Fire remote sensing monitoring
Traditional fire monitoring relies on ground manual and aircraft, with single means, limited patrol area and narrow vision. Artificial ground patrol will be limited by terrain, landform and other factors, can not cover a large area, the monitoring range is limited. Although aircraft inspection can cover a large area, it requires a lot of resources and time, and the cost is high. At the same time, the traditional monitoring time is long, the monitoring period is long, limited by time and space, can not realize the timely response to the fire, easy to cause the fire expansion. Remote sensing satellites can cover a large area and realize the comprehensive monitoring of fire without the limitation of terrain and landform. Real-time data can be obtained, the monitoring frequency is high, and the change of fire point and fire situation can be found in time. Combined with UAV, camera, radar and other technologies, real-time acquisition and analysis of forest fire fire point, fire, fire line and other information, the spread trend of fire, dangerous areas, etc., forecast and early warning. Xiaobian collated the fire remote sensing monitoring articles published in recent years for your appreciation!
Theme Keywords： remote sensingforest fireLandsattransfer learningtopographic featurestime seriesswidden agriculturespatiotemporal predictionsegmentation thresholdpre-judgment of hidden dangers
- The Paper
- Abstract：Volcano monitoring is essential for predicting volcanic eruptions and implementing early warning measures. Traditional ground-based monitoring methods cannot fully cover all volcanoes. Satellite remote sensing technology, with its advantages of global coverage and high temporal and spatial resolutions, is an important complement for near-real-time monitoring of volcanic activities, especially for the detection of lava flows and volcanic thermal anomalies.This study presents the current status of typical sensors used for infrared remote sensing of volcanic hotspots and summarizes the methodology for detecting volcanic hotspots by using satellite infrared data. First, the history of thermal infrared satellite data monitoring and satellite system development is summarized. Notably, various types of algorithms and satellite systems have been applied to make the monitoring of volcanic activities at the global scale efficient and accurate. Second, the development of volcanic hotspot identification algorithms is analyzed, and existing volcanic hotspot identification algorithms are classified into four categories in accordance with the different characteristics of the volcano used and its surrounding features (spatial/temporal). The four algorithm categories are spatial feature, temporal feature, comprehensive feature, and artificial intelligence algorithms. The spatial feature algorithms are categorized into fixed and dynamic threshold methods on the basis of different methods of threshold selection (fixed/dynamic threshold). On the basis of the classification above, we describe the current status of the volcanic hotspot identification algorithms and summarize their data, scope of application, and application limitations to provide a comprehensive classification and assessment for understanding and improving volcano hotspot detection technology. Such classification and assessment are crucial for the development of future volcano thermal remote sensing theories and technologies.Subsequent research should improve the adaptability of the algorithms to different volcanic environments, combine the advantages of traditional algorithms and artificial intelligence, and utilize historical data and time-series analyses to identify volcanic hotspots accurately. In addition, the fusion of high-resolution and multispectral satellite data can improve the spatial and spectral resolutions of volcanic activity monitoring, thus capturing the microfeatures of volcanoes accurately. These improvements will enhance the comprehensiveness and accuracy of volcanic hotspot monitoring and provide reliable support for the monitoring, early warning, and prevention of geologic hazards.Keywords：volcanic lava flows;thermal remote sensing;infrared satellite data;volcano monitoring;thermal anomalies;hotspot automatic detection;algorithm classification;disaster prevention and reduction1154|2547|0
<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2025-04-21 - Abstract：The utilization of remote sensing data for forest fire spread simulation is predicated on the acquisition of critical information such as combustibles, topography, and ignition points through remote sensing technology. These data, when integrated with forest fire spread models, are pivotal for predicting the trajectory of forest fires and serve as a vital reference for forest fire prevention and emergency response operations. This study harnesses two fire spread simulation tools, Cell2Fire and FARSITE, to simulate two forest fire events occurring in the Liangshan of Sichuan and the Naji Forestry Site in Inner Mongolia, comparing the precision of these simulators.Addressing the challenge of suboptimal accuracy in continuous simulations, the study proposes an innovative reinitialization method for burned areas. This