A novel, post-Soviet fire disturbance regime drives bird diversity and abundance on the Eurasian steppe.

Many grassland ecosystems and their associated biodiversity depend on the interactions between fire and land-use, both of which are shaped by socioeconomic conditions. The Eurasian steppe biome, much of it situated in Kazakhstan, contains 10% of the world's remaining grasslands. The break-up of the Soviet Union in 1991, widespread land abandonment and massive declines in wild and domestic ungulates led to biomass accumulation over millions of hectares. This rapid fuel increase made the steppes a global fire hotspot, with major changes in vegetation structure. Yet, the response of steppe biodiversity to these changes remains unexplored. We utilized a unique bird abundance dataset covering the entire Kazakh steppe and semi-desert regions together with the MODIS burned area product. We modeled the response of bird species richness and abundance as a function of fire disturbance variables-fire extent, cumulative burned area, fire frequency-at varying grazing intensity. Bird species richness was impacted negatively by large fire extent, cumulative burned area, and high fire frequency in moderately grazed and ungrazed steppe. Similarly, overall bird abundance was impacted negatively by large fire extent, cumulative burned area and higher fire frequency in the moderately grazed steppe, ungrazed steppe, and ungrazed semi-deserts. At the species level, the effect of high fire disturbance was negative for more species than positive. There were considerable fire legacy effects, detectable for at least 8 years. We conclude that the increase in fire disturbance across the post-Soviet Eurasian steppe has led to strong declines in bird abundance and pronounced changes in community assembly. To gain back control over wildfires and prevent further biodiversity loss, restoration of wild herbivore populations and traditional domestic ungulate grazing systems seems much needed.

Deep learning-based burned forest areas mapping via Sentinel-2 imagery: a comparative study.

In order to evaluate the effects of forest fires on the dynamics of the function and structure of ecosystems, it is necessary to determine burned forest areas with high accuracy, effectively, economically, and practically using satellite images. Extraction of burned forest areas utilizing high-resolution satellite images and image classification algorithms and assessing the successfulness of varied classification algorithms has become a prominent research field. This study aims to indicate on the capability of the deep learning-based Stacked Autoencoders method for the burned forest areas mapping from Sentinel-2 satellite images. The Stacked Autoencoders, used in this study as an unsupervised learning method, were compared qualitatively and quantitatively with frequently used supervised learning algorithms (k-Nearest Neighbors (k-NN), Subspaced k-NN, Support Vector Machines, Random Forest, Bagged Decision Tree, Naive Bayes, Linear Discriminant Analysis) on two distinct burnt forest zones. By selecting burned forest zones with contrasting structural characteristics from one another, an objective assessment was achieved. Manually digitized burned areas from Sentinel-2 satellite images were utilized for accuracy assessment. For comparison, different classification performance and quality metrics (Overall Accuracy, Mean Squared Error, Correlation Coefficient, Structural Similarity Index Measure, Peak Signal-to-Noise Ratio, Universal Image Quality Index, and KAPPA metrics) were used. In addition, whether the Stacked Autoencoders method produces consistent results was examined through boxplots. In terms of both quantitative and qualitative analysis, the Stacked Autoencoders method showed the highest accuracy values.

Greenhouse gas emissions from agricultural residue burning have increased by 75 % since 2011 across India.

Quantifying crop residue burning across India is imperative, owing to its adverse impacts on public health, the environment, and agricultural productivity. Specific information about the extent and characteristics of agricultural crop burning can verify the emission potential of agricultural systems and thereby facilitate targeted dissemination of agricultural innovations and support policymakers in mitigating the harmful effects. With a focus on district-level burning estimates, our study provides a comprehensive seasonal analysis of agricultural burning in India, including burned area, dry matter burned, and gaseous emissions for seven major crops from 2011 to 2020. To quantify the actual residues burned, we developed a remote sensing-based approach that incorporates the monitoring of agricultural burned area to quantify the actual residues burned. Including this satellite measure of the burned area greatly improves emissions estimates and minimizes error compared to typical approaches, which instead use an assumed fraction of total residues that are burned for each crop type. We estimated that emissions have increased by approximately 75 % for CO and Greenhouse gasses - CO2, CH4 and N2O - from 2011 to 2020. Total CO2e emissions increased from ~19,340 Gg.yr-1 in 2011 to ~33,834 Gg.yr-1 in 2020. Most emissions occurred during end of the Kharif season, followed by Rabi, caused by the burning of rice and wheat residues. Among the Indian states, Punjab has the highest burning activity, with 27 % (2.0 million hectares) of its total cultivated area burned in 2020. Interestingly, Madhya Pradesh has emerged as the second-largest contributor, accounting for 30 % of the total burned area across India in 2020. Our study demonstrates how satellite data can be used to map agricultural residue burning at scale, and this information can provide crucial insights for policy framing, targeting, and interventions to manage agricultural residues without compromising air quality and climate.

