Forest Fire Prediction Research Articles

Forest Fire Prediction using Machine Learning

Abstract: Forest Fire Prediction is a key component of forest fire control. This is a major environmental problem that decreases resources such as water that causes global warming and water pollution. Fire Detection is a key element for controlling such incidents. Prediction of forest fire is expected to reduce the impact of forest fire in the future. It plays a major role in resource allocation, mitigation and recovery efforts. Forest fire prediction system is based on predictive modeling that predicts the fire on the basis of weather conditions that user provides as an input to the system. This project presents a description and analysis of forest fire prediction methods based on machine learning and discusses about a comparative study of different models for predicting forest fire such as Decision Tree Classifier, Random Forest Classifier .We use Flask framework to develop web application and imported the NumPy and Panda’s modules to access and perform operations on data sets. We have taken several data sets and train the system in order to predict the fire by taking Temperature, Oxygen, Humidity as parameters to the system.

Environmental drivers and spatial prediction of forest fires in the Western Ghats biodiversity hotspot, India: An ensemble machine learning approach

In ecologically sensitive areas like the Western Ghats, fire is largely considered to be the most significant management concern. Therefore, identifying and predicting the fire susceptible areas is crucial for managing protected area networks. In this study, we investigated the factors affecting forest fires and modelled the fire susceptibility in a human-dominated landscape in the central Western Ghats, India, by employing remote sensing, geographic information systems and machine learning techniques. We used six machine learning models (MLMs), including artificial neural network, random forest, generalized linear model, maximum entropy, multivariate adaptive regression splines, gradient boosting machine and the ensemble model, to relate the occurrence data of fires to 14 predictor variables categorized as climate, topography, vegetation, and anthropogenic disturbances to generate fire susceptibility maps. The area under the receiver operating characteristic (AUC-ROC), true skill statistic (TSS) curves, accuracy and continuous Boyce index (CBI) were used to assess the model's accuracy. We found that all models achieved acceptable performance while validating the independent test data. However, ensemble model had the best overall performance with an AUC-ROC = 0.93; TSS = 0.72; ACCURACY = 0.89; and CBI = 0.95. The land-use and land-cover, and the distance to the agricultural fields and settlements were the key determinants of fire occurrences in the study area based on the ensemble model. The results showed that the landscape has a high to very high fire risk for about 26.5% of the total area. Forest fires have mostly occurred in the northeastern regions of the landscape, which are dominated by deciduous forests and plantations, whereas western regions are less susceptible to fires. Finally, we developed an ensemble-based fire susceptibility map with a resolution of 10 m that may be used as a tool for the prevention of forest fires. Further, it aids the relevant authorities in mitigating the disastrous effects and safeguarding the environment. The results demonstrated that ensemble-based MLMs can be employed effectively for fire prediction in other regions.

An Accurate Forest Fire Recognition Method Based on Improved BPNN and IoT

Monitoring and early warning technology for forest fires is crucial. An early warning/monitoring system for forest fires was constructed based on deep learning and the internet of things. Forest fire recognition was improved by combining the size, color, and shape characteristics of the flame, smoke, and area. Complex upper-layer fire-image features were extracted, improving the input conversion by building a forest fire risk prediction model based on an improved dynamic convolutional neural network. The proposed back propagation neural network fire (BPNNFire) algorithm calculated the image processing speed and delay rate, and data were preprocessed to remove noise. The model recognized forest fire images, and the classifier classified them to distinguish images with and without fire. Fire images were classified locally for feature extraction. Forest fire images were stored on a remote server. Existing algorithms were compared, and BPNNFire provided real-time accurate forest fire recognition at a low frame rate with 84.37% accuracy, indicating superior recognition. The maximum relative error between the measured and actual values for real-time online monitoring of forest environment indicators, such as air temperature and humidity, was 5.75%. The packet loss rate of the forest fire monitoring network was 5.99% at Longshan Forest Farm and 2.22% at Longyandong Forest Farm.

