Advances in Forest Fire Detection, Prediction and Behavior: A Comprehensive Survey

: Forest fires are a major environmental challenge that poses a threat to both human life and ecological health. To effectively prevent and manage forest fires, it is crucial to have reliable detection, prediction and behavior analysis systems in place. This study provides a comprehensive survey of the different approaches and techniques used for forest fire detection, prediction and behavior analysis. It covers ground-based and aerial surveillance systems, remote sensing technologies, machine learning-based approaches and social media-based systems. The paper also discusses the challenges and limitations of current systems and provides insights into future directions for research and development in this field. Overall, this study highlights the importance of leveraging multiple data sources and analysis methods to improve our understanding of forest fire behavior and develop effective strategies for managing this environmental threat.


Introduction
Forest fires are a serious threat to the environment, wildlife and human lives around the world.Early detection, accurate prediction and effective management of forest fires are crucial to minimizing their impact.In recent years, advances in technology and machine learning have provided new opportunities for improving forest fire detection, prediction and behavior analysis.There are various methods and technologies available for forest fire detection, including ground-based systems, aerial surveillance and remote sensing techniques.One new contribution to forest fire detection is the use of machine learning algorithms and satellite data to identify and track potential wildfires.By analyzing changes in vegetation and thermal anomalies in satellite imagery, machine-learning algorithms can detect and classify potential fires with high accuracy (Jain et al., 2020).Furthermore, the use of drones equipped with thermal cameras and sensors can also aid in the detection of forest fires.These drones can quickly and efficiently survey large areas of forest and identify areas of high heat or smoke (Guimarães et al., 2020).
In addition to these technological advancements, community involvement is also crucial in detecting forest fires.Citizen reporting and monitoring systems, such as the Firewatch app (Planet, 2016), can help individuals report potential fires and receive alerts about nearby fires in real time.
Overall, the integration of machine learning algorithms, satellite data, drones, WSN, IOT and community involvement in forest fire detection can significantly improve the speed and accuracy of identifying and preventing forest fires.Here is a brief summary of some of the current methods.

Forest Fire Detection Techniques
Over the years numerous have been developed to detect forest fires, ranging from ground-based systems to aerial surveillance methods.The following techniques are used for detecting forest fires as shown in Fig. 1: • Ground-based systems: Ground-based systems include fixed and mobile sensors that can detect smoke, heat and other indicators of fire.Examples include infrared cameras, flame detectors and thermal sensors These systems are relatively inexpensive and can provide real-time data, but they are limited by their range and require a network of sensors to cover large areas (Ambrosia and Wegener, 2009;Ryu and Kwak, 2022) • Aerial surveillance: Aerial surveillance involves using drones or aircraft to monitor forests and detect signs of fire.This method can cover a larger area than ground-based systems and can provide highresolution imagery.However, it requires specialized equipment and trained personnel to operate and the imagery must be analyzed in real time (Daud et al., 2022;Mehmood et al., 2022) • Remote sensing: Remote sensing involves using satellites and other remote sensors to detect heat and smoke from forest fires.This method can cover large areas and is particularly useful in detecting fires in remote or inaccessible regions.However, it can be expensive and is limited by cloud cover and other environmental factors (Santos et al., 2021;Dennison et al., 2014) • Machine learning-based approaches: Machine learning-based approaches involve training algorithms to detect forest fires using data from sensors or satellite imagery.These methods can improve the accuracy and speed of detection and can be applied to various types of data.However, they require large amounts of training data and can be limited by the quality of the input data (Sathishkumar et al., 2023;Zhang et al., 2016) • Social media-based approaches: Social mediabased approaches involve using social media platforms to detect and monitor forest fires.These methods rely on crowdsourced data and can provide real-time information about the location and severity of fires.However, they are limited by the availability and reliability of social media data (Slavkovikj et al., 2014) Fig. 1: Forest fire detection techniques

Forest Fire Prediction Techniques
Forest fire prediction techniques aim to identify the likelihood of a fire occurring in a particular area and estimate its potential size and behavior.Below are some of the common techniques used for predicting forest fires as shown in Fig. 2 Overall, forest fire prediction techniques rely on a combination of data sources and analysis methods to provide a comprehensive understanding of the risk of forest fires.By combining these techniques, fire managers can develop effective strategies for preventing and managing forest fires.