Agricultural expansion dominates rapid increases in cropland fires in Asia.

Asia contains 58 % of the global population and approximately 39 % of the world's cropland, making evaluation of the spatiotemporal variability patterns in cropland fire critical for understanding the interplay between crop residue burning and human activity in Asia. Although agricultural expansion and intensification have contributed to an overall decline in vegetation fires worldwide since the late 1990s, burned area by cropland fires in Asia has expanded by more than 19 % over the past two decades. India accounts for about 32 % of cropland fires in Asia, and the burned area has increased by more than two-thirds since 2003, particularly increased around 80 % during the two major cereal harvest seasons of March-May and October-November. In comparison, cropland fires have increased by <6 % in China since 2003, and there has been a marked downward trend in burned area in June due to the intensive implementation of the nationwide bans on open-land crop straw burning. The expansion of agricultural harvests is primarily responsible for the rapid increase in cropland fires in Asia, notably in India, where agricultural intensification is occurring with population growth and economic development in recent years, and crop straw burning should be strictly controlled in the future.

Characterization of global fire activity and its spatiotemporal patterns for different land cover types from 2001 to 2020.

Fire is a widespread phenomenon that plays an important role in Earth's ecosystems. This study investigated the global spatiotemporal patterns of burned areas, daytime and nighttime fire counts, and fire radiative power (FRP) from 2001 to 2020. The month with the largest burned area, daytime fire count, and FRP presented a bimodal distribution worldwide, with dual peaks in early spring (April) and summer (July and August), while the month with the largest nighttime fire count and FRP showed a unimodal distribution, with a peak in July. Although the burned area showed decline at the global scale, a significant increase occurred in temperate and boreal forest regions, where nighttime fire occurrence and intensity have consistently increased in recent years. The relationships among burned area, fire count, and FRP were further quantified in 12 typical fire-prone regions. The burned area and fire count exhibited a humped relationship with FRP in most tropical regions, whereas the burned area and fire count constantly increased when the FRP was below approximately 220 MW in temperate and boreal forest regions. Meanwhile, the burned area and FRP generally increased with the fire count in most fire-prone regions, indicating an increased risk of more intense and larger fires as the fire count increased. The spatiotemporal dynamics of burned areas for different land cover types were also explored in this study. The results suggest that the burned areas in forest, grassland, and cropland showed dual peaks in April and from July to September while the burned areas in shrubland, bareland, and wetlands usually peaked in July or August. Significant increases in forest burned area were observed in temperate and boreal forest regions, especially in the western U.S. and Siberia, whereas significant increases in cropland burned area were found in India and northeastern China.

Efficacy of hydrosurgical excision combined with skin grafting in the treatment of deep partial-thickness and full-thickness burns: A two-year retrospective study.