method is based on the Burned Area Index (BAI) spectral index combined with region-growing approach. The study compares the simulation precision of FARSITE and Cell2Fire across three methodologies: continuous simulation, VIIRS reinitialization, and the proposed burned area reinitialization.In the Liangshan fire simulations, Cell2Fire projected a faster spread of fire compared with FARSITE, although FARSITE consistently outperformed Cell2Fire in simulation accuracy across all methods. In the simulations for the Naji Forestry Site, FARSITE’s VIIRS reinitialization and burned area reinitialization methods yielded higher precision than Cell2Fire, but its continuous simulation accuracy considerably diminished and underperformed relative to Cell2Fire. When comparing the accuracy of the three simulation methods, the SC values and confusion matrix indicators showed a general agreement, with the higher precision outcomes predominantly emerging from the burned area reinitialization method applied during the mid to later stages of the fires, with most SC values and F1 scores exceeding 0.8.The conclusions of this study are as follows (1) The simulation outcomes from FARSITE and Cell2Fire show minor discrepancies, with FARSITE demonstrating superior accuracy in most scenarios compared with Cell2Fire. (2) Utilizing the BAI and regional growth methodologies enables the automatic extraction of burned areas. This approach is not only simple and reliable in extracting results but also saves time and labor costs, making it more suitable for practical applications in forest fire spread simulations. (3) The reinitialization method for burned areas, which integrates automatically extracted burned areas with VIIRS fire spot data for forest fire spread simulations, remarkably enhances the mitigation of issues related to insufficient fire spots in the early stages of a fire and the accumulation of errors during prolonged simulations. The simulation outcomes utilizing this method exhibit the highest accuracy within the simulations conducted by Cell2Fire and FARSITE, suggesting that the selection of an appropriate simulation approach may be more crucial to improving the accuracy of simulation results than the choice of a forest fire spread simulator.Keywords：FARSITE;Cell2Fire;VIIRS;forest fire;fire spread simulation;“5·7” Liangshan fire;“5·17” Naji fire994|2131|1<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2026-01-05
- Abstract：In recent years, the problem of forest wildfires in Sichuan Province has emerged as a matter of great concern. These wildfires have occurred with alarming frequency, presenting a formidable threat not only to the local ecological security but also to the lives and property of the people and the courageous rescue workers who put themselves in harm’s way. This in-depth study is committed to conducting a comprehensive exploration of the temporal and spatial characteristics of forest and grassland fires in Sichuan Province over the extensive period from 2001 to 2021. The ultimate aim is to provide highly valuable and actionable information that can serve as a solid foundation for making well-informed decisions regarding fire prevention and control strategies. By understanding these characteristics, it is hoped that effective measures can be implemented to minimize the occurrence and impact of wildfires, thereby safeguarding the delicate balance of the ecosystem and protecting the well-being of the local population.This extensive research is firmly grounded in a diverse array of multi-source remote sensing fire products, such as MCD64A1, Fire_CCI51, and MCD14ML. Through meticulous extraction of effective fire points, a wealth of regional fire data is painstakingly obtained. Leveraging the power of a sophisticated geographic information system, the temporal trend and spatial distribution of forest and grassland fires are thoroughly examined. This involves analyzing patterns over time and identifying areas of concentration or dispersion. Additionally, mathematical statistics and an adaptable fuzzy neural network are skillfully employed to meticulously analyze the complex relationship between climatic, combustible, and topographical environmental factors and the occurrence of fires. By using these advanced techniques, researchers can gain a deeper understanding of the underlying causes and contributing factors of wildfires, enabling more targeted prevention and response efforts.Research findings reveal that from 2001 to 2014, both the frequency of fires and the area affected by them demonstrated an upward trend. This indicates a growing concern for fire management and prevention. Fires occurred with notable frequency from January to May, suggesting a seasonal pattern that can be used to inform preventive measures during these high-risk months. In terms of the spatial distribution of grassland fires, it exhibits a distinct heterogeneity, with a concentration mainly in the southwest region of Sichuan Province. This spatial pattern may be influenced by a combination of factors such as vegetation type, climate, and human activities. Intriguingly, in the northeast of China, grassland fires have witnessed a remarkable increase in recent times. This finding highlights the need for