Forest Fire Behaviour Analysis Techniques
Forest fire behavior analysis techniques are used to understand and predict the behavior of a forest fire, including its rate of spread, direction, intensity and likelihood of ignition.Here are some common techniques used for forest fire behavior analysis as shown in Fig. 3.
• Fire behavior modeling: This technique involves using mathematical models to simulate the spread and behavior of a fire.Models can take into account factors such as topography, fuel characteristics and weather conditions to predict fire behavior (Kucuk and Sevinc, 2023;Alkhatib, 2013) • Fire behavior assessment: This technique involves direct observation and analysis of a fire to assess its behavior.Firefighters and fire behavior analysts use visual cues such as flame length, smoke color and wind direction to assess the fire's behavior and predict its spread (Andrews, 1986) • Fuel mapping: This technique involves creating detailed maps of the types and distribution of fuels in an area to predict the potential behavior of a fire.By understanding the types and densities of fuels, analysts can predict how quickly a fire will spread and how intense it will be (Scott, 2005;Heisig et al., 2022) • Weather monitoring: Weather conditions such as temperature, wind speed and humidity can greatly impact fire behavior.By monitoring weather conditions in real-time, fire behavior analysts can adjust their predictions and provide more accurate information to firefighters and other emergency responders (Miyanishi, 2001;Cohen, 2000) • Remote sensing: This technique involves using satellites or other remote sensing technologies to gather data on forest fires.This can include information on the location, size and behavior of the fire, which can be used to create more accurate fire behavior models and predictions (Chuvieco et al., 2010) Fig. 3: Forest fire behavior analysis techniques Overall, understanding the behavior of a forest fire is crucial for firefighting and emergency response efforts.By using a combination of techniques like those listed above, analysts can provide valuable information to firefighters and emergency responders to help them contain and control the fire as quickly and safely as possible.

Ground-Based Techniques Used for Forest Fire Detection
Ground-based systems are one of the methods used for forest fire detection.These systems typically use a network of sensors, cameras, or other instruments placed on the ground to detect smoke, heat, or other indicators of a forest fire.Here are some of the most common types of ground-based systems used for forest fire detection:

•
Infrared cameras: Infrared cameras are devices that detect infrared radiation, which is emitted by objects based on their temperature.These cameras can detect heat signatures from wildfires and can be used to monitor fires at night when visibility is limited (Ambrosia and Wegener, 2009) Examples of ground-based systems used for forest fire detection that incorporate infrared cameras: − Fire watch: Fire watch is a ground-based system that uses infrared cameras to detect forest fires.The system is equipped with multiple sensors, including infrared cameras, to provide a 360°-view of the surrounding area.The system uses algorithms to analyze the thermal images captured by the infrared cameras and identify potential fire hotspots (IQ-FireWatch, 2023) FLIR aerial firefighting system: The FLIR aerial firefighting system is a ground-based system that uses infrared cameras to detect forest fires from a distance.The system includes a thermal imaging camera mounted on a telescoping mast, which can be raised to a height of up to 30 feet.The camera can detect temperature differences as small as 0.1° Celsius, allowing it to detect fires before they become visible to the naked eye (Georgiades et al., 2019) − DRS technologies tamarisk: The DRS technologies tamarisk system is a portable ground-based system that uses infrared cameras to detect forest fires.The system includes a thermal imaging camera mounted on a tripod, which can be easily transported and set up in the field.The camera can detect hotspots as small as 0.5 square meters, allowing it to quickly detect and locate fires (Jeong and Kim, 2020) • Flame detectors: Flame detectors use optical sensors to detect the light emitted by flames.These detectors can be used to detect fires at their early stages when the flames are small and may not be visible to the naked eye (Basu et al., 2018) Examples of ground-based systems used for forest fire detection that incorporate flame detectors: − Wasp wildfire detection system: The wasp wildfire detection system is a ground-based system that uses a combination of flame detectors and infrared sensors to detect forest fires.The system is mounted on a telescoping mast and can be used to monitor large areas.
The flame detectors can detect the unique signature of a forest fire and alert firefighters to the presence of a fire (WASP, 2020) − Fire sentry FS24X flame detector: The fire sentry FS24X flame detector is a portable ground-based system that uses a sophisticated infrared sensor to detect forest fires.The system can detect fires from up to 200 feet away and can be easily transported and set up in the field.The flame detector can distinguish between real fires and false alarms caused by things like sunlight, headlights and other heat sources (Lisakov et al., 2015) − Viper perimeter protection system: The viper perimeter protection system is a ground-based system that uses a combination of flame detectors, infrared sensors and video cameras to detect forest fires.The system is designed to monitor large areas and can be integrated with other sensors and alarms to provide comprehensive wildfire detection and protection (Viper perimeter protection system, 2023) • Thermal sensors: Thermal sensors detect changes in temperature and can be used to detect heat signatures from wildfires.These sensors can be installed on towers or other structures and can be connected to a central monitoring system Examples of ground-based systems used for forest fire detection that incorporate thermal sensors: − AlertWildfire: AlertWildfire is a ground-based system that uses thermal sensors to detect forest fires.The system is made up of a network of cameras equipped with thermal sensors that can detect heat signatures associated with forest fires.The cameras are strategically placed in areas prone to wildfire and are connected to a central monitoring system that can alert firefighters and other emergency responders in real time (AlertWildfire, 2024) − ThermEye: ThermEye is a portable ground-based system that uses thermal sensors to detect forest fires.The system is designed to be quickly deployed in the field and can be used to monitor large areas.The sensors are mounted on a tripod and can detect temperature differences as small as 0.1° Celsius, allowing it to detect fires before they become visible to the naked eye drone equipped with a thermal camera that can detect the heat signature of a forest fire.The drone is designed to be launched from a catapult and can stay aloft for up to 24 h, making it a useful tool for monitoring large areas.The drone can be operated remotely and can transmit real-time video and data back to a command center (Insitu, 2005) These are just a few examples of aerial surveillance systems that incorporate drones for forest fire detection.Drones are useful for monitoring forested areas and can provide early warning of potential fires, allowing firefighters and other first responders to respond quickly and effectively.