INTRODUCTION: Deep partial-thickness and full-thickness burn wounds often undergo tangential excision or escharectomy to expose healthy tissue, combined with skin grafting to promote wound healing. However, conventional tangential excision with the humby knife leads to inevitable damage to the dermis while excising burn tissue due to the lack of precision. Indeed, the preservation of dermal tissue is a key factor in determining wound healing and scar quality. The precision and tissue selectivity of the Versajet Hydrosurgical System has been established for excising burn tissue while preserving dermal tissue. In this study, we retrospectively compared the efficacy of "Hydrosurgical excision combined with skin grafting" and "Conventional tangential excision combined with skin grafting" in treating deep partial-thickness and full-thickness burn wounds to demonstrate that hydrosurgery improved the treatment of deep partial-thickness and full-thickness burns. METHODS: A total of 86 patients with deep partial-thickness and/or full-thickness burns with a total burn surface area (TBSA) ≤ 25% from July 2018 to July 2020 were included in this study and were divided into experimental (hydrosurgical excision combined with skin grafting, n = 43) and control (conventional tangential excision combined with skin grafting, n = 43) groups. Parameters were analyzed, including the intraoperative blood loss volume per unit area of grafted skin, surgery duration, wound healing time, skin graft survival, and the treatment costs per unit of burned area. Scar assessment was performed at 1 year with the modified Vancouver Scar Scale linked with TBSA (mVSS-TBSA). RESULT: No significant difference was found in male to female ratio, age, weight, TBSA, burn depth, skin grafting area (SKA), skin grafting methods, cases treated with carbon dioxide fractional laser or incidence of inhalation injury, and the incidence of hypovolemic shock between two groups(p > 0.05). Compared with the control group, patients treated with hydrosurgical excision combined with skin grafting experienced less intraoperative blood loss volume per unit area of grafted skin (p < 0.05). The mVSS-TBSA of patients that underwent hydrosurgical excision combined with skin grafting was significantly improved in comparison to the control group (p < 0.01). No significant difference was found in surgery duration, wound healing time, skin graft survival and treatment costs per unit of burned area between the two groups (p > 0.05). CONCLUSION: Hydrosurgical excision combined with skin grafting reduced intraoperative blood loss volume per unit area of grafted skin, improved scarring 1-year after injury, and did not increase the treatment costs per unit of burned area. This technique provides a novel alternative for managing deep partial-thickness and full-thickness burn wounds.

Drivers and implications of the extreme 2022 wildfire season in Southwest Europe.

Wildfire is a common phenomenon in Mediterranean countries but the 2022 fire season has been extreme in southwest Europe (Portugal, Spain and France). Here we provide a preliminary but comprehensive analysis of 2022's wildfire season in southwest Europe. Burned area has exceeded the 2001-2021 median by a factor of 52 in some regions and large wildfires (>500 ha) started to occur in June-July, earlier than the traditional fire season. These anomalies were associated with record-breaking values of fuel dryness, atmospheric water demand and pyrometeorological conditions. Live fuel moisture content was below the historical minima for almost 50 % of the season in some regions. A few large wildfires were responsible for 82 % of the burned area and, in turn, 47 % of the area burned occurred in protected areas. Shrublands, transitional woodlands and conifer forests (but not eucalypt plantations) were the land cover types most affected by extreme fires. As climate change intensifies, we can expect such fire seasons to become the new normal in large parts of the continent, potentially leading to major negative impacts on rural economies. These results highlight the need for landscape level fuel management also in protected areas, to avoid fire-induced biodiversity losses and landscape scale degradation. Our results have important policy implications and indicate that fire prevention should be explicitly addressed within continental forest legislation and strategies.

North-south antiphase of wildfire activity across the pyroregions of continental China driven by NAO and the Antarctic oscillation.

Wildfires are a natural disturbance in many parts of the world, but fire regimes are changing as a result of anthropogenic pressures. A key uncertainty towards anticipating future changes in burned area lies in understanding the effects of climate teleconnections (CTs). Here we test how different CTs impact burned area in China, a large country comprising different biomes and where similar fire-suppression and post-fire afforestation policies are implemented. We observed diverging temporal trends in burned area across the different pyroregions of China, from increases in the Northeastern grasslands and mixed forests pyroregion to decreases in the Southern tropical forests pyroregion. This North-South antiphase in fire activity was being partly driven by joint effects of the North Atlantic Oscillation and the Antarctic Oscillation, which exerted contrasting effects on fire weather across latitude. El Niño Southern Oscillation and the other examined teleconnections had minor effects over burned area. The increasing burned area in the NE-mixed forests pyroregion indicates that mega-fires may increase under global warming but their occurrence may be modulated by potential strengthening or weakening of NAO and AAO.