a broader understanding of fire dynamics on a national scale. In the correlation analysis of various influencing factors, a high degree of correlation is observed between forest fires and fuel water content. This suggests that changes in fuel moisture levels can have a significant impact on the likelihood and severity of forest fires. Environmental variables are clearly identified as the primary driving factors behind the temporal and spatial characteristics of forest fires. For grassland fires, although there is a strong correlation with meteorological factors, it is reasonably speculated that human factors also exert a substantial influence on the characteristics of grassland fires. This could include activities such as land use changes, agricultural practices, and accidental ignitions.Based on the detailed analysis of the temporal and spatial characteristics of fires in Sichuan Province, this study provides a solid and reliable decision-making basis for formulating forest and grassland fire prevention and control policies, establishing early warning systems, and enhancing monitoring efforts in this region. By understanding the patterns and drivers of wildfires, policymakers and fire management agencies can develop more effective strategies to protect the precious ecological environment and safeguard the lives and property of the people. This will undoubtedly contribute to a more sustainable future for the region, ensuring that the beauty and biodiversity of Sichuan’s forests and grasslands are preserved for generations to come.Keywords：remote sensing;Sichuan province;MCD64A1;Fire_CCI51;MCD14ML;forest grassland fire;spatial distribution;time trend;spatio-temporal characteristics1807|5932|1<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2024-12-30
- Abstract：In the past decades, remote sensing methods of forest fire monitoring were mainly ground patrol, visual interpretation of aerial images, and remote sensing satellite observation with low spatial and temporal resolution. Nowadays, mobile measurement backpack system, light and small UAV, image fusion technology with high spatial and temporal resolution, and near real-time data sharing platform are driving remote sensing into broader forest fire application scenarios. The spatiotemporal-spectral resolution of a single data source is difficult to improve simultaneously as restricted by satellite orbit, observation mode, and sensor performance. The monitoring results may also be constrained by environmental factors such as cloud and rain. This condition leads to reduced monitoring accuracy and inability to collect reliable and detailed fire data to meet the emergency needs of fire location and continuous monitoring. Determining the spatial and temporal characteristics of forest fires is important for disaster prevention and control. A large number of new-generation sub-meter satellite platforms and sensors are currently used, and intelligent remote sensing inversion methods are constantly optimized. With the support of these technologies, the current fire monitoring capability based on multi-source remote sensing methods has the advantages of low cost, near real-time performance, multi-scale, wide coverage, and high precision. The monitoring, analysis, and continuous tracking of forest fires with multi-source remote sensing data can provide effective prediction and evaluation for forest fires.In general, in pre-fire, based on traditional fire risk factors such as meteorological, topographical, and human factors, multi-source remote sensing data and inversion optimization algorithm of fuel parameters are used to provide more accurate three-dimensional characteristic information of vegetation, including fuel moisture content, canopy height, and forest biomass. In during-fire, the accuracy and timeliness of a single remote sensing data source need to be improved due to spatial and temporal heterogeneity of ground objects. Matching the spatial and temporal domains between polar-orbiting meteorological satellite fire detection results and the geostationary satellite fire intensity monitoring results can make up for the shortcomings of a single remote sensing data source and realize dynamic monitoring of forest fires with high spatial and temporal resolution. Satellite monitoring is limited by revisit cycles and dense cloud cover in some cases. This problem can be effectively solved by data complementation, fusion, or using airborne or ground platform monitoring. Fire intensity monitoring results can also be used as dynamic input data for biomass burning in atmospheric dispersion models, which provides the basis for fire emission dispersion simulation. In post-fire, optical, radar, and LiDAR data can be combined to improve the ability to gauge environmental changes caused by fire.For the rapid development of multi-source remote sensing technology, this study summarizes current fire risk assessment, fuel parameter inversion, fire detection, fire behavior analysis, burned area identification, fire intensity evaluation, and vegetation recovery monitoring. In general, future research is expected to be based on the synergy of multi-source remote sensing technologies. This synergy can be made through the optimization and integration of new remote sensing analysis methods to further understand fire patterns and improve fire monitoring ability.Keywords：multi-source remote sensing;forest fire;image fusion;fire risk assessment;fire detection;fire emissions;fire environment;fire damage2549|8392|1<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2026-02-02
