Mapping Burn Extent of Large Wildland Fires from Satellite Imagery Using Machine Learning Trained from Localized Hyperspatial Imagery

Devastating large outdoor fires have been responsible for the destruction of vast amounts of infrastructure and the loss of human life. Wildland fires that spread into urban areas, known as wildland-urban interface (WUI) fires, are capable of enormous destruction. WUI fires are distinct from wildland fires; WUI fires include the combustion of both vegetative and human-made fuels and occur where population centers exist, whereas wildland fires include the combustion of vegetative fuels and occur in uninhabited areas. The rise of densely populated urban areas has also seen the development of large urban fires. In addition, the rise of informal settlement communities in Southeast Asia and Africa continues to result in large outdoor fires capable of great destruction. As a wildland fire reaches an urban area, structure-to-structure fire spread processes will occur via the same mechanisms as those in informal settlement fires and urban fires: radiant heat, direct flame contact, and firebrands. Firebrands are in fact the main culprit to destroy structures in large outdoor fires. An additional complication is related to the showers of firebrands produced in large outdoor fires from vegetative and structural fuel combustion, and in the presence of the combustion of these fuel types, there is also radiant heat produced that may only enhance ignition processes. These combined ignition processes have been overlooked and are the subject of this experimental investigation. To this end, a custom experimental apparatus that produces firebrand showers, in the presence of applied radiant heat, was installed inside a wind facility. The firebrand generator was tuned to produce firebrands typical of those quantified from the combustion of structures and the coupled influences of firebrand showers and radiant heat, under various wind speeds was determined. The total firebrand mass for smoldering ignition (SI) decreased as the wind speed and preheating time increased. In all experiments, SI was observed to transition to flaming ignition (FI). To better understand the experimental observations, a simple model was developed and is presented.

Aim of study: Large wildland fires (LWF) are major disturbance processes affecting many ecosystems each year. In last decades, socio-economic changes have contributed to major changes in land uses. This study assess trends in number, burned area and average size of large wildfires (> 100 ha) from 1978 to 2010 in Spain.Area of study: This work analyzes three clearly different regions of Spain (Mediterranean coast, MC, Mediterranean Interior, MI, Northwestern Spain, NW).Material and Methods: We studied historical wildland fire data from Spain’s EGIF database (General Statistics on Wildland Fires). We selected only wildland fires larger than 100 ha. All LWF were analyzed to test trends in number of fires, burned area and mean fire size.Main results: The number of LWF decreased in all regions but the burned area only decreased in MC and NW regions. However, both the number of LWF and the burned area did not decrease after 1995 in any region. The average size of LWF did not change in any of the three regions. Fires larger than 500 ha were very significant due to the high percentage of area burned in relation to the total area burned by fires larger than 100 ha (79.3 % in MC, 63.9 % in MI, and 35.7% in NW).Research highlights: After 1995, the number of LWF and burned area did not decrease. Additional actions are required including learned lessons from past LWF spread, and better trained fire suppression workers and more fuel management.Keywords: large wildland fires; trends; forest management; Spain.

