Integrating FARSITE and GIS for Enhanced Forest Fire Spread Prediction and Simulation

Integrating FARSITE and GIS for Enhanced Forest Fire Spread Prediction and Simulation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Integrating FARSITE and GIS for Enhanced Forest Fire Spread Prediction and Simulation Elham Goleiji, Hamide Aliani, Seyed Mohsen Hosseini This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5764687/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 18 Jun, 2025 Read the published version in Natural Hazards → Version 1 posted 5 You are reading this latest preprint version Abstract Given the widespread fires in forests worldwide, which result in the destruction of ecosystems, effective management of these forests—particularly in relation to fire behavior and its spread—is essential. Forest fire prediction is a worldwide problem. There are different models to predict the spread of fires, especially in forest areas but these models are usually difficult to apply for real forest fire cases and the accuracy of prediction is lower than reality. The aim of this research is to identify and predict fire spread behavior and the process of fire spread in natural areas using the FARSITE model. To conduct the research, data including topography, vegetation, and meteorological information, were compiled the study area. Using GIS tools and the FARSITE model, a simulation map of fire spread behavior in the studied forest was generated. The model tested against the spread of fires under the field conditions where real fire had occurred in forest area. To assess the accuracy of the simulation, the Kappa coefficient method was employed. The simulation results showed that 34.93 hectares of the observed fire area were accurately simulated as a burned area, demonstrating a high degree of consistency with the actual fire spread. The overestimated area was minimal (0.85 hectares), while 37.7 hectares of the fire spot were not simulated as burned. The findings of this study can inform fire spread behavior prediction models and guide the development of preventive and control measures for fire risk management in forests. Forest fire risk management Fire spread simulation FARSITE Geographical Information System Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 9 Figure 10 Figure 16 Figure 17 Figure 29 Introduction Wildfires represent a major environmental hazard, particularly in forested and grassland ecosystems, where they can cause extensive ecological and economic damage. The occurrence and intensity of wildfires are shaped by multiple factors, such as climate fluctuations, fuel properties, and terrain conditions. (Wang 2023; Ajin 2016; Mohammadi 2010; Goleiji 2017; Zigner 2020; Mirdeylami 2014). Forests, as one of the most vital renewable natural resources, Play a vital role in preserving biodiversity, regulating carbon cycles, and stabilizing regional climate patterns. (Zhang 2009; Jahdi 2015; Ghobadi 2012; Aliani 2016). However, wildfires pose a significant threat to these ecosystems, making fire modeling and predictive analysis essential for effective wildfire management (Nasanbat 2016 ; Marozas 2007; Paz 2011; Adab 2013). The modeling of fire spread dynamics is an essential component of wildfire risk assessment, as it enables the prediction of fire behavior under varying environmental conditions. The spread of fire is determined by a range of interconnected factors, such as vegetation type, fuel load, weather conditions, and topographical features. To aid in the analysis of fire propagation, several simulation models have been created, encompassing physical, empirical, and machine learning-based methods such as FIRESTAR (Cruz and Alexander 2013 ), FIRETEC (Marshall et al. 2020 ), and the Phoenix RapidFire model (Kevin et al. 2008 ) Depend on the core principles of combustion and heat transfer. Empirical models, including the widely used Rothermel model (Rothermel 1972 ), the McArthur model in Australia (Stocks et al. 2004 ), and the Canadian Fire Danger Rating System (CFFDRS) (Bowman et al. 2021 ), establish relationships between fire behavior and environmental factors based on observational data. In recent years the use of machine learning techniques has transformed fire prediction and risk assessment across various scientific disciplines. (Hegde and Rokseth 2020; Chen et al. 2019; Jain et al. 2020 ; Jordan and Mitchell 2015 ). Notably, convolutional neural networks (CNNs) have emerged as effective tools for forecasting the future status of fire-affected areas (Li et al. 2020 ; Radke et al. 2019 ). The development of advanced machine learning models, such as the Convolutional Long Short-Term Memory (ConvLSTM) networks, has demonstrated superior performance in predicting fire behavior compared to traditional methods (Burge et al. 2020). Fire simulation models, including FARSITE and Minimum Travel Time (MTT; Finney 1998 , 2002), are essential for assessing and managing fire risks (Keane 2001). Numerous studies have focused on fire risk assessment and fire spread simulation, highlighting the importance of accurate fire