- Abstract：Using remote-sensing technology to obtain information about burned areas is important for ecological environment monitoring. High-resolution data are more suitable for extracting small-scale burned areas. To develop the fire monitoring ability of domestic remote-sensing data and improve the extraction efficiency and accuracy of a small-scale burned area, two GF-1 WFV images (before and after fires) and multi temporal FY-3D MERSI fire products are used to extract burned areas for two study areas, respectively, located in the Tibetan Autonomous County of Muli and Xichang City, Sichuan Province.The reference true values are obtained by human-computer interaction for verification. The results of burned areas extracted by neural network classification are compared with the result of the proposed method. Our results show that the accuracy of burned areas detected by the proposed method is higher than that by neural network classification, and the Kappa coefficients in two study areas are 0.82 and 0.87, respectively. The regions of commission and omission are usually distributed at the edge of the burned area patch. The distribution of burned area in Xichang is more compact than that in Muli, so the accuracy of burned area mapping in Xichang is higher.The method can fully combine the advantages of the two kinds of data, reduce the uncertainty and time cost caused by sample selection, and extract the small-scale burned area quickly and accurately. Fully exploiting the temporal, spatial, and spectral characteristics of fire points and burned areas can compensate for the shortcomings of GF-1 WFV images in temporal and spectral resolution. Meanwhile, the method can fully combine the two kinds of data and minimize the impact of the difference of spatial resolution. In the future, the method can be improved using a higher accuracy of fire-point products. The accuracy of the reference true value of the burned area can be improved through field investigation.The method is primarily divided into two partsrough extraction and fine extraction. In rough extraction, according to the relationship between fire points and the formation of burned areas, the fire-point pixels are selected and expanded into the rough range of burned areas by combining temporal, spatial, and spectral characteristics. Temporal characteristic refers to fire points with concentrated occurrence time that easily form burned areas; spatial characteristic refers to fire points with concentrated location that easily form burned areas, and burned pixels are usually adjacent to fire-point pixels; spectral characteristics refer to pixels with higher NDVI difference before and after fire, which may be burned pixels. In fine extraction, the land-cover types included in the burned area are determined according to the number of fire-point pixels. The segmentation threshold is determined using the iterative-threshold method for each land-cover type. Burned pixels and unburned pixels in each land-cover type are classified using the segmentation threshold. The small patches are removed to obtain the result of burned-area extraction.Keywords：remote sensing;burned area;fire point product;FY-3D MERSI;GF-1 WFV;NDVI;segmentation threshold1559|1935|2<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2024-03-22
- Abstract：In recent years, the frequent occurrence of forest fires has considerably affected people’s normal work and life and the natural ecosystem. Fire hazard assessment is crucial to the prevention of forest fire and the allocation of fire resources. This study collects historical forest fire events in China from 2002 to 2020. The events are distributed in five climate regions in China, namely, plateau mountain, temperate continental, temperate monsoon, subtropical monsoon, and tropical monsoon climate regions. The meteorological factors, vegetation indices, and topographic factors in the different regions are integrated, and the random forest method is used to establish a comprehensive forest fire hazard assessment model. Fire influencing factors are calculated from different data products, fire events are selected using FIRMS images, meteorological factors are calculated using ERA5-land data, topographic factors are calculated using Shuttle Radar Topography Mission’s digital elevation model products, and vegetation indices are computed using the MODIS reflectance product MCD43A4. The fire hazard assessment model can predict the time series of fire hazards and evaluate the spatial distribution of these hazards. The fire occurrence location revealed by the test data differs from that from the training data. Test case results show that the accuracy of the established fire hazard assessment model is high, and the area under the receiver operating characteristic curve reaches 0.84, which produces good results in the time series prediction and spatial distribution assessment of forest