•
Satellites: Satellites equipped with sensors that can detect changes in temperature, reflectance and other indicators of forest fires can be used for forest fire detection.These sensors can detect the heat emitted by fires, the smoke plumes and the changes in vegetation that can occur as a result of a fire (Alkhatib, 2014) Examples of aerial surveillance systems used for forest fire detection that incorporate satellites: − MODIS (moderate resolution imaging spectroradiometer): MODIS is a key instrument aboard two NASA Earth Observing System (EOS) satellites that can detect active fires and hot spots around the world.The sensors on MODIS satellites detect the amount of heat being radiated from the Earth's surface, which can be used to identify the location of active fires.MODIS has 1 KM and 500 M of spatial resolution and it can detect fires as small as 10 m across and provides near-real-time data to firefighters and other first responders (Alkhatib, 2014) − Landsat: The Landsat program is a series of Earthobserving satellites that provide multispectral data for monitoring and managing natural resources.The sensors on Landsat satellites can detect changes in vegetation, which can be an indicator of potential fire risk.The data from Landsat can be used to create fire risk maps and identify areas that may be at high risk of wildfire (Thapa et al., 2021) − Sentinel-2: Sentinel-2 is a European Union earth observation mission that provides high-resolution multispectral imagery of the Earth's surface.The sensors on sentinel-2 satellites can detect changes in vegetation and identify areas that may be at high risk of wildfire.The data from sentinel-2 can be used to monitor forested areas and identify potential fires, providing early warning to firefighters and other first responders (Thapa et al., 2021) − Fire towers: Fire towers are tall structures that are typically located on hilltops or other high points and used for fire detection and monitoring.Fire spotters can use binoculars or other optical instruments to scan the forest for signs of smoke or fire.Fire towers can be connected to a central monitoring system to facilitate communication with firefighters and other emergency responders (Alkhatib, 2014) Aerial surveillance provides an important complement to ground-based systems for forest fire detection.By using a combination of ground-based and aerial surveillance systems, forest managers can develop a comprehensive strategy for detecting and responding to wildfires.Aerial surveillance can also be used for forest fire prediction, in addition to detection.Here are some examples of how aerial surveillance is used for forest fire prediction (Thapa et al., 2021;Partheepan et al., 2023): • Mapping vegetation and fuel types: Aerial imagery can be used to map the types of vegetation and fuel present in forested areas.This information can be used to predict the likelihood of a fire occurring and how quickly it may spread.For example, areas with high fuel loads and dense vegetation are more likely to experience fast-spreading fires • Monitoring weather conditions: Aerial surveillance can also be used to monitor weather conditions, such as temperature, humidity, wind speed and direction.This information can be used to predict how weather conditions will affect the likelihood and behavior of fires.For example, high winds can cause fires to spread quickly, while high humidity can reduce the risk of fires • Early warning systems: Aerial surveillance can be used to detect early signs of fire, such as smoke plumes or hot spots.This information can be used to trigger early warning systems, which can alert firefighters and emergency responders to the potential threat • Real-time monitoring: Aerial surveillance can be used to provide real-time monitoring of fires as they occur.This information can be used to predict the behavior of the fire and determine the best course of action for controlling and suppressing it Aerial surveillance plays an important role in forest fire prediction, in addition to detection.By using a combination of ground-based systems, aerial surveillance and remote sensing, forest managers can develop a comprehensive strategy for preventing and responding to wildfires.