Forest Fire Prediction Based on Long- and Short-Term Time-Series Network

Modeling and prediction of forest fire occurrence play a key role in guiding forest fire prevention. From the perspective of the whole world, forest fires are a natural disaster with a great degree of hazard, and many countries have taken mountain fire prediction as an important measure for fire prevention and control, and have conducted corresponding research. In this study, a forest fire prediction model based on LSTNet is proposed to improve the accuracy of forest fire forecasts. The factors that influence forest fires are obtained through remote sensing satellites and GIS, and their correlation is estimated using Pearson correlation analysis and testing for multicollinearity. To account for the spatial aggregation of forest fires, the data set was constructed using oversampling methods and proportional stratified sampling, and the LSTNet forest fire prediction model was established based on eight influential factors. Finally, the predicted data were incorporated into the model and the predicted risk map of forest fires in Chongli, China was drawn. This paper uses metrics such as RMSE to compare with traditional machine learning methods, and the results show that the LSTNet model proposed in this paper has high accuracy (ACC 0.941). This study illustrates that the model can effectively use spatial background information and the periodicity of forest fire factors, and is a novel method for spatial prediction of forest fire susceptibility.

An Ensemble Model for Forest Fire Occurrence Mapping in China

Assessing and predicting forest fires has long been an arduous task. Nowadays, the rapid advancement of artificial intelligence and machine learning technologies have provided a novel solution to forest fire occurrence assessment and prediction. In this research, we developed a novel hybrid machine-learning-technique algorithm to improve forest fire prediction based on random forest (RF), gradient-boosting decision tree (GBDT), support vector machine (SVM), and other machine learning models. The dataset we employed was satellite fire point data from 2010 to 2018 from the Chinese Department of Fire Prevention. The efficacy and performance of our methods were examined by validating the model fit and predictive capability. The results showed that the ensemble model LR (logistic regression)-RF-SVM-GBDT outperformed the single RFSVMGBDT model and the LR-RF-GBDT integrated framework, displaying higher accuracy and greater robustness. We believe that our newly developed hybrid machine-learning algorithm has the potential to improve the accuracy of predicting forest fire occurrences, thus enabling more efficient firefighting efforts and saving time and resources.

Forest Fire Prediction using Image Processing through Deep Learning

The important advantage of photograph processing from a fire detection device is its predictive software. This fireplace detection gadget can be hooked up nearly everywhere in a commercial building, shopping mall and many different public regions. This device makes use of cameras to come across fires. Therefore, so as to locate fireplace, we do no longer want different sensible. The digicam input machine is processed after which the technique is processed to stumble on fires. Thermal signatures and fireplace lights patterns are detected within the pix to decide if there is a fire and take appropriate action.

A geographical similarity-based sampling method of non-fire point data for spatial prediction of forest fires

Understanding the mechanisms and risks of forest fires by building a spatial prediction model is an important means of controlling forest fires. Non-fire point data are important training data for constructing a model, and their quality significantly impacts the prediction performance of the model. However, non-fire point data obtained using existing sampling methods generally suffer from low representativeness. Therefore, this study proposes a non-fire point data sampling method based on geographical similarity to improve the quality of non-fire point samples. The method is based on the idea that the less similar the geographical environment between a sample point and an already occurred fire point, the greater the confidence in being a non-fire point sample. Yunnan Province, China, with a high frequency of forest fires, was used as the study area. We compared the prediction performance of traditional sampling methods and the proposed method using three commonly used forest fire risk prediction models: logistic regression (LR), support vector machine (SVM), and random forest (RF). The results show that the modeling and prediction accuracies of the forest fire prediction models established based on the proposed sampling method are significantly improved compared with those of the traditional sampling method. Specifically, in 2010, the modeling and prediction accuracies improved by 19.1% and 32.8%, respectively, and in 2020, they improved by 13.1% and 24.3%, respectively. Therefore, we believe that collecting non-fire point samples based on the principle of geographical similarity is an effective way to improve the quality of forest fire samples, and thus enhance the prediction of forest fire risk.

