An official website of the United States government
Wildfires in the Arctic-boreal zone have increased in frequency over recent decades, carrying substantial ecological, social, and economic consequences. Remote sensing is crucial for mapping burned areas, monitoring wildfire dynamics, and evaluating their impacts. However, existing high-latitude burned area products suffer from significant discrepancies, particularly in Siberia, and their coarse spatial resolutions limit accuracy and utility. To address these gaps, we developed a convolutional neural network model to map burned areas at a 30-meter resolution across the Arctic-boreal zone using Landsat and Sentinel-2 imagery. Our model achieved promising results, with an Intersection Over Union (IOU) of 0.77 and an F1 score of 0.85 on unseen test data, performing better in North America (IOU=0.84) than Eurasia (IOU=0.72) due to differences in fire regimes and data quality. Predictions for six representative years showed our model’s burned area closely matched the median values of Landsat, MODIS, and VIIRS-based products, although alignment varied annually and spatially. Visual assessments indicated our approach was generally more accurate, notably in detecting unburned vegetation islands within fire perimeters missed by other products. This research has numerous potential applications, such as analyzing feedback between vegetation and burn patterns, characterizing spatial dynamics of unburned islands, and improving carbon emission estimates through detailed burn severity assessments. Here we have provided the primary series of scripts used to achieve the above results. In these scripts we use historical vector fire polygons to download imagery from Landsat 5, 7, 8, 9 and Sentinel-2 to train a deep learning model called a UNet++ in the Arctic-boreal zone. Imagery is downloaded from Google Earth Engine, while all other processing is done locally. The series of 6 scripts describes main steps from downloading training data, pre-processing it, training the model, and applying the model across the Arctic Boreal Zone. All scripting is done in python through .py scripts and Jupyter notebooks (.ipynb). Our study area includes Alaska, Canada and Eurasia, and we trained our model on all historical fire polygons from 1985-2020.
more »
« less
Annual and Seasonal Patterns of Burned Area Products in Arctic-Boreal North America and Russia for 2001–2020
Boreal and Arctic regions have warmed up to four times quicker than the rest of the planet since the 1970s. As a result, boreal and tundra ecosystems are experiencing more frequent and higher intensity extreme weather events and disturbances, such as wildfires. Yet limitations in ground and satellite data across the Arctic and boreal regions have challenged efforts to track these disturbances at regional scales. In order to effectively monitor the progression and extent of wildfires in the Arctic-boreal zone, it is essential to determine whether burned area (BA) products are accurate representations of BA. Here, we use 12 different datasets together with MODIS active fire data to determine the total yearly BA and seasonal patterns of fires in Arctic-boreal North America and Russia for the years 2001–2020. We found relatively little variability between the datasets in North America, both in terms of total BA and seasonality, with an average BA of 2.55 ± 1.24 (standard deviation) Mha/year for our analysis period, the majority (ca. 41%) of which occurs in July. In contrast, in Russia, there are large disparities between the products—GFED5 produces over four times more BA than GFED4s in southern Siberia. These disparities occur due to the different methodologies used; dNBR (differenced Normalized Burn Ratio) of short-term composites from Landsat images used alongside hotspot data was the most consistently successful in representing BA. We stress caution using GABAM in these regions, especially for the years 2001–2013, as Landsat-7 ETM+ scan lines are mistaken as burnt patches, increasing errors of commission. On the other hand, we highlight using regional products where possible, such as ABoVE-FED or ABBA in North America, and the Talucci et al. fire perimeter product in Russia, due to their detection of smaller fires which are often missed by global products.