Machine Learning-Based Approaches Used for Forest Fire Detection
Machine learning-based approaches are increasingly being used for forest fire detection due to their ability to automate the process of analyzing large amounts of data from various sources.Here are some examples of how machine learning is used for forest fire detection (Masoud et al., 2019;Grari et al., 2022): • Image classification: Machine-learning algorithms can be trained to classify images from ground-based systems, aerial surveillance and remote sensing data into fire and non-fire categories.These algorithms use deep neural networks to learn patterns and features of fire and can identify even small smoke plumes that may not be visible to the naked eye • Anomaly detection: Machine-learning algorithms can be used to detect anomalies in the data that may indicate the presence of a forest fire.These algorithms can detect sudden changes in temperature, humidity, wind direction, or other environmental factors that may be indicative of a fire fusion of multi-modal data: Machine learning algorithms can be used to fuse data from different sources such as satellite imagery, drone footage and ground-based cameras to provide a more comprehensive and accurate detection of forest fires • Early warning systems: Machine-learning algorithms can be used to develop early warning systems that can detect fires in their early stages and alert fire departments and other emergency responders.These systems can use historical data, weather data and other environmental data to predict the likelihood of a forest fire occurring in a given area • Fire spread prediction: Machinel e a r n i n g algorithms can be used to predict the spread of a forest fire based on data such as topography, vegetation type, wind direction and humidity levels.This information can be used to inform decisions on how to best manage the fire Overall, machine learning-based approaches are an effective way to detect forest fires and provide timely alerts to emergency responders, enabling them to take action quickly and potentially prevent the spread of the fire.By combining machine learning with other detection methods, forest managers can develop a comprehensive strategy for detecting and responding to forest fires.Machine learning-based approaches are also increasingly being used for forest fire prediction, which involves estimating the likelihood and severity of a potential fire before it occurs.Here are some examples of how machine learning is used for forest fire prediction ( Bot and Borges, 2022):

•
Fire risk assessment: Machine-learning algorithms can be used to predict the likelihood of a forest fire based on factors such as weather patterns, fuel moisture content and historical fire data.These algorithms can learn from past fires and weather conditions to identify areas at high risk of fire and provide recommendations for preventative measures such as controlled burns or vegetation management • Fuel moisture content prediction: Machinelearning algorithms can be used to predict the fuel moisture content, which is a key factor that affects the severity and spread of a forest fire.These algorithms can use weather data, vegetation type and other environmental factors to predict the moisture content of the vegetation and provide early warning of potential fires

Social Media-Based Approaches for Forest Fire Detection
Social media-based approaches for forest fire detection have become increasingly popular in recent years, especially as more people share information and images about fires on social media platforms.Here are some examples of how social media can be used for forest fire detection (Wu et al., 2020;Sakaki et al., 2010;Mukkamala, 2018;Jaber et al., 2018) Crowdsourced data collection: Social media platforms can be used to crowdsource data about the fire, such as the location of the fire and its size.
Emergency responders can use this information to improve their understanding of the fire and develop better strategies for responding to it • Public awareness: Social media can be used to raise public awareness about the dangers of forest fires and encourage individuals to take steps to prevent them.This can include sharing educational content about fire safety and providing updates on the current fire situation in a particular area Social media-based approaches can be a valuable tool for detecting forest fires and providing timely alerts to emergency responders.Social media-based approaches for forest fire prediction are still in their infancy and are not yet widely used.However, there are some potential applications of social media for predicting forest fires (Wu et al., 2020;Sakaki et al., 2010;Mukkamala, 2018) • Early warning systems: Social media can be used to identify trends and patterns that may be indicative of a forest fire.For example, an increase in posts about smoke or flames in a particular area could signal the start of a fire.Machine learning algorithms can be used to analyze this data and provide early warning alerts to emergency responders • Crowdsourced data collection: Social media can be used to crowdsource data about environmental conditions that may be relevant to forest fire prediction.For example, users can share information about the presence of dry or dead vegetation in a particular area, which can increase the risk of a fire • Public awareness: Social media can be used to raise public awareness about the importance of preventing forest fires and encourage individuals to take steps to 13 reduce their risk.This can include sharing educational content about fire prevention and providing updates on the current fire situation in a particular area Overall, social media-based approaches have the potential to provide valuable information for predicting forest fires.However, more research is needed to develop effective algorithms for analyzing social media data and integrating it with other prediction methods.Additionally, privacy concerns and data reliability issues must be addressed before social media-based approaches can be widely adopted.

