire as opposed to fire embracer tree species (e.g. black spruce, jack pine, lodgepole pine).

Over the past half-century, the boreal region has experienced a temperature change 1.5 times higher than the global average, which includes warmer summers (hot, dry and windy), a longer growing season, and frequent drought conditions. Consequently, wildfires in the boreal forest have increased in frequency, severity, and acres burned. In the last two decades (2001-2020: 31.4 million acres) wildfires in Alaska have burned twice as many acres than the previous two decades (1981-2000: 14.1 million acres). In 2019, Alaska experienced 368 lightning-caused fires that burned 2.6 million acres. A significant number of fires burned in the Wildland Urban Interface (WUI) causing considerable property damage, prompted health issues from poor air quality, and heavily strained the limited local and national firefighting resources. With this rapid increase of fires, accurately assessing burn severity to understand fire behavior is essential for fire managers. In recent years, there have been rapid improvements in remote sensing methods and satellite image availability that have the potential to massively improve burn severity assessments of the boreal forest. In this study, we utilized pre- and post-fire Sentinel-2 satellite imagery of the 2019 Nugget Creek and Shovel Creek Burn Scars located in Interior Alaska to assess burn severity across the burn scars and test the effectiveness of machine learning classifiers (Random Forest and Support Vector Machine). We then compared the results of these machine learning classifiers to more traditional spectral indices-based burn severity (NBR, dNBR, NDVI and dNDVI) assessments. We validated all burn severity maps using the Composite Burn Index (a field-based assessment of burn severity). We found that both machine learning classifiers and the traditional spectral indices can accurately assess burn severity, however, machine learning classifiers (83% accuracy) outperform standard spectral indices (73% accuracy) when an adequate amount of field data has been collected.

Over multiple fire events, carbon (C) and nitrogen (N) accumulate in the soil organic layers of boreal forests. However, climate warming has led to more severe and frequent fires. If these accumulated pools are combusted, ecosystems could transition from net sinks to net sources for these elements. In addition, the soil organic matter that escapes combustion contributes to soil respiration and N mineralization, which regulates post-fire recovery of plant biomass. Thus, loss of accumulated organic matter may affect post-fire carbon and nutrient cycling, as well as net element balance over the fire cycle. After the 2014 record fire year in the Northwest Territories, we sampled soil organic matter from 41 black spruce (Picea mariana) stands across seven fire scars. We performed a 90-day soil incubation to investigate C and N cycling attributes of soils that differed in radiocarbon age, including soils that had no evidence of C accumulated over multiple fire cycles, and soils that accumulated C over many fire cycles. We found that 65% of soils respired carbon older than a single fire return interval and that older soil respired less CO2 to the atmosphere than younger soil. We also found that nitrogen mineralization occurred in older soils while microbial immobilization occurred in younger soils. Our findings suggest that older soil organic matter plays an important role in post-fire ecosystem recovery and may be particularly important for supplying nitrogen to regenerating vegetation.

After record high spring temperatures in 2020, Siberia experienced an unusually severe fire season, especially where extreme temperatures persisted throughout the summer. High fire activity in in northern Siberia (north of 65 degrees) was associated with severe summer drought that led to extreme fire behavior and long-lasting fires. Fires started at lower latitudes (50-55 degrees N) in March and expanded to mid latitudes (55-60 degrees N) in April. The bulk of the fire activity in June and July occurred from 60-70 degrees N, with the largest burned area during this period observed north of 65 degrees, near or above the Arctic Circle. Most of this far northern fire was in areas classified as forest, some of which was intermixed with peat bogs. In August and September, actively burning area decreased and most fire activity was at latitudes below 65 degrees N. Areas of active fire were decreased rapidly in early October. Over the fire season, about 29 million ha burned; approximately 27 percent of this was above 65 degrees North latitude. By comparison, in 2001 over 36 percent of burned area was in this far northern region. Because of the historically high interannual and spatial variability of wildland fires in Siberia, it is not possible to conclusively attribute the summer 2020 fires to changing climate, although severe fire seasons in northern Siberia can be expected to be more frequent in the future.