Dynamic Response of Shrubbery Fire to Meteorological Changes in Yunnan Province

This paper takes shrubland fires as the research object and selects typical shrubland forest fire cases that occurred throughout Yunnan Province from 1954 to 2018 as the research samples, and the shrubland fire loss elements and meteorological data are studied by SPSS correlation analysis and MATLAB gray prediction model to analyze shrubland fires and meteorological elements. The results showed that there were significant correlations between shrubland fires and meteorological data; the shrubland fire over fire area, forest damage area, and economic loss composite data index showed positive correlations with annual maximum temperature, annual average minimum temperature, annual average temperature, and annual sunshine hours; the annual average relative humidity, precipitation days, and the composite index of the same day meteorological data showed significant negative correlations, etc. In exploring the main meteorological factors affecting the distribution of forest fires, we establish a meteorological fire model to provide a real and credible scientific basis and theoretical foundation for future forest fire prevention and prediction.

A Survey of the Machine Learning Models for Forest Fire Prediction and Detection

A Survey of the Machine Learning Models for Forest Fire Prediction and Detection

Deterministic–Probabilistic Prediction of Forest Fires from Lightning Activity Taking into Account Aerosol Emissions

Forest fires arise from anthropogenic load and lightning activity. The formation of a thunderstorm front is due to the influence of a number of factors, including the emission of aerosol particles from forest fires. The purpose of this study is mathematical modeling of heat and mass transfer in vegetation firebrand carried out from a forest fire front, taking into account the formation of soot particles to predict forest fire danger from thunderstorm activity. Research objectives: (1) development of a deterministic mathematical model of heat and mass transfer in a pyrolyzed firebrand of vegetation, taking into account soot formation; (2) development of a probabilistic criterion for assessing forest fire danger from thunderstorms, taking into account aerosol emissions; (3) scenario modeling of heat and mass transfer and the formation of a thunderstorm front; (4) and the formulation of conclusions and proposals for the practical application of the developed deterministic–probabilistic approach to the prediction of forest fires from thunderstorms, taking into account aerosol emissions. The novelty of this study lies in the development of a new model of heat and mass transfer in a pyrolyzed vegetation firebrand and a new probabilistic criterion for forest fire danger due to thunderstorm activity, taking into account aerosol emission. The distributions of temperature and volume fractions of phases in a firebrand are obtained for various scenarios. Scenarios of surface fires, crown forest fires, and a fire storm are considered for typical types of coniferous vegetation. Cubic firebrands are considered in the approximation of a two-dimensional mathematical model. To describe the heat and mass transfer in the firebrand structure, a differential heat conduction equation is used with the corresponding initial and boundary conditions, taking into account the kinetic scheme of pyrolysis and soot formation. Variants of using the developed mathematical model and probabilistic criterion in the practice of protecting forests from fires are proposed. Key findings: (1) linear deterministic–probabilistic mathematical model to assess forest fire occurrence probability taking into account aerosol emission and lightning activity; (2) results of mathematical modeling of heat and mass transfer in firebrand taking into account soot formation; (3) and results of scenario modeling of forest fire occurrence probability for different conditions of lightning activity and aerosol emission.

Developing a forest fire prediction model based on deep learning and forecast a fire risk map of Turkey

Orman yangınları; doğal hayatı, biyolojik ekosistemi tehdit eden ve ormancılık faaliyetlerini sekteye uğratarak ekonomik kayıplara neden olan ana sorunlardan birisidir. Yangınlar; bitki örtüsü, su ve hava gibi doğal kaynakların zarar görmesine veya tamamen yok olmasına neden olabilir. Ayrıca ormanların çevresinde bulunan yerleşim veya tarım alanlarına da hasar vererek can ve mal kayıplarına sebep olabilir. Bu yüzden orman yangınlarıyla mücadele edilmesi ve tahribatın asgari düzeye indirilmesi için kaynakların kullanılmasında doğru kararların verilmesi ve etkili planlamaların yapılması önem arz etmektedir. Bu çalışmada orman yangınlarıyla mücadelede kullanılmak için Orman Yangını Analiz Tahmin (OYAT) modeli oluşturulmuştur. OYAT; resmi kaynaklardan elde edilen bitki örtüsü, iklim değişiklikleri, beşeri etmenler ve daha önceki yangın verilerine dayandırılarak geliştirilmiştir. Bu veriler derin öğrenme tekniği ile işlenerek OYAT modeli oluşturulmuştur. OYAT kullanılarak, analiz edilen veriler ile orman yangını tahmini yapılır ve bölgesel yangın risk haritası elde edilir. OYAT coğrafi bilgi sistemleri (CBS) ile uyumlu çalışabilmesi sayesinde kolay kullanıma sahip, dinamik veriler ile güncellenebilen ve görselleştirilerek saklanabilen bir yapıya sahiptir. Türkiye’deki 2013-2019 yılları arasındaki orman yangını verileri OYAT modelinin değerlendirilmesinde kullanılmıştır. OYAT ile oluşturulan risk haritalarının %98 doğruluk oranına sahip olduğu gözlenmiştir. İzleyen yıllarda veri havuzunun büyümesiyle birlikte, OYAT modelinin yangın önleme için yapılacak planlamalarda daha verimli ve başarılı olacağı öngörülmektedir.