fire hazards. Moreover, the predicted fire hazard value is close to the precalibrated fire hazard value. The results of the time series prediction and evaluation of fire hazard spatial distribution are good and close to the values in actual situations. Furthermore, the proposed model ranks the importance of the factors affecting the occurrence of fires. The most important factor is the annual diurnal sequence, which reflects the seasonal factor, followed by moisture and vegetation growth. The importance of topographic factors is low. Importance ranking can help in understanding the driving effects of different factors on the occurrence of fires and identifying which factors exert substantial effects on the occurrence of forest fires. Although the areas of forest fire occurrence and the factors that affect fire occurrence differ, the change rule of the fire hazard value is similar, that is, the fire hazard value is high in the week before the fire and low in other times. The spatial distribution of fire hazards is reasonable, and the fire hazard value in the fire area gradually increases from two months before the fire to the day of the fire. Moreover, the fire hazard value in the same area one year before the fire is much lower than the fire hazard value on the day of the fire, indicating an accurate assessment of the fire hazard situation. The proposed forest fire hazard assessment model involves comprehensive indicators, which can accurately assess fire hazard situations. It can be applied to different regions in China to partially solve the problem of regional restrictions.Keywords：forest fire;fire hazard;remote sensing;Random Forest;hazard monitoring1695|4513|0<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2025-04-21
- Abstract：Swidden agriculture is a widespread but controversial traditional land-use type in the tropics, especially in mountainous Laos with high percentage of forest cover. Driven by population growth, forestry policies, and climate change, swidden agriculture has been experiencing rapid evolution itself and drastic transformations into commercial plantations, such as rubber plantation. However, the remote sensing monitoring of tropical swidden agriculture has always been challenged, primarily because of the spatial and temporal dynamics in agricultural and forest cover, marginal feature compared with modern agriculture, and fragmentation and random distribution of swidden patches, hence with many unsettled issues and very limited information on its involved population, exact distribution and spatio-temporal dynamics. To explore the application potentials of machine learning algorithms in monitoring swidden agriculture, with two Landsat Operational Land Imager (OLI) images acquired in April, or the peak of the 2016 dry season, a support machine algorithm (SVM) was modified by masking out the information of construction land to improve the classification accuracy, or an overall accuracy of 95% and a Kappa coefficient of 0.81, followed by the examination of spatial (e.g., district-level) differences of freshly-opened swidden in Phongsaly Province, Laos, and their characteristics to local settlements and varied-level roads as well as topographical features. The results showed that: (1) Swidden agriculture remains an important land use type in Phongsaly because the newly-opened swidden was about 987.93 km2 (6.10% of the province) in 2016. More swidden patches were detected in the south and west parts of the province, with a fragmented distribution. (2) The area of newly-opened swiddens at district level ranged between 100—210 km2, with Samphanh District ranking the first (1/5) and Boonneua District the last (1/10). (3) Approximately 90% of newly-opened swiddens were concentrated within five km to residential points, particularly within four km. Similarly, these swiddens exhibited a distance decay pattern along the minor roads, tracks and major roads, in particular within a distance of five km of minor roads. (4) The newly-opened swiddens were mainly distributed in low mountain area (500—1000 m) with slope gradients of 15°—25° and aspects of southeast, showing slight variations among the districts of Phongsaly Province. This study provides reference for exploring machine learning algorithms in remote sensing monitoring of swidden agriculture in transition in the tropics.Keywords：swidden agriculture;Support Vector Machine (SVM);Landsat;accessibility analysis;topographic features;Laos3010|2501|5<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2022-12-09