Wildfire impacts on surface water quality parameters: Cause of data variability and reporting needs.

Surface runoff mobilizes the burned residues and ashes produced during wildfires and deposits them in surface waters, thereby deteriorating water quality. A lack of a consistent reporting protocol precludes a quantitative understanding of how and to what extent wildfire may affect the water quality of surface waters. This study aims to analyze reported pre- and post-fire water quality data to inform the data reporting and highlight research opportunities. A comparison of the pre-and post-fire water quality data from 44 studies reveals that wildfire could increase the concentration of many pollutants by two orders of magnitude. However, the concentration increase is sensitive to when the sample was taken after the wildfire, the wildfire burned area, discharge rate in the surface water bodies where samples were collected, and pollutant type. Increases in burned areas disproportionally increased total suspended solids (TSS) concentration, indicating TSS concentration is dependent on the source area. Increases in surface water flow up to 10 m3 s-1 increased TSS concentration but any further increase in flow rate decreased TSS concentration, potentially due to dilution. Nutrients and suspended solids concentrations increase within a year after the wildfire, whereas peaks for heavy metals occur after 1-2 years of wildfire, indicating a delay in the leaching of heavy metals compared to nutrients from wildfire-affected areas. The concentration of polycyclic aromatic hydrocarbons (PAHs) was greatest within a year post-fire but did not exceed the surface water quality limits. The analysis also revealed inconsistency in the existing sampling protocols and provides a guideline for a modified protocol along with highlighting new research opportunities. Overall, this study underlines the need for consistent reporting of post-fire water quality data along with environmental factors that could affect the data so that the post-fire water quality can be assessed or compared between studies.

Climate and socioeconomic drivers of biomass burning and carbon emissions from fires in tropical dry forests: A Pantropical analysis.

Global burned area has declined by nearly one quarter between 1998 and 2015. Drylands contain a large proportion of these global fires but there are important differences within the drylands, for example, savannas and tropical dry forests (TDF). Savannas, a biome fire-prone and fire-adapted, have reduced the burned area, while the fire in the TDF is one of the most critical factors impacting biodiversity and carbon emissions. Moreover, under climate change scenarios TDF is expected to increase its current extent and raise the risk of fires. Despite regional and global scale effects, and the influence of this ecosystem on the global carbon cycle, little effort has been dedicated to studying the influence of climate (seasonality and extreme events) and socioeconomic conditions of fire regimen in TDF. Here we use the Global Fire Emissions Database and, climate and socioeconomic metrics to better understand long-term factors explaining the variation in burned area and biomass in TDF at Pantropical scale. On average, fires affected 1.4% of the total TDF' area (60,208 km2 ) and burned 24.4% (259.6 Tg) of the global burned biomass annually at Pantropical scales. Climate modulators largely influence local and regional fire regimes. Inter-annual variation in fire regime is shaped by El Niño and La Niña. During the El Niño and the forthcoming year of La Niña, there is an increment in extension (35.2% and 10.3%) and carbon emissions (42.9% and 10.6%). Socioeconomic indicators such as land-management and population were modulators of the size of both, burned area and carbon emissions. Moreover, fires may reduce the capability to reach the target of "half protected species" in the globe, that is, high-severity fires are recorded in ecoregions classified as nature could reach half protected. These observations may contribute to improving fire-management.

El área global quemada se redujo en casi una cuarta parte entre 1998 y 2015. Los bosques secos contienen una gran proporción de esos incendios globales, pero existen diferencias importantes dentro de ellos, por ejemplo, las sabanas y los bosques secos tropicales (SBC). Las sabanas, son un bioma propenso y adaptado al fuego, y que en los últimos años han reducido su área quemada. Mientras que el fuego en la SBC es uno de los factores más críticos que impactan la biodiversidad y las emisiones de carbono. Además, bajo escenarios de cambio climático, se espera que la SBC aumente su extensión actual y aumente el riesgo de incendios. A pesar de los efectos a escala regional y global, y la influencia de este ecosistema en el ciclo global del carbono, se le ha dedicado poco esfuerzo a estudiar la influencia del clima (estacionalidad y eventos extremos) y las condiciones socioeconómicas del régimen de incendios. Aquí usamos la base de datos global de emisiones de incendios y métricas climáticas y socioeconómicas para comprender mejor los factores a largo plazo que explican la variación en el área quemada y la biomasa a escala Pantropical. En promedio, los incendios afectaron el 1,4% del área total de la SBC (60 208 km2 ) y quemaron el 24,4% (259,6 Tg) de la biomasa global quemada anualmente a escala Pantropical. Los moduladores climáticos influyen en gran medida en los regímenes de incendios locales y regionales. La variación interanual del régimen de incendios está determinada por El Niño y La Niña. Durante El Niño y el año subsecuente de La Niña, se produce un incremento en la extensión (35,2% y 10,3%) y en las emisiones de carbono (42,9% y 10,6%). Los indicadores socioeconómicos como la gestión de la tierra y la población fueron moduladores del tamaño tanto del área quemada como de las emisiones de carbono. Además, los incendios pueden reducir la capacidad de alcanzar el objetivo de "protección de la mitad de las especies" en el mundo, es decir, los incendios de alta gravedad se registran en ecorregiones clasificadas como naturaleza que podría alcanzar la protección de la mitad de su biodiversidad. Estas observaciones pueden contribuir a mejorar la gestión de incendios.

Heatwaves and fire in Pantanal: Historical and future perspectives from CORDEX-CORE.

The Pantanal biome, at the confluence of Brazil, Bolivia and Paraguay, is the largest continental wetland on the planet and an invaluable reserve of biodiversity. The exceptional 2020 fire season in Pantanal drew particular attention due to the severe wildfires and the catastrophic natural and socio-economic impacts witnessed within the biome. So far, little progress has been made in order to better understand the influence of climate extremes on fire occurrence in Pantanal. Here, we evaluate how extreme hot conditions, through heatwave events, are related to the occurrence and the exacerbation of fires in this region. A historical analysis using a statistical regression model found that heatwaves during the dry season explained 82% of the interannual variability of burned area during the fire season. In a future perspective, an ensemble of CORDEX-CORE simulations assuming different Representative Concentration Pathways (RCP2.6 and RCP8.5), reveal a significant increasing trend in heatwave occurrence over Pantanal. Compared to historical levels, the RCP2.6 scenario leads to more than a doubling in the Pantanal heatwave incidence during the dry season by the second half of the 21st century, followed by a plateauing. Alternatively, RCP8.5 projects a steady increase of heatwave incidence until the end of the century, pointing to a very severe scenario in which heatwave conditions would be observed nearly over all the Pantanal area and during practically all the days of the dry season. Accordingly, favorable conditions for fire spread and consequent large burned areas are expected to occur more often in the future, posing a dramatic short-term threat to the ecosystem if no preservation action is undertaken.

Seasonal Variations of Solar-Induced Fluorescence, Precipitation, and Carbon Dioxide Over the Amazon.

Previous studies suggested that the Amazon, the largest rainforest on Earth, changes from a CO2 sink to a CO2 source during the dry/fire season. However, the biospheric contributions to atmospheric CO2 are not well understood during the two main seasons, the dry/fire season and the wet season. In this article, we utilize Orbiting Carbon Observatory 2 (OCO-2) Solar-Induced Fluorescence (SIF) to explore photosynthetic activity during the different seasons. The spatiotemporal variability of OCO-2 SIF, OCO-2 CO2, precipitation, and burned area are investigated over the Amazon from September 2014 to December 2019. Averaging over the entire Amazon region, we found a positive temporal correlation (0.94) between OCO-2 SIF and Global Precipitation Climatology Project precipitation and a negative temporal correlation (-0.64) between OCO-2 SIF and OCO-2 CO2, consistent with the fact that precipitation enhances photosynthesis, which results in higher values for SIF and rate of removal of CO2 from the atmosphere above the Amazon region. We also observed seasonality in the spatial variability of these variables within the Amazon region. During the dry/fire (August-October) season, low SIF values, low precipitation, high vapor pressure deficit (VPD), large burned areas, and high atmospheric CO2 are mainly found over the southern Amazon region. In contrast, during the wet season (January-March), high SIF values, high precipitation, low VPD, smaller burned areas, and low CO2 are found over both the central and southern Amazon regions. The seasonal difference in SIF suggests that photosynthetic activity is reduced during the dry/fire season relative to the wet season as a result of low precipitation and high VPD, especially over the southern Amazon region, which will contribute to more CO2 in the atmosphere during the dry/fire season.