Detection of forest-fire smoke plumes by satellite imagery

Forest and wildland fires are a natural part of ecosystems worldwide, but large fires in particular can cause societal, economic and ecological disruption. Fires are an important source of greenhouse gases and black carbon that can further amplify and accelerate climate change. In recent years, large forest fires in Sweden demonstrate that the issue should also be considered in other parts of Fennoscandia. This final report of the project “Forest fires in Fennoscandia under changing climate and forest cover (IBA ForestFires)” funded by the Ministry for Foreign Affairs of Finland, synthesises current knowledge of the occurrence, monitoring, modelling and suppression of forest fires in Fennoscandia. The report also focuses on elaborating the role of forest fires as a source of black carbon (BC) emissions over the Arctic and discussing the importance of international collaboration in tackling forest fires. The report explains the factors regulating fire ignition, spread and intensity in Fennoscandian conditions. It highlights that the climate in Fennoscandia is characterised by large inter-annual variability, which is reflected in forest fire risk. Here, the majority of forest fires are caused by human activities such as careless handling of fire and ignitions related to forest harvesting. In addition to weather and climate, fuel characteristics in forests influence fire ignition, intensity and spread. In the report, long-term fire statistics are presented for Finland, Sweden and the Republic of Karelia. The statistics indicate that the amount of annually burnt forest has decreased in Fennoscandia. However, with the exception of recent large fires in Sweden, during the past 25 years the annually burnt area and number of fires have been fairly stable, which is mainly due to effective fire mitigation. Land surface models were used to investigate how climate change and forest management can influence forest fires in the future. The simulations were conducted using different regional climate models and greenhouse gas emission scenarios. Simulations, extending to 2100, indicate that forest fire risk is likely to increase over the coming decades. The report also highlights that globally, forest fires are a significant source of BC in the Arctic, having adverse health effects and further amplifying climate warming. However, simulations made using an atmospheric dispersion model indicate that the impact of forest fires in Fennoscandia on the environment and air quality is relatively minor and highly seasonal. Efficient forest fire mitigation requires the development of forest fire detection tools including satellites and drones, high spatial resolution modelling of fire risk and fire spreading that account for detailed terrain and weather information. Moreover, increasing the general preparedness and operational efficiency of firefighting is highly important. Forest fires are a large challenge requiring multidisciplinary research and close cooperation between the various administrative operators, e.g. rescue services, weather services, forest organisations and forest owners is required at both the national and international level.

In Canada, the average annual area of burned forest has increased from around 1 million ha in the 1970's to over 2.5 million ha in the 1990's. A previous study has identified the link between anomalous mid-tropospheric circulation at 500 hPa over northern North America and wildland fire severity activity in various large regions of Canada over the entire May to August fire season. In this study, a northern North American study region of the hemispheric gridded 5° latitude by 10° longitude 500 hPa dataset is identified and analysed from 1959 to 1996 for a sequence of six monthly periods through the fire season, beginning in April and ending in September. Synoptic types, or modes of upper air behavior, are determined objectively by the eigenvector method employing K-means cluster analysis. Monthly burned areas from the Canadian Large Fire Database (LFDB) for the same period, 1959 to 1996, are analysed in conjunction with the classified monthly 500hPa synoptic types. Relationships between common monthly patterns of anomalous upper flow and spatial patterns of large fire occurrence are examined at the ecozone level. Average occurrence of a monthly synoptic type associated with very large area burned is approximately 18% of the years from 1959 to 1996. The largest areas burned during the main fire (May to August) season occur in the western Boreal and Taiga ecozones - the Taiga Plains, Taiga Shield, Boreal West Shield and Boreal Plains. Monthly burned areas are also analysed temporally in conjunction with a calculated monthly zonal index (Zi m ) for two separate areas defined to cover western and eastern Canada. In both western and eastern Canada, high area burned is associated with synoptic types with mid-tropospheric ridging in the proximity of the affected region and low Zi m with weak westerlies and strong meridional flow over western Canada.

Large outdoor fires have become commonplace all over the world. The International Organization for Standardization (ISO) defines large outdoor fires as an urban fire, tsunami-generated fire, volcano-generated fire, wildland-urban interface (WUI) fire, wildland fire, or informal settlement fire, where the total burnout area is significant. Perhaps of all the large outdoor fires, it is wildland fires that spread into urban areas, simply called WUI fires that attract the most attention. A glance at the recent headlines in the summer of 2022 reveals numerous catastrophic WUI fires all over Europe. Across the Atlantic Ocean in the USA, there is yet another destructive WUI fire raging in the USA state of California. With the increasing risks from a changing climate, these large outdoor fire disasters are only going to become more and more commonplace all over the world. More homes will be lost and more lives will be lost. It is the authors opinion that a targeted, multi-disciplinary approach is needed to address the large outdoor fire problem. In this short, invited paper to Horizons in Mechanical Engineering, it is argued that large outdoor fire problem is a fascinating and challenging research area and that engineers have the necessary skills and training to impact a problem that influences millions upon millions of people all over the world. An important danger, present in all large outdoor fires, are firebrands. Firebrands are introduced for non-specialist readers, and the most recent literature is reviewed. Several challenges are discussed, in particular, areas where engineers may help move the needle forward on this globally important topic.