- Abstract：In recent years, forest fires occur frequently around the world, which severely damage the structure and function of the forest ecosystem. The initial assessment of burn severity could provide a quantitative basis for rapid implementations of post-fire restoration measures. In the last decades, remote sensing-based models have become an appropriate choice to assess burn severity, which generally require a certain amount of field survey data. However, this requirement could not be sufficiently satisfied in the first moments after fire, since the field survey work would cost a substantial amount of time and labor. The absence of field survey data in the initial assessment of burn severity would largely limit the efficient application of remote sensing technologies. In this study, a transfer learning algorithm (i.e., SSTCA, semi-supervised Transfer Component Analysis) was employed to propose an initial assessment model of burn severity to improve the time-efficiency of traditional remote sensing-based models. Firstly, the SSTCA algorithm was applied to project a series of new features from original spectral features of remotely sensed data. Based on these projected features, a Support Vector Regression (SVR) model was then trained using historical field survey data from source areas (i.e., Bear fire on June 27, 2002 and Mule fire on July 11, 2002). Thereafter, the SSTCA-SVR model was transferred to the initial assessment of burn severity of a target area (i.e., Lushan fire on March 30, 2020). Finally, the performance of this proposed model was quantitatively compared with those of some traditional models (i.e., dNDVI-, dLST-, dNBR-, and SVR-based models). Results showed that original spectral features of remote sensing images over source and target areas were quite different. After the SSTCA projection, projected features of source and target samples have a similar distribution pattern in the new features-based space. Meanwhile, in the initial assessment of burn severity, dNDVI- and dNBR-based models have overestimated burn severity levels with low accuracies (i.e., overall accuracy was from 20.80% to 24.80% and Kappa value was between 0.01 and 0.06). Compared with them, the dLST-based model has a better performance with an overall accuracy of 34.80% and a Kappa value of 0.19. Although SVR-based model has shown a promising performance with an overall accuracy of 58.00% and a Kappa value of 0.48, this model has overestimated the burn severity levels in some regions of burned areas. The assessment results of burn severity levels using SSTCA-SVR model has the best performance with an overall accuracy of 71.20% and a Kappa value of 0.64. We conclude that the application of a transferring learning algorithm would be helpful for building an assessment model of burn severity with a good transferring ability. In this way, more accurate results could be obtained in the initial assessment of burn severity, and the response of post-fire management might be accelerated after forest fires.Keywords：burn severity;transfer learning;initial assessment;Landsat;forest fire;Lushan1559|2368|5<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2022-11-18
- Abstract：As the largest land cover, forests play an important role in human living environment, biological habitat, and global carbon cycle. Forest health is directly related to global ecological security and sustainable development of human society. In recent years, urban construction, disasters, forest management and deforestation, and other factors have caused different degrees of disturbance to forests. It is important to determine the exact time point and spatial range of forest burned area for forest damage assessment, management, carbon accounting, and forest restoration management. Owing to the continuity of spatial distribution of forest burned areas, most of the existing methods of forest burned area extraction use the two-step treatment strategy of first classification and then post-processing to suppress the effect of false alarm pixels. In this paper, a spatiotemporal detection method, Stacked ConvLSTM, is proposed for the detection of forest fire tracks in time series. This method avoids subjective post-processing operations on the basis of maintaining better spatial continuity of the results, and achieves end-to-end extraction of forest burned area information, which improves the extraction accuracy of forest fire-burning land. This paper proposes to use Stacked ConvLSTM to detect forest disturbance in time and space. Combined with the characteristics of ConvLSTM in extracting temporal and spatial characteristics from long-term historical series, it can predict the change trend of vegetation in a period of time in the future, and accurately determine the time point and spatial range of forest disturbance. ConvLSTM is an LSTM variant proposed on the basis of LSTM. The full connection state from input layer to hidden layer and from hidden layer to hidden layer of LSTM is replaced by convolution connection, which can make full use of spatial information. Compared with single-pixel-based methods, ConvLSTM can extract the spatiotemporal structure information of time series images at the same time, which is better for spatiotemporal analysis. In this paper, Stacked ConvLSTM is used to detect the temporal and spatial distribution of forest burned areas, predict the change trend of vegetation in a period of time in the future, and determine the presence of forest burned areas by comparing with the newest time-series images. With MODIS long time series data, based on the historical time series of Yinanhe Forest Farm of Zhanhe Forestry Bureau in Heilongjiang Province and Beidahe Forest Farm of Bilahe Forestry Bureau in Inner Mongolia from 2001—2008 and 2001—2016, the extraction results of burned areas were compared with Stacked LSTM and bfast algorithm. The Stacked ConvLSTM, Stacked LSTM, and bfast algorithms were used