Conclusion
As we reach the end of this survey paper on forest fire detection, prediction and behavior analysis techniques, we can see that the topic of forest fires is complex and multifaceted, with numerous challenges to address.Forest fires can have devastating effects on natural resources, wildlife and human communities.However, by using a combination of ground-based systems and aerial surveillance technologies, as well as machine learningbased approaches, it is possible to detect and predict forest fires and to understand their behavior in real time.
Ground-based systems such as infrared cameras, flame detectors, thermal sensors, weather stations, smoke detectors and acoustic sensors can be used to detect fires and provide real-time data to firefighters and other first responders.Aerial surveillance systems such as drones and satellites can provide a broad view of forested areas and can detect fires that may be difficult to see from the ground.These systems provide early warning to firefighters and other first responders, allowing them to respond quickly and effectively to forest fires.
Machine learning-based approaches such as image classification techniques can be used to identify fires in near-real-time, helping to detect fires earlier and provide early warning to first responders.These approaches can also be used to understand the behavior of forest fires and predict their spread, allowing for more effective and targeted firefighting efforts.
There is also a need for continued research into the underlying causes of forest fires, including climate change and human activities, as well as the development of effective strategies for preventing and managing forest fires.This will require collaboration between scientists, policy makers and stakeholders at all levels, as well as continued engagement with local communities and indigenous peoples who have valuable knowledge and expertise in managing and preventing forest fires.
The authors would like to express our sincere gratitude to the civil defense directorate and the ministry of agriculture for their invaluable support and assistance throughout the research process.Their guidance and expertise have been instrumental in the successful completion of this study.We would also like to acknowledge the civil defense directorate for providing the support and required resources.We are deeply grateful for their dedication and commitment to our research.Also, the authors would like to thank Al-Zaytoonah University of Jordan for funding this project and providing support during the project period.

•
(Krüll et al., 2012)ather stations can be used to monitor weather conditions that can contribute to the spread of wildfires.They can measure temperature, humidity, wind speed and other environmental factors that can affect the behavior of fires.Examples of ground-based systems used for forest fire detection that incorporate weather stations: The fire guard smoke detection system is a portable ground-based system that uses smoke detectors to detect forest fires.The system can be easily transported and set up in the field and is designed to monitor large areas.The smoke detectors are connected to a central monitoring system that can alert firefighters and other emergency responders in real-time(Krüll et al., 2012)-VESDA smoke detection system: The VESDA smoke detection system is a ground-based system that uses advanced smoke detection technology to detect forest fires.The system is equipped with multiple smoke detectors that are sensitive enough to detect smoke particles in the air, even before a fire is visible.The system can be integrated with other sensors and alarms to provide comprehensive wildfire detection and protection (VESDA, 2023) Fire watch: Fire watch is a ground-based system that uses a combination of thermal and acoustic sensors to detect forest fires.The acoustic sensors can detect the sound of a forest fire and the thermal sensors can detect the heat signature of a fire.The system uses advanced algorithms to analyze the data collected by the sensors and identify potential fire hotspots − DASH: Detection and Alert System for High-risk wildfires (DASH) is a ground-based system that uses acoustic sensors to detect forest fires.The system is equipped with multiple sensors that can detect the sound of a forest fire and can differentiate between the sound of a forest fire and other sources of noise.The system is designed to be quickly deployed in the field and can be used to monitor large areas (Wildfire Defense Systems, 2021)Ground-based systems are useful for detecting and monitoring forest fires in localized areas, such as campgrounds, residential neighborhoods, or areas close to roads.However, they may not be practical for largescale monitoring of remote forest areas due to their limited range.A combination of ground-based systems and other methods such as aerial surveillance and remote sensing may provide the most effective solution for forest fire detection and prevention.Ground-based systems can also be used for forest fire prevention, in addition to detection.Here are some examples of how ground-based systems are used for forest fire prevention(National Park  Service, 2024; USDA, 2019; National Interagency Fire  Center, 1973; US Department of the Interior, 2018): • Smoke detectors: Smoke detectors are sensors that can detect the presence of smoke and other particulate matter in the air.They can be used to detect forest fires by monitoring the concentration of smoke in the air. :