Interest in small unmanned aircraft systems (sUAS) for topographic mapping has significantly grown in recent years, driven in part by technological advancements that have made it possible to survey small- to medium-sized areas quickly and at low cost using sUAS aerial photography and digital photogrammetry. Although this approach can produce dense point clouds of topographic measurements, they have not been tested extensively to provide insights on accuracy levels for topographic mapping. This case study examines the accuracy of a sUAS-derived point cloud of a parking lot located at the Citizens Bank Arena (CBA) in Ontario, California, by comparing it to ground control points (GCPs) measured using global navigation satellite system (GNSS) data corrected with real-time kinematic (RTK) and to data from a terrestrial laser scanning (TLS) survey. We intentionally chose a flat surface due to the prevalence of flat scenes in sUAS mapping and the challenges they pose for accurately deriving vertical measurements. When the GNSS-RTK survey was compared to the sUAS point cloud, the residuals were found to be on average 18 mm and −20 mm for the horizontal and vertical components. Furthermore, when the sUAS point cloud was compared to the TLS point cloud, the average difference observed in the vertical component was 2 mm with a standard deviation of 31 mm. These results indicate that sUAS imagery can produce point clouds comparable to traditional topographic mapping methods and support other studies showing that sUAS photogrammetry provides a cost-effective, safe, efficient, and accurate solution for topographic mapping.

While Smokey’s presence is ubiquitous throughout the country – at events, at parades and in commercials – and has certainly helped reduce the number of catastrophic, human-caused wildland fire, the issue of arson, both by the intentional and the unintentional arsonist, is still a significant problem. The Eagle Creek Fire continues to prove to be a poster child for consequences of arson. Powerlines can also be a source of wildland fires and are considered one form on unintentional arson. The 2018 Camp Fire in California was started by power transmission sparks created during a high-wind event. One prime example of a large wildland fire with both intentional and unintentional arson is Rodeo-Chediski Fire that occurred in Arizona in 2002. Land management agencies including the United States Forest Service, the National Park Service and the Bureau of Land Management have historically had, and still have today, many challenges regarding management policies as well as public education and outreach regarding wildland fire.

Anticipating the consequences of climatic change for fire requires understanding of the causes of variation in historical fire regimes. We assessed the influence of annual and decadal variation in climate on fire regimes of ponderosa pine-dominated forests in eastern Oregon and Washington using existing, annually dated tree-ring reconstructions (1687–1994). In four watersheds, we compared the extent of low-severity fires (total area burned each year) to precipitation and the Southern Oscillation Index, a measure of variation in El Niño-Southern Oscillation (ENSO), which affects weather in this region. At the annual scale, large fires burned during dry years and El Niño years (low SOI) in all watersheds while small fires burned regardless of variation in these climate parameters. Large fires also burned during relatively wet years and La Niña years (high SOI) in one watershed, indicating that local factors can override regional climate controls in some locations. Climate from previous years did not influence current year's fire extent. The influence of ENSO on fire regimes in this region has not previously been demonstrated at these multicentury, regional scales. At the decadal scale, fire extent varied with precipitation, perhaps in response to variation in such climate features as the Pacific Decadal Oscillation. Several decades of low fire extent in the watersheds during the early 1800s was synchronous with a lack of fire at other sites in North and South America, probably in response to a change in the global climate that included a lessening in the frequency and/or intensity of ENSO events.