to extract forest burned areas from MODIS time series in both regions, and to compare the detection results with the Fire_CCI 5.1 burned areas products released by ESA. Results show that, firstly, from the visual effect, in study area Ⅰ, the error detection of Stacked ConvLSTM is fewer than that of Stacked LSTM and bfast algorithm and maintains high continuity in spatial distribution. In study Area Ⅱ, Stacked ConvLSTM detected a more complete area of fire. Secondly, in study area Ⅰ , Stacked ConvLSTM was 0.120 and 0.405 more accurate than Stacked LSTM and bfast algorithms, respectively. Moreover, the recall rate, accuracy, and Fire_CCI 5.1 F1-score were higher. In study area Ⅰ , the accuracy of Stacked ConvLSTM is 0.924 had a higher recall rate, accuracy, and F1-score than Stacked LSTM, bfast algorithms, and Fire_CCI 5.1. The detection accuracy of ConvLSTM model in space is higher than that of the other two methods, and its continuity of detection results in space is better. The detection effect of ConvLSTM model is equivalent to that of Stacked LSTM in time, but both of them are closer to the real fire time point than bfast algorithm. Results show that Stacked ConvLSTM has advantages in obtaining the change trend of forest long-term historical series for spatiotemporal prediction, and improves the detection accuracy of forest fire to a certain extent.Keywords：Stacked ConvLSTM;time series;spatiotemporal prediction;forest burned area2067|2874|4<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2022-11-18
- Abstract：Forest fires seriously affect the environment and social economy, e.g., damaging infrastructure, causing economic losses, and endangering human health. Effective simulation and prediction of forest fire growth are greatly important. Fire behavior models can provide analytical schemes for characterizing and predicting the speeds and directions of fire spread. However, fire spread models are subject to assumptions and limitations that inherently produce compounding errors during simulations. Satellite remote sensing monitoring of forest fire can be used to analyze the spatial dynamic change process of large-scale fires. It is an economical and effective technology for obtaining fire information in a large range and a short period. It can also provide fire location information for fire spread models.This study proposes a new approach of fire spread simulations based on the assessment of simulated fire growth discrepancies by using satellite active fire data. The FARSITE fire spread simulator was used to simulate the spread of forest fires that occurred on May 17, 2017 in Chenbaerhuqi, Inner Mongolia autonomous region, China, and the S-NPP\VIIRS forest active fire data were applied into the FARSITE simulator for calibration and re-initialization. The Landsat-8 and GF-1 data were used to generate the data required by the FARSITE. The fire field for different time periods was monitored by the multisource satellite data Sentinel-2A, GF-1 and GF-4 data. 375 m VIIRS active fire monitoring data were employed for re-initializing FARSITE fire simulation. We combined the satellite fire data and fire spread model for reducing errors of simulation results caused by the condition limitation of fire model, and the Sørensen’s coefficient (SC) was employed to evaluate the accuracy of fire spread simulation results at FARSITE before and after reinitializing the simulator for VIIRS active fire data.The re-initialization results of the FARISTE simulator by VIIRS active fire data showed that the simulation accuracy in each simulation process gradually decreased along with time. The distribution of simulation results indicated that the simulation findings after re-initialization were consistent with the actual fire perimeter monitored by high or moderate resolution remote sensing data. The highest precision in the process using active fire data increased by 56.89%, and the final accuracy increased by 45.45%. The final SC value increased from 54.14% to 78.76% when the satellite data were used to re-initialize the FARSITE fire simulation system, with increment of 42.76%. The maximum SC value was 87.8% for VIIRS active fire data re-initialization during simulation. The re-initialization approach meaningfully improved the accuracy of fire simulation.The use of satellite remote sensing active fire data and the re-initialization of FARSITE limited the further expansion of the error of fire spread model and improved the reliability and accuracy of forest fire simulation. This innovative approach represents a potential scheme for reducing the error of large-scale fire simulation results that can improve the reliability of fire spread model. This method provides an effective data assimilation method for fire prediction. It also provides a basis for fire management departments to manage forests and develop fire suppressing plans. In this study, the actual ground and air fire-fighting forces change the results of fire spread. They are an important factor for the deviation between the simulation results and the actual results.Keywords：remote sensing;forest fires;forest fires spread model;VIIRS;FARSITE;fire behavior simulations3177|4347|11<HTML><L-PDF><Enhanced-PDF><Meta-XML>Updated：2022-09-29