CANTO: An actor model-based distributed fog framework supporting neural networks training in IoT applications

The large volumes of Internet of Things (IoT) data transmission to and from the cloud leads to one of cloud-centric processing’s major drawbacks: latency. Fog computing gives a solution to this by bringing the processing closer to the edge devices. Although a lot of work has been done which makes use of fog nodes like aggregators and offloaders, data intensive tasks like machine learning are still, for the most part, being performed on cloud. In a fog network, there are multiple fog devices, albeit resource-constrained ones. We can make use of distributed processing to collectively utilize these resources. So, we propose CANTO, a general distributed framework which uses the actor model to train neural networks on a network of fog devices in an IoT setting. Since the actors communicate using messages and act according to the type of message they receive, the framework provides a new message type wherein we can specify parameters like the dataset, the size of each dataset part, the activation function, learning rate, etc. The framework is containerized and deployed in a docker swarm. The framework is demonstrated with the use case on IoT-based forest fire prediction. In addition, the effectiveness of the framework is demonstrated with respect to accuracy, latency and load distribution.

Evaluation of geographically weighted logistic model and mixed effect model in forest fire prediction in northeast China

IntroductionForest fires seriously threaten the safety of forest resources and human beings. Establishing an accurate forest fire forecasting model is crucial for forest fire management.MethodsWe used different meteorological and vegetation factors as predictors to construct forest fire prediction models for different fire prevention periods in Heilongjiang Province in northeast China. The logistic regression (LR) model, mixed-effect logistic (mixed LR) model, and geographically weighted logistic regression (GWLR) model were developed and evaluated respectively.ResultsThe results showed that (1) the validation accuracies of the LR model were 77.25 and 81.76% in spring and autumn fire prevention periods, respectively. Compared with the LR model, both the mixed LR and GWLR models had significantly improved the fit and validated results, and the GWLR model performed best with an increase of 6.27 and 10.98%, respectively. (2) The three models were ranked as LR model < mixed LR model < GWLR model in predicting forest fire occurrence of Heilongjiang Province. The medium-and high-risk areas of forest fire predicted by the GWLR model were distributed in western and eastern parts of Heilongjiang Province in spring, and western part in autumn, which was consistent with the observed data. (3) Driving factors had strong temporal and spatial heterogeneities; different factors had different effects on forest fire occurrence in different time periods. The relationship between driving factors and forest fire occurrence varied from positive to negative correlations, whether it’s spring or autumn fire prevention period.DiscussionThe GWLR model has advantages in explaining the spatial variation of different factors and can provide more reliable forest fire predictions.

Predictive model of spatial scale of forest fire driving factors: a case study of Yunnan Province, China