Fire frequency, extent, and size exhibit a strong linkage with climate conditions and play a vital role in the climate system. Previous studies have shown that the frequency of large fires in the western United States increased significantly since the mid‐1980s due to climate warming and frequent droughts. However, less work has been conducted to examine burned area and fire emissions of large fires at a national scale, and the underlying mechanisms accounting for the increases in the frequency of large fires are far from clear. In this study, we integrated remote‐sensed fire perimeter and burn severity data sets into the Dynamic Land Ecosystem Model to estimate carbon emissions from large fires (i.e., fires with size larger than 1000 acres or 4.05 km2) in conterminous United States from 1984 to 2012. The results show that average area burned by large fires was 1.44 × 104 km2 yr−1 and carbon emissions from large fires were 17.65 Tg C yr−1 during the study period. According to the Mann‐Kendall trend test, annual burned area and pyrogenic carbon emissions presented significant upward trends at the rates of 810 km2 yr−1 and 0.87 Tg C yr−1, respectively. Characteristic fire size (fire size with the largest contribution to the total burned area) in the period of 2004–2012 increased by 176.1% compared to the period of 1984–1993. We further found that the larger fires were associated with higher burn severity and occurred more frequently in the warmer and drier conditions. This finding implies that the continued warming and drying trends in the 21st century would enhance the total burned area and fire emissions due to the contributions of larger and more severe wildfires.

Abstract. Against the backdrop of global change, in terms of both climate and demography, there is a pressing need for monitoring of the global water cycle. The publicly available global database is very limited in its spatial and temporal coverage worldwide. Moreover, the acquisition of in situ data and their delivery to the database have been in decline since the late 1970s, be it for economical or political reasons. Given the insufficient monitoring from in situ gauge networks, and with no outlook for improvement, spaceborne approaches have been under investigation for some years now. Satellite-based Earth observation with its global coverage and homogeneous accuracy has been demonstrated to be a potential alternative to in situ measurements. This paper presents HydroSat as a database containing geometric quantities of the global water cycle from geodetic satellites. HydroSat provides time series and their uncertainty in water level from satellite altimetry, surface water extent from satellite imagery, terrestrial water storage anomaly represented in equivalent water height from satellite gravimetry, lake and reservoir water volume anomaly from a combination of satellite altimetry and imagery, and river discharge from either satellite altimetry or imagery. The spatial and temporal coverage of these datasets varies and depends on the availability of geodetic satellites. These products, which are complementary to existing products, can contribute to our understanding of the global water cycle within the Earth system in several ways. They can be incorporated for hydrological modeling, they can be complementary to current and future spaceborne observations, and they can define indicators of the past and future state of the global freshwater system. HydroSat is publicly available through http://hydrosat.gis.uni-stuttgart.de (last access: 18 May 2022​​​​​​​). Moreover, a snapshot of all the data (taken in April 2021) is available in GFZ Data Services at https://doi.org/10.5880/fidgeo.2021.017 (Tourian et al., 2021).

Increased spatial resolution has been shown to be an important factor in enabling machine learning to map burn extent and severity with extremely high accuracy. Unfortunately, the acquisition of drone imagery is a labor-intensive endeavor, making the capture of drone imagery impractical for large catastrophic fires, which account for the majority of the area burned each year in the western US. To overcome this difficulty, satellites, such as PlanetScope, are now available which can produce imagery with remarkably high spatial resolution (approximately three meters). In addition to having higher spatial resolution, PlanetScope imagery contains up to eight bands in the visible and near-infrared spectra. This study examines the efficacy of each of the eight bands observed in PlanetScope imagery using a variety of feature selection methods, then uses these bands to map the burn extent and biomass consumption of three wildland fires. Several classifications are produced and compared based on the available bands, resulting in highly accurate maps with slight improvements as additional bands are utilized. The near-infrared band proved contribute most to increased mapping accuracy, while the green 1 and yellow bands contributed the least.