more »
« less
- PAR ID:
- 10557859
- Publisher / Repository:
- MDPI
- Date Published:
- Journal Name:
- Remote Sensing
- Volume:
- 16
- Issue:
- 17
- ISSN:
- 2072-4292
- Page Range / eLocation ID:
- 3306
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Despite the low annual temperatures and short growing seasons that are characteristic of high northern latitudes (HNL), wildland fire is the dominant ecological disturbance within the region's boreal forest, the world's largest terrestrial biome. The boreal forest, also known as Taiga, is the band of mostly coniferous trees that stretches across the area north of the July 13°C isotherm in North America and Eurasia. Wildland fires also impact the tundra regions bordering the Taiga. This brief report updates our previous contribution to Arctic Report Card 2017. It summarizes evidence on variability and trends in fire disturbance in HNL, describes the fuels that characterize boreal and tundra ecosystems, and outlines how climate and subseasonal fire weather conditions in HNL influence the extent of area burned in a given year.more » « less
-
Abstract. As the northern high-latitude permafrost zone experiences accelerated warming, permafrost has become vulnerable to widespread thaw. Simultaneously, wildfire activity across northern boreal forest and Arctic/subarctic tundra regions impacts permafrost stability through the combustion of insulating organic matter, vegetation, and post-fire changes in albedo. Efforts to synthesis the impacts of wildfire on permafrost are limited and are typically reliant on antecedent pre-fire conditions. To address this, we created the FireALT dataset by soliciting data contributions that included thaw depth measurements, site conditions, and fire event details with paired measurements at environmentally comparable burned and unburned sites. The solicitation resulted in 52 466 thaw depth measurements from 18 contributors across North America and Russia. Because thaw depths were taken at various times throughout the thawing season, we also estimated end-of-season active layer thickness (ALT) for each measurement using a modified version of the Stefan equation. Here, we describe our methods for collecting and quality-checking the data, estimating ALT, the data structure, strengths and limitations, and future research opportunities. The final dataset includes 48 669 ALT estimates with 32 attributes across 9446 plots and 157 burned–unburned pairs spanning Canada, Russia, and the United States. The data span fire events from 1900 to 2022 with measurements collected from 2001 to 2023. The time since fire ranges from 0 to 114 years. The FireALT dataset addresses a key challenge: the ability to assess impacts of wildfire on ALT when measurements are taken at various times throughout the thaw season depending on the time of field campaigns (typically June through August) by estimating ALT at the end-of-season maximum. This dataset can be used to address understudied research areas, particularly algorithm development, calibration, and validation for evolving process-based models as well as extrapolating across space and time, which could elucidate permafrost–wildfire interactions under accelerated warming across the high-northern-latitude permafrost zone. The FireALT dataset is available through the Arctic Data Center (https://doi.org/10.18739/A2RN3092P, Talucci et al., 2024).more » « less
-
Abstract. The snow cover extent across the Northern Hemisphere has diminished, while the number of lightning ignitions and amount of burned area have increased over the last 5 decades with accelerated warming. However, the effects of earlier snow disappearance on fire are largely unknown. Here, we assessed the influence of snow disappearance timing on fire ignitions across 16 ecoregions of boreal North America. We found spatially divergent trends in earlier (later) snow disappearance, which led to an increasing (decreasing) number of ignitions for the northwestern (southeastern) ecoregions between 1980 and 2019. Similar northwest–southeast divergent trends were observed in the changing length of the snow-free season and correspondingly the fire season length. We observed increases (decreases) over northwestern (southeastern) boreal North America which coincided with a continental dipole in air temperature changes between 2001 and 2019. Earlier snow disappearance induced earlier ignitions of between 0.22 and 1.43 d earlier per day of earlier snow disappearance in all ecoregions between 2001 and 2019. Early-season ignitions (defined by the 20 % earliest fire ignitions per year) developed into significantly larger fires in 8 out of 16 ecoregions, being on average 77 % larger across the whole domain. Using a piecewise structural equation model, we found that earlier snow disappearance is a good direct proxy for earlier ignitions but may also result in a cascade of effects from earlier desiccation of fuels and favorable weather conditions that lead to earlier ignitions. This indicates that snow disappearance timing is an important trigger of land–atmosphere dynamics. Future warming and consequent changes in snow disappearance timing may contribute to further increases in western boreal fires, while it remains unclear how the number and timing of fire ignitions in eastern boreal North America may change with climate change.more » « less
-
Abstract. Fire is the dominant disturbance agent in Alaskan and Canadianboreal ecosystems and releases large amounts of carbon into the atmosphere.Burned area and carbon emissions have been increasing with climate change,which have the potential to alter the carbon balance and shift the regionfrom a historic sink to a source. It is therefore critically important totrack the spatiotemporal changes in burned area and fire carbon emissionsover time. Here we developed a new burned-area detection algorithm between2001–2019 across Alaska and Canada at 500 m (meters) resolution thatutilizes finer-scale 30 m Landsat imagery to account for land coverunsuitable for burning. This method strictly balances omission andcommission errors at 500 m to derive accurate landscape- and regional-scaleburned-area estimates. Using this new burned-area product, we developedstatistical models to predict burn depth and carbon combustion for the sameperiod within the NASA Arctic–Boreal Vulnerability Experiment (ABoVE) coreand extended domain. Statistical models were constrained using a database offield observations across the domain and were related to a variety ofresponse variables including remotely sensed indicators of fire severity,fire weather indices, local climate, soils, and topographic indicators. Theburn depth and aboveground combustion models performed best, with poorerperformance for belowground combustion. We estimate 2.37×106 ha (2.37 Mha) burned annually between 2001–2019 over the ABoVE domain (2.87 Mhaacross all of Alaska and Canada), emitting 79.3 ± 27.96 Tg (±1standard deviation) of carbon (C) per year, with a mean combustionrate of 3.13 ± 1.17 kg C m−2. Mean combustion and burn depthdisplayed a general gradient of higher severity in the northwestern portionof the domain to lower severity in the south and east. We also found larger-fire years and later-season burning were generally associated with greatermean combustion. Our estimates are generally consistent with previousefforts to quantify burned area, fire carbon emissions, and their drivers inregions within boreal North America; however, we generally estimate higherburned area and carbon emissions due to our use of Landsat imagery, greateravailability of field observations, and improvements in modeling. The burnedarea and combustion datasets described here (the ABoVE Fire EmissionsDatabase, or ABoVE-FED) can be used for local- to continental-scaleapplications of boreal fire science.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3390/rs16173306