Forest fires are among the major natural disasters that destroy the balance of forest ecosystems. The construction of a forest fire prediction model to investigate the driving mechanism of fire drivers on forest fires can help reveal the mechanism of forest fire occurrence and its risk, and thus contribute to the prevention and control of forest fires. However, previous studies on the mechanisms of forest fire drivers have not considered the effect of differences in spatial scale of action of forest fire drivers on the predicted effect. Therefore, the present study proposes a spatial prediction model of forest fires that considers the spatial scale effect of forest fire drivers to predict forest fire risk. First, based on historical forest fire data and geographic environmental data in the Yunnan Province, geographically weighted logistic regression (GWLR) was used to determine the forest fire drivers and to estimate the probability of forest fire occurrence at locations where fire observations are absent. Then, multi-scale geographically weighted regression (MGWR) was used to explore the spatial scales of action of different drivers on forest fires. The results show that meteorological factors such as relative humidity, air temperature, air pressure, sunshine hours, daily precipitation, wind speed, topographic factors such as elevation, slope, and aspect, anthropogenic factors such as population density and road network, as well as vegetation type, were significantly correlated with forest fires; thus, they are identified as important factors influencing occurrence of forest fires in the Yunnan Province. The MGWR model regression results show that the role of different forest fire drivers on forest fire occurrence has spatial scale differences. The spatial scale of drivers such as altitude, aspect, wind speed, temperature, slope, and distance from the road to the fire point was larger and their spatial influence was relatively stable, with spatial heterogeneity having less influence on the model evaluation results. The spatial scale of drivers such as relative humidity, sunshine, air pressure, precipitation, population density, and vegetation type were smaller, and spatial heterogeneity had a more obvious influence on the model evaluation results. This study provides a reference for selecting drivers and evaluating their spatial scale effects to construct predictive regional forest fire models.

Forest Fire Occurrence Prediction in China Based on Machine Learning Methods

Forest fires may have devastating consequences for the environment and for human lives. The prediction of forest fires is vital for preventing their occurrence. Currently, there are fewer studies on the prediction of forest fires over longer time scales in China. This is due to the difficulty of forecasting forest fires. There are many factors that have an impact on the occurrence of forest fires. The specific contribution of each factor to the occurrence of forest fires is not clear when using conventional analyses. In this study, we leveraged the excellent performance of artificial intelligence algorithms in fusing data from multiple sources (e.g., fire hotspots, meteorological conditions, terrain, vegetation, and socioeconomic data collected from 2003 to 2016). We have tested several algorithms and, finally, four algorithms were selected for formal data processing. There were an artificial neural network, a radial basis function network, a support-vector machine, and a random forest to identify thirteen major drivers of forest fires in China. The models were evaluated using the five performance indicators of accuracy, precision, recall, f1 value, and area under the curve. We obtained the probability of forest fire occurrence in each province of China using the optimal model. Moreover, the spatial distribution of high-to-low forest fire-prone areas was mapped. The results showed that the prediction accuracies of the four forest fire prediction models were between 75.8% and 89.2%, and the area under the curve (AUC) values were between 0.840 and 0.960. The random forest model had the highest accuracy (89.2%) and AUC value (0.96). It was determined as the best performance model in this study. The prediction results indicate that the areas with high incidences of forest fires are mainly concentrated in north-eastern China (Heilongjiang Province and northern Inner Mongolia Autonomous Region) and south-eastern China (including Fujian Province and Jiangxi Province). In areas at high risk of forest fire, management departments should improve forest fire prevention and control by establishing watch towers and using other monitoring equipment. This study helped in understanding the main drivers of forest fires in China over the period between 2003 and 2016, and determined the best performance model. The spatial distribution of high-to-low forest fire-prone areas maps were produced in order to depict the comprehensive views of China’s forest fire risks in each province. They were expected to form a scientific basis for helping the decision-making of China’s forest fire prevention authorities.

Predicting soil moisture conditions across a heterogeneous boreal catchment using terrain indices

Abstract. Soil moisture has important implications for drought and flooding forecasting, forest fire prediction and water supply management. However, mapping soil moisture has remained a scientific challenge due to forest canopy cover and small-scale variations in soil moisture conditions. When accurately scaled, terrain indices constitute a good candidate for modelling the spatial variation of soil moisture conditions in many landscapes. In this study, we evaluated seven different terrain indices at varying digital elevation model (DEM) resolutions and user-defined thresholds as well as two available soil moisture maps, using an extensive field dataset (398 plots) of soil moisture conditions registered in five classes from a survey covering a (68 km2) boreal landscape. We found that the variation in soil moisture conditions could be explained by terrain indices, and the best predictors within the studied landscape were the depth to water index (DTW) and a machine-learning-generated map. Furthermore, this study showed a large difference between terrain indices in the effects of changing DEM resolution and user-defined thresholds, which severely affected the performance of the predictions. For example, the commonly used topographic wetness index (TWI) performed best on a resolution of 16 m, while TWI calculated on DEM resolutions higher than 4 m gave inaccurate results. In contrast, depth to water (DTW) and elevation above stream (EAS) were more stable and performed best on 1–2 m DEM resolution. None of the terrain indices performed best on the highest DEM resolution of 0.5 m. In addition, this study highlights the challenges caused by heterogeneous soil types within the study area and shows the need of local knowledge when interpreting the modelled results. The results from this study clearly demonstrate that when using terrain indices to represent soil moisture conditions, modelled results need to be validated, as selecting an unsuitable DEM resolution or user-defined threshold can give ambiguous and even incorrect results.