Wildland fires that grow out of the initial attack phase are responsible for most of the damage and burned area. We model the allocation of fire suppression resources (ground crews, engines, bulldozers, and airdrops) to these large fires. The fireline at a given future time is partitioned into homogeneous segments on the basis of fuel type, available resources, risk, and other factors. Each is assigned a utility value corresponding to the importance of holding the segment. For a given resource allocation, the probability of holding the segment is modeled in terms of the (random) width of fireline built. The task is then to find the allocation that maximizes the expected total utility. With certain restrictions, it proves possible to formulate the optimization as a linear programming problem. Use of the model is demonstrated with a case study of a large fire representative of conditions on the Los Padres National Forest in southern California. One feature is that different assumptions about the uncertainty in the predictions of constructed fireline widths can lead to differences in the optimal resource allocations. Thus, if one inappropriately took the uncertainty to be zero (the deterministic case), the resulting allocation may well not be the optimal one. This illustrates the potential advantage of probabilistic modeling over the previous deterministic approach. For. Sci. 38(4):842-853

Abstract. It is critical to assess bushfire impact rapidly and accurately because bushfires play a significant role in forest degradation and present a threat to ecosystems and human lives. Over the past decades, several supervised algorithms of burn severity mapping have been proposed, facing the significant drawback of time-consuming labeling. Moreover, there is no robust framework for burn severity mapping through fusing multi-sensor, multi-resolution, and multi-temporal remote sensing imagery from satellite and aerial platforms. Therefore, this paper presents an unsupervised two-step pipeline: processing 2D data followed by 3D data for burn severity mapping, both of which are acquired from either aircraft or satellites. For the 2D data processing, our proposed unsupervised burned area detection (UsBA detection) model enhances burned area mapping accuracy by integrating Ultra-High Resolution (UHR) aerial imagery with bi-temporal medium-resolution PlanetScope imagery, using a Segment Anything Model (SAM)-assisted UNetFormer (pre-trained on the target-style public dataset – LoveDA Rural) for refinement. The model demonstrates superior burned area segmentation, evidenced by improved evaluation metrics calculated from labeled test sites. For the 3D analysis, the burned areas extracted from 2D processing are further assessed using pre- and post-event airborne laser data. We implement a voxel-based workflow, including necessary steps such as ground filtering through Superpoints in RANSAC Planes (SiRP) method and biomass change analysis. The results indicate that the 3D branch provides a reliable lower bound of the actual damage map, because the vegetation growth between two measurements remains, in essence, undetected. The proposed framework offers a more accurate and robust solution for burn severity mapping utilizing combined 2D and 3D data, evaluated on a multi-source dataset from a real bushfire event that occurred in Bushland Park, South Australia.

The United States (US) has a wildland fire resources shortage. During peak fire season, the US struggles to fill fire leadership and support positions to manage the 68,000 wildland fires that occur on an annual basis. This wildland fire resource shortage happens on a yearly basis; with more large fires burning out of control and because of the competition for fire resources, many crucial positions go unfilled. During this presentation I will describe the problems that have driven the shortage of fire resources: older resource pool reaching retirement age, fewer government employees available to respond to fires, and the length of time to complete incident qualifications for positions. In the past, the US fire agencies have only deployed resources that are qualified under the National Wildfire Coordination Group (NWCG) Qualifications, which limits response of qualified resources outside the NWCG, specifically those positions that have a shortage problem, such as logistics, finance, and planning.This year FEMA released the National Qualification System (NQS) under the National Incident Management System. The purpose of NQS is to establish a minimum standard for qualification and certification for emergency responders positions across the US. The core tenets of NQS are minimum "discipline neutral" qualification requirements, performance standards, training, and qualification standards for each position for all incidents. NQS allows Endorsements which is a process in which the requirements of an NQS position are set to certain type of incident. An Endorsement is utilized for positions that require specialized, incident-specific experience in the type of incident to which an individual might respond. For example, an Operations Section Chief must have experience and training in wildland fire to get a Wildfire (WF) Endorsement. With limited positions having Endorsements, all other NQS positions are qualified to support all types of incidents. Participants will hear about NWGC's procedure to establish WF endorsements and the process to utilize discipline neutral positions. The WF endorsement process will create an organized system to bring in other non-fire resources to support wildland fire incidents which, in turn, will open up more resources to respond to wildfires and mitigate the wildland fire resource shortage