Building a small fire database for Sub-Saharan Africa from Sentinel-2 high-resolution images.

Coarse resolution sensors are not very sensitive at detecting small fire patches, making current estimations of global burned areas (BA) very conservative. Using medium or high-resolution sensors to generate BA products becomes then a priority, particularly in areas where fires tend to be small and frequent. Building on previous work that developed a small fire dataset (SFD) for Sub-Saharan Africa for 2016, this paper presents a new version of the dataset for 2019 using the two Sentinel-2 satellites (A and B) and VIIRS active fires. Total estimated BA was 4.8 Mkm2. This value was much higher than estimations from two global, coarser-spatial resolution BA products based on MODIS data for the same area and period: 80 % greater than estimates from FireCCI51 (based on MODIS 250 m bands) and 120 % larger than MCD64A1 (based on MODIS 500 m bands). The main differences were observed in those months with higher fire occurrence (November to January for the Northern Hemisphere regions and June to September for the Southern Hemisphere ones). Accuracy assessment of the SFD product was based on a novel sampling strategy designed to obtain independent fire reference perimeters. Validation results showed remarkable high accuracy values comparing to existing global BA products. Overall omission errors (OE) were estimated as 8.5 %, commission errors (CE) as 15.0 %, with a Dice Coefficient of 87.7 %. All of these estimations implied significant improvements over the global, coarser spatial resolution BA products (OE > 50 % and CE > 20 % for the same area and period), as well as over the previous SFD product for 2016 of the same area, generated from a single Sentinel-2 satellite and MODIS active fires (OE = 26.5 % and CE = 19.3 %). Temporal accuracies greatly increased as well with the new product, with 92.5 % of fires detected within the first 10 days of occurrence.

Agricultural burned area detection using an integrated approach utilizing multi spectral instrument based fire and vegetation indices from Sentinel-2 satellite.

This study presents a methodology that focuses on detecting agricultural burned areas using Sentinel-2 multispectral data at 10 m. We developed a simple, locally adapted, straightforward approach of multi-index threshold to extract post-winter agricultural burned areas at high resolution for 2019-21. Further, we design a new method for virtual sample collection using already validated fire location data and visual interpretation conditioned using strict selection criteria to improve sample accuracy. Sampling accuracy showed near-perfect agreement with an average Cohen's Kappa value of 0.98. We retrieved monthly ABAs at a resolution of 10 m, and these products were validated against reference burned sample plots identified using visual interpretation of Planet (3m) satellite data. Overall, we found that our method performed well, with an F1 score of 83.63% and low commission (20%) and omission (7%) errors. When compared to global burnt area products, validation accuracy demonstrated an exceptional subpixel scale detecting capability. The study also addresses the complexity of residue burnings and burn signatures' volatile nature by performing multilevel masking and temporal corrections.•A novel remotely sensed data aided virtual sampling approach to acquire burned and unburned samples.•An integrated method to extract smallholder agricultural burned area using Sentinel-2 multispectral data at a high resolution of 10 m.

Dynamic prediction of global monthly burned area with hybrid deep neural networks.

Wildfires not only severely damage the natural environment and global ecological balance but also cause substantial losses to global forest resources and human lives and property. Unprecedented fire events such as Australia's bushfires have alerted us to the fact that wildfire prediction is a critical scientific problem for fire management. Therefore, robust, long-lead models and dynamic predictions of wildfire are valuable for global fire prevention. However, despite decades of effort, the dynamic, effective, and accurate prediction of wildfire remains problematic. There is great uncertainty in predicting the future based on historical and existing spatiotemporal sequence data, but with advances in deep learning algorithms, solutions to prediction problems are being developed. Here, we present a dynamic prediction model of global burned area of wildfire employing a deep neural network (DNN) approach that produces effective wildfire forecasts based on historical time series predictors and satellite-based burned area products. A hybrid DNN that combines long short-term memory and a two-dimensional convolutional neural network (CNN2D-LSTM) was proposed, and CNN2D-LSTM model candidates with four different architectures were designed and compared to construct the optimal architecture for fire prediction. The proposed model was also shown to outperform convolutional neural networks (CNNs) and the fully connected long short-term memory (FcLSTM) approach using the refined index of agreement and evaluation metrics. We produced monthly global burned area spatiotemporal prediction maps and adequately reflected the seasonal peak in fire activity and highly fire-prone areas. Our combined CNN2D-LSTM approach can effectively predict the global burned area of wildfires 1 month in advance and can be generalized to provide seasonal estimates of global fire risk.

Lightning patterns in the Pantanal: Untangling natural and anthropogenic-induced wildfires.

The identification of fire causes and characteristics is of fundamental importance to better understand fire regimes and drivers. Particularly for Brazil, there is a gap in the quantification of lightning-caused fires. Accordingly, this work is a novel probabilistic assessment of the spatial-temporal patterns of lightning-ignited wildfires in the Pantanal wetland. Here, remote sensing information such as VIIRS active fires, MODIS burned area (BA) and STARNET lightning observations from 2012 to 2017, were combined to estimate the location, number of scars and amount of BA associated with atmospheric discharges on a seasonal basis. The highest lightning activity occurs during summer (December-February), and the lowest during winter (June-August). Conversely, the highest fire activity occurred during spring (September-November) and the lowest during autumn (March-May). Our analysis revealed low evidence of an association between fires and lightning, suggesting that human-related activities are the main source of ignitions. Weak evidence of natural-caused fire occurrence is conveyed by the low spatial-temporal match of lightning and fire throughout the studied period. Natural-caused fires accounted for only 5% of the annual total scars and 83.8% of the BA was human-caused. Most of the fires with extension larger than 1000 ha were not related to lighting. Lightning-fires seem an important element of the summer fire regime given that around half of the total BA during this season may be originated by lightning. By contrast, in the rest of the year the lightning-fires represent a minor percentage of the fire activity in the region. The density of lightning-ignited fires varies considerably, being higher in the north part of the Pantanal. This work provides a basis for a better understanding of lightning-related fire outbreaks in tropical ecosystems, particularly wetlands, which is fundamental to improve region-based strategies for land management actions, ecological studies and modeling climatic and anthropogenic drivers of wildfires.

Putting fire on the map of Brazilian savanna ecoregions.

The Brazilian savanna (Cerrado) is considered the most floristically diverse savanna in the world, home to more than seven thousand species. The region is a mosaic of savannas, grasslands and forests whose unique biophysical and landscape attributes are on the basis of a recent ecoregional map, paving the way to improved region-based strategies for land management actions. However, as a fire-prone ecosystem, Cerrado owes much of its distribution and ecological properties to the fire regime and contributes to an important parcel of South America burned area. Accordingly, any attempt to use ecoregion geography as a guide for management strategies should take fire into account, as an essential variable. The main aim of this study is to complement the ecoregional map of the Cerrado with information related to the fire component. Using remotely sensed information, we identify patterns and trends of fire frequency, intensity, seasonality, extent and scar size, and combine this information for each ecoregion, relying on a simple classification that summarizes the main fire characteristics over the last two decades. Results show a marked north-south fire activity gradient, with increased contributions from MATOPIBA, the latest agricultural frontier. Five ecoregions alone account for two thirds of yearly burned area. More intense fires are found in the Arc of Deforestation and eastern ecoregions, while ecoregions in MATOPIBA display decreasing fire intensity. An innovative analysis of fire scars stratified by size class shows that infrequent large fires are responsible for the majority of burned area. These large fires display positive trends over many ecoregions, whereas smaller fires, albeit more frequent, have been decreasing in number. The final fire classification scheme shows well defined spatially-aggregated groups, where trends are found to be the key factor to evaluate fire within their regional contexts. Results presented here provide new insights to improve fire management strategies under a changing climate.

Machine Learning Methods and Synthetic Data Generation to Predict Large Wildfires.

Wildfires are becoming more frequent in different parts of the globe, and the ability to predict when and where they will occur is a complex process. Identifying wildfire events with high probability of becoming a large wildfire is an important task for supporting initial attack planning. Different methods, including those that are physics-based, statistical, and based on machine learning (ML) are used in wildfire analysis. Among the whole, those based on machine learning are relatively novel. In addition, because the number of wildfires is much greater than the number of large wildfires, the dataset to be used in a ML model is imbalanced, resulting in overfitting or underfitting the results. In this manuscript, we propose to generate synthetic data from variables of interest together with ML models for the prediction of large wildfires. Specifically, five synthetic data generation methods have been evaluated, and their results are analyzed with four ML methods. The results yield an improvement in the prediction power when synthetic data are used, offering a new method to be taken into account in Decision Support Systems (DSS) when managing wildfires.

Human and climate drivers of global biomass burning variability.

Biomass burning is one of the most critical factors impacting vegetation and atmospheric trends, with important societal implications, particularly when extreme weather conditions occur. Trends and factors of burned area (BA) have been analysed at regional and global scales, but little effort has been dedicated to study the interannual variability. This paper aimed to better understand factors explaining this variation, under the assumption that the more human control of fires the more frequently they occur, as burnings will be less dependent of weather cycles. Interannual variability of BA was estimated from the coefficient of variation of the annual BA (BA_CV) estimated from satellite data at 250 m, covering the period from 2001 to 2018. These data and the explanatory variables were resampled at 0.25-degree resolution for global analysis. Relations between this variable and explanatory factors, including human and climate drivers, were estimated using Random Forest (RF) and generalized additive models (GAM). BA_CV was negatively related to BA_Mean, implying that areas with higher average BA have lower variability as well. Interannual BA variability decreased when maximum temperature (TMAX) and actual and potential evapotranspiration (AET, PET) increased, cropland and livestock density increased and the human development index (HDI) values decreased. GAM models indicated interesting links with AET, PET and precipitation, with negative relation with BA_CV for the lower ranges and positive for the higher ones, the former indicating fuel limitations of fire activity, and the latter climate constrains. For the global RF model, TMAX, AET and HDI were the main drivers of interannual variability. As originally hypothesised, BA_CV was more dependent on human factors (HDI) in those areas with medium to large BA occurrence, particularly in tropical Africa and Central Asia, while climatic factors were more important in boreal regions, but also in the tropical regions of Australia and South America.

The role of fire in global forest loss dynamics.

Fires, among other forms of natural and anthropogenic disturbance, play a central role in regulating the location, composition and biomass of forests. Understanding the role of fire in global forest loss is crucial in constraining land-use change emissions and the global carbon cycle. We analysed the relationship between forest loss and fire at 500 m resolution based on satellite-derived data for the 2003-2018 period. Satellite fire data included burned area and active fire detections, to best account for large and small fires, respectively. We found that, on average, 38 ± 9% (± range) of global forest loss was associated with fire, and this fraction remained relatively stable throughout the study period. However, the fraction of fire-related forest loss varied substantially on a regional basis, and showed statistically significant trends in key tropical forest areas. Decreases in the fraction of fire-related forest loss were found where deforestation peaked early in our study period, including the Amazon and Indonesia while increases were found for tropical forests in Africa. The inclusion of active fire detections accounted for 41%, on average, of the total fire-related forest loss, with larger contributions in small clearings in interior tropical forests and human-dominated landscapes. Comparison to higher-resolution fire data with resolutions of 375 and 20 m indicated that commission errors due to coarse resolution fire data largely balanced out omission errors due to missed small fire detections for regional to continental-scale estimates of fire-related forest loss. Besides an improved understanding of forest dynamics, these findings may help to refine and separate fire-related and non-fire-related land-use change emissions in forested ecosystems.