Knowledge Graph Representation Learning-Based Forest Fire Prediction

Forest fires destroy the ecological environment and cause large property loss. There is much research in the field of geographic information that revolves around forest fires. The traditional forest fire prediction methods hardly consider multi-source data fusion. Therefore, the forest fire predictions ignore the complex dependencies and correlations of the spatiotemporal kind that usually bring valuable information for the predictions. Although the knowledge graph methods have been used to model the forest fires data, they mainly rely on artificially defined inference rules to make predictions. There is currently a lack of a representation and reasoning methods for forest fire knowledge graphs. We propose a knowledge-graph- and representation-learning-based forest fire prediction method in this paper for addressing the issues. First, we designed a schema for the forest fire knowledge graph to fuse multi-source data, including time, space, and influencing factors. Then, we propose a method, RotateS2F, to learn vector-based knowledge graph representations of the forest fires. We finally leverage a link prediction algorithm to predict the forest fire burning area. We performed an experiment on the Montesinho Natural Park forest fire dataset, which contains 517 fires. The results show that our method reduces mean absolute deviation by 28.61% and root-mean-square error by 53.62% compared with the previous methods.

Forest Fire Prediction Using K-Mean Clustering and Random Forest Classifier

Prediction of fire forest will be needing several parameters that located on the same location and on the same time frame. This is important to have data on the same location and same time periods, since forest fire mostly triggered by weather or temperature condition on certain area on the certain time. Since there are many parameters involved, the preprocessing will need to made the data have standard structure. The clustering process needed to give a label to the data with five class label, very low risk, low risk, medium risk, high risk, very high risk. Based on the clustered data, the data training and data test given for random forest classifier for model development, the composition of the data training and data test is 70:30. The accuracies of both algorithm is very good 100%, precision, recall and f1-score also have very high score 100%. This meant that the forest fire prediction model will produce a good prediction.

Enhanced interpolation with semi-supervised algorithm and greedy forwarding technique for forest fire prediction over wireless sensor and actuator networks

Enhanced interpolation with semi-supervised algorithm and greedy forwarding technique for forest fire prediction over wireless sensor and actuator networks

Spatio-Temporal Knowledge Graph Based Forest Fire Prediction with Multi Source Heterogeneous Data

Forest fires have frequently occurred and caused great harm to people’s lives. Many researchers use machine learning techniques to predict forest fires by considering spatio-temporal data features. However, it is difficult to efficiently obtain the features from large-scale, multi-source, heterogeneous data. There is a lack of a method that can effectively extract features required by machine learning-based forest fire predictions from multi-source spatio-temporal data. This paper proposes a forest fire prediction method that integrates spatio-temporal knowledge graphs and machine learning models. This method can fuse multi-source heterogeneous spatio-temporal forest fire data by constructing a forest fire semantic ontology and a knowledge graph-based spatio-temporal framework. This paper defines the domain expertise of forest fire analysis as the semantic rules of the knowledge graph. This paper proposes a rule-based reasoning method to obtain the corresponding data for the specific machine learning-based forest fire prediction methods, which are dedicated to tackling the problem with real-time prediction scenarios. This paper performs experiments regarding forest fire predictions based on real-world data in the experimental areas Xichang and Yanyuan in Sichuan province. The results show that the proposed method is beneficial for the fusion of multi-source spatio-temporal data and highly improves the prediction performance in real forest fire prediction scenarios.