behavior modeling for effective fire management strategies. In contrast to North America and Europe, countries like Iran frequently lack detailed data on fire history, fuel characteristics, and environmental conditions, which can affect the precision of fire modeling systems. The reliability of these models hinges on three critical factors: the quality of input data, the theoretical framework of fire behavior models, and the efficiency of fire spread algorithms. (WenBin and Perera 2008; Jahdi 2015). FARSITE is a widely used fire spread simulation tool, but its effectiveness in Iran remains underexplored. While studies like Salis ( 2007 ), Nyatondo ( 2010 ), and Jahdi (2015) have validated its use in Mediterranean and montane regions, its application in Iranian forests lacks thorough validation. This study integrates FARSITE with GIS to enhance localized fire spread predictions, leveraging environmental data for more precise and context-specific modeling. While FARSITE has been widely used for fire spread simulations, its application to the unique environmental conditions of Iranian forests remains limited. This study integrates GIS techniques with FARSITE to assess its accuracy under varying conditions, identifying strengths and limitations. By developing detailed vegetation maps and fuel models, we aim to improve fire spread predictions and management strategies. Additionally, the findings will offer valuable insights for policymakers and fire management agencies, contributing to enhanced fire prevention and mitigation efforts. This research fills a critical gap in fire modeling literature for northern Iranian forests and advances the understanding of fire dynamics in these ecosystems. This study hypothesizes that integrating GIS with the FARSITE model will improve the precision of fire spread simulations in the study area. We aim to evaluate FARSITE’s applicability in modeling historical fire events, map regional vegetation types, and develop fuel models tailored to local conditions. The hypotheses are: FARSITE simulations will accurately replicate over 50% of the observed fire spread patterns in the Darisan Forest. Incorporating local environmental data, including vegetation and climate, will significantly improve model accuracy. The developed fuel models will effectively represent fire behavior across different vegetation types. This study seeks to answer the following research questions: • How accurately does FARSITE simulate fire spread in the Darisan wildfire event? • What key factors drive fire propagation in this forest ecosystem? • How can GIS integration enhance fire spread prediction and support wildfire management? This study examines the effectiveness of the FARSITE model in simulating wildfire spread during the Darisan event, identifies the primary factors influencing fire propagation, and evaluates the contribution of GIS in enhancing fire prediction and management. The objectives are: Evaluating FARSITE’s accuracy in predicting fire spread. Identifying environmental and topographical factors influencing wildfire behavior. Enhancing fire modeling through GIS-based spatial analysis. By achieving these goals, this research aims to advance wildfire modeling and improve understanding of fire dynamics in complex forest ecosystems. Case study The study area (Darishan forest) is in Dohezar forest area in Northwestern of Iran (Fig. 1 ). This study focuses on the Darishan Forest within the Dohezar region (UTM Zone 39N). The selection of the case study area is based on two primary factors. First, the Darishan wildfire was the largest and most destructive fire event in local history, making it a critical case study for understanding fire spread dynamics and validating simulation models. Comprehensive fire records are available, including ignition locations, burned area perimeters, and wildfire weather data, Sourced from regional natural resources and watershed management organization of province. Secondly, the study area is located within the Hyrcanian Forests in northwestern Iran, known for their distinctive ecological features harboring significant vegetation species such as Fagus orientalis, Carpinus betulus, Quercus castanaefolia, Alnus subcordata, Acer insign, Tilia begonifolia, Buxus hyrcana, Juglans regia, and Parrotia persica. This region boasts a rich diversity of flora, contributing to high biodiversity and significant conservation value and providing critical habitat for rare and endemic species and playing vital roles in climate regulation, soil conservation, and water resource management. The mix of varied vegetation types, substantial fuel loads, and intricate topography highlights the need for localized fire spread simulations. With the growing challenges of deforestation and climate change, studying wildfire behavior in this region is essential for refining fire management strategies, improving predictive models, and promoting sustainable forest conservation efforts. Methodology Among the available fire simulation models, the Fire Area Simulator (FARSITE) has been widely adopted for predicting fire behavior in diverse ecosystems (Finney 1998 , 2002). FARSITE integrates key environmental variables, including fuel models, weather conditions, and topographical data, to generate spatiotemporal fire spread projections. Previous applications of FARSITE have demonstrated its effectiveness in simulating fire spread under various landscape conditions, including Mediterranean forests (Salis 2007 ) and mountainous terrain (Nyatondo 2010 ). In this study, we employ FARSITE to model fire spread behavior in the Darishan area of the Dohezar and Sehezar forests region, incorporating detailed geospatial datasets through Geographic Information Systems (GIS). The integration of GIS allows for the precise mapping of fuel types, vegetation distribution, and terrain features, thereby enhancing the accuracy of fire simulations. By leveraging high-resolution spatial data, we aim to assess the extent to which FARSITE can accurately reconstruct the fire spread patterns observed in historical wildfire event in study area. Mathematical Basis for Fire Spread Modeling Fire modeling relies on fundamental mathematical equations that describe the energy released during combustion (Albini 1979 ). Fire spreads as energy from the flame’s transfers to adjacent flammable materials, raising their temperature until ignition occurs. Accurately modeling this energy transfer enhances the ability to predict the spatial and temporal progression of fire (Rothermel 1991 ). The most important mathematical equations used in fire modeling focus on the energy generated during a fire (Albini 1979 ). The process of fire spread involves the release of energy from the flames and the transfer of part of this energy to adjacent flammable materials, which increases their temperature until they reach the ignition point and begin to burn (Albini 1979 ). Modeling the flow of energy enhances the ability to predict the propagation and spatio-temporal evolution of fire (Rothermel 1991 ). Many field studies and observations generally support the elliptical shape of fire propagation (Richard 1995 ). For simulation purposes, it is suggested to consider points or vertices at regular intervals around the fire perimeter, each representing a new small source from which the fire begins to spread. These two-dimensional vertices, defined by X and Y coordinates, are considered as the peripheral points of fire propagation (Richard 1995 ). This fire growth simulation model, grounded in Huygens' principle of wave propagation, suggests that environmental conditions are calculated individually for each vertex on the fire's perimeter. The elliptical shape and direction of the fire front are then determined according to the time taken for each calculation step. Richard developed formulas based on Huygens’ principle to model fire propagation, which can be incorporated into the FARSITE simulation (Richard 1995 ). These formulas describe the elliptical wave of the fire front using vertices along the fire perimeter. Huygens' principle posits that each vertex acts as the origin of an expanding oval, growing independently. The data required for each vertex includes: a) The position of each vertex on the fire front, denoted by xsx_sxs​ and ysy_sys​ b) The maximum extent of fire spread (θ), determined by the slope vector and wind direction c) The elliptical shape of the fire, primarily determined by the dominant conditions at a given vertex, specified by the parameters aaa, bbb, and ccc. The equation used in the FARSITE model to calculate fire spread rate was developed by Rothermel (Rothermel 1979). It computes the rate of fire spread (RRR) in meters per minute for each vertex on a plane parallel to the Earth's surface. \(\:R=\frac{{I}_{R}\xi\:\left(1+{{\Phi\:}}_{W}+\:{{\Phi\:}}_{S}\right)\:}{{\rho\:}_{b}\epsilon\:{\mathcal{Q}}_{ig}}\) Eq. (1) In the above equation, III represents the intensity of the reaction (kJ/min/m²), ϕ\phiϕ is the flux ratio of propagation, ρ\rhoρ is the volume of dry material (kg/m³), NNN is the number of factors affecting heat generation, TignitionT_{\text{ignition}}Tignition​ is the temperature before ignition (kJ/kg), and θ\thetaθ and vvv are the slope and wind speed, respectively. Parameters Affecting the FARSITE Model The Process of Implementing the FARSITE Model Step 1: Preparing Spatial Data by ArcGIS To implement the fire spread model, the first step is preparing spatial data. ArcGIS software was used to process and prepare the necessary spatial data. These data layers were subsequently transformed into ASCII format for integration with the simulation model. Step 2: Creating the Landscape Layer by combination of GIS The converted spatial layers were imported into the model and integrated to generate a comprehensive "landscape" layer. This layer included essential spatial characteristics of the study area, such as topography, vegetation, and fuel properties. Step 3: Inputting Meteorological Data Meteorological data was collected from the Synoptic Station near the study area at 3-hour intervals. These data were then incorporated into the model to account for weather conditions influencing fire behavior. Step 4: Sensitivity Analysis and Fire Simulation A sensitivity analysis was conducted to assess the model’s response to key parameters. Using FARSITE version 4.1.0, fire spread was simulated based on spatial and meteorological data. Step 5: validity of model Finally, for validity of model, The Kappa index is calculated. Input Data required Simulations in the FARSITE model require a set of spatial data, which are classified into three main environmental factors that influence fire behavior: Ecological factors /topography Weather conditions Vegetation These factors are often referred to as the "fire environment triangle." Ecological factors/ Topography Ecological factors, including topographic criteria such as slope, geographic aspect, and elevation, play an essential role in fire risk assessment. Topography can significantly impact vegetation composition and distribution, microclimate, soil moisture, and, ultimately, fire behavior (Syphard 2008). The relationship between altitude and fire risk is influenced by temperature, humidity, and vegetation. As the altitude increases, the risk of fire tends to decrease. In contrast, steeper slopes can affect soil moisture content and heighten fire risk. In dense areas, the slope of the terrain brings the flames closer to the ground, thereby increasing the likelihood of fire (Whelan 1995). Areas receiving more sunlight are typically more prone to intense fires (Franklin et al. 2000). Topography exhibits significant spatial variation but is treated as constant over time in simulations. It is defined by three distinct layers: height, slope, and slope direction. Height refers to the altitude above sea level and is required to calculate the burned surface during the simulation process. Slope and slope direction are used to calculate fire spread, flame angle, and adjust the fire spread surface on a horizontal plane. Topographic data, including height, slope, and direction, were derived from 1:25,000 scale maps supplied by the Iran National Cartographic Center in raster format. The data were then used to create a digital elevation model (DEM), which provided elevation, slope, and aspect data for the simulation. These data were then converted to ASCII format using ArcGIS software to prepare them for input into the FARSITE model. Weather condition For weather data, climate statistics from the station with the most similar atmospheric conditions to the study area were used. included key meteorological variables, such as temperature, relative humidity, wind speed, wind direction, and precipitation. Temperature, relative humidity, and precipitation data from a specified location were incorporated into the FARSITE model using a weather stream file. To maintain consistency and accuracy, climate statistics were obtained from the synoptic station that most closely matched the atmospheric conditions of the study area. Of the available stations, the Ramsar synoptic station, located 25 km west of the study area's boundary, was identified as the most representative. Its 3-hour interval data formed the basis for the weather inputs in the simulation. This dataset included crucial meteorological variables such as temperature, relative humidity, wind speed, wind direction, and rainfall. Vegetation Vegetation acts as a flammable fuel bed, driving fire spread and significantly influencing fire characteristics, including intensity, spread rate, and flame height (Pyne 1996). Among environmental factors, vegetation is the only one that can directly control fire behavior through management operations. The characteristics of vegetation can be categorized into several types, including fuel type, vegetation canopy, and forest mass height. Vegetation Canopy A canopy map of the area is essential for calculating shading and wind speed reduction factors for all fuel models. The vegetation canopy refers to the horizontal area of the ground surface directly covered by tree crowns. Canopy cover maps of the region, which are vital for assessing the extent of shading on the forest floor, were acquired. The vegetation canopy represents the horizontal area of the ground surface directly covered by the tree crowns. Canopy cover maps of the region, which are crucial for determining the level of shading on the forest floor, were obtained. These maps are crucial for calculating the moisture content of combustible materials and for models that predict fire behavior. Furthermore, canopy cover affects the amount of radiant energy the forest receives. The canopy cover maps were classified based on data provided by the Forests and Rangelands, and Watershed Management Organization of Iran, and were then converted into a format compatible with the FARSITE model. Fuel Model Fuel is one of the primary elements in the burning process, whether the fire is at the ignition or spread stage. Mathematical equations and predictive systems designed to model fire behavior require data related to the burning area surface and fuel bed. This data includes fuel load, bulk density, fuel particle size, heat content, and moisture content limitations (Scott 2005 ). To simplify the use of these factors in models and systems, they are presented as a set of fuel models. Rothermel ( 1972 ) defined a fuel model as a complete set of inputs used to define a flammable bed in a mathematical model of fire spread (Rothermel 1979). He described the fuel model as a tool that allows users to predict fire behavior and patterns with greater accuracy (Anderson 1982). At the first level, fuel models are classified into base and initial groups based on cover type, with seven distinct groups. This classification is like the one proposed by Anderson in 1982 (Anderson 1982). To select a standard fuel model, the models developed by Scott and Burgan were utilized. This updated set of fuel models, designed for fire behavior analysis, is compatible with any simulation system based on Rothermel's (1972) fire spread model (Rothermel, 1972 ). Since fuel properties are typically complex, fire managers have attempted to simplify them by categorizing the physical parameters and spatial distribution of fuel into different classes, known as ‘fuel models’ (Anderson 1982; Burgan and Rothermel 1984). These fuel characteristics' spatial distribution is often represented as fuel type maps. Various classification systems are used to group vegetation types based on their fire behavior. Fifty-three standard fuel models have been developed for the Rothermel ( 1972 ) fire spread model. According to Nyatondo ( 2010 ), these include the original 13 fuel models described by Anderson (1982) and the additional 40 models outlined by Scott and Burgan ( 2005 ). In this study, the selection of appropriate fuel models was based on the similarities between the vegetation characteristics observed in the field and the descriptions of the standard fuel models (Anderson 1982; Scott and Burgan 2005 ) as shown in Table 1 . The vegetation cover types identified during fieldwork were reclassified according to the selected fuel models, reflecting the fuel conditions observed on-site. Table 1 Description of the standard fuel models used for simulation in the wildfire case studies from 53 standard fuel models (Anderson 1982; Scott and Burgan 2005 ) Observed vegetation Fire carrying fuel type, model name and code Fuel model number Fuel model description River Insufficient wildland fuel to carry wildland fire under any condition (non-burnable, NB8) 98 Land covered by open bodies of water such as lakes, rivers and oceans Grass (Low density) Humid-climate grass (Grass, GR5) 105 Grass and herb fuel load is light; fuel bed depth is about 1 to 2 feet. Grass (Medium density) Continuous humid-climate grass (Grass, GR6) 106 Load is greater than GR5 but depth is about the same. Grass-shrub Grass and shrubs combined (Grass-Shrub, GS3) 123 Moderate grass/shrub load, average grass/shrub depth less than 2 feet Natural Forest (Medium density and timber-shrub) Moderate litter load with shrub component (Timber-Understory, TU2) 162 High extinction moisture. Spread rate is moderate; flame length low Natural forest/Mixed Forest (Medium density and timber-grass shrub) Grass and shrub mixed with litter from forest canopy (Timber Understory, TU3) 163 Extinction moisture is high. Spread rate is high; flame length moderate Hardwood plantation (Medium density) Moderate load broadleaf litter (Timber Litter, TL6) 186 Less compact than TL2. Spread rate is moderate; flame length low Hardwood plantation (High density) Very high load, fluffy broadleaf litter (Timber Litter, TL9) 189 Spread rate is moderate; flame length moderate. In the final step of evaluating the simulation results, the Kappa index was employed to assess the model's validity. The Kappa index is calculated using the following matrices and equations (Eq. 2). \(\:\text{K}=\frac{\sum\:_{\text{i}=1}^{\text{n}}{\text{P}}_{\text{i}\text{i}}-\sum\:_{\text{i}=1}^{\text{n}}{\text{q}}_{\text{i}\text{i}}}{1-\sum\:_{\text{i}=1}^{\text{n}}{\text{q}}_{\text{i}\text{i}}}\) Eq. (2) Where Pii ​ is the observed probability and ​ qii is the expected probability. Results and Discussion The study was conducted in the Darishan Forest, located within the Dohezar forest area, in watershed region 33 at latitude 36˚19′22″ N and longitude 50˚29′57″ E. The Hyrcanian Forests, which encompass Darishan, are renowned for their rich biodiversity. In December 2000, the study area was affected by a fire that burned 72.63 hectares, and this event was used to simulate fire behavior in the region. Data that is used in study area Topography: Topographic data, including elevation, slope, and aspect, were extracted from 1:25,000 scale maps provided by the Iran National Cartographic Center in raster format. Figures 2 , 3 , and 4 illustrate the slope, elevation, and aspect of the study area. The data were then used to create a digital elevation model (DEM), which supplied the elevation, slope, and aspect data for the simulation. The data were subsequently converted to ASCII format using ArcGIS software to prepare them for input into the FARSITE model. Weather data: Table 2 displays the 3-hour intervals of temperature, rainfall, wind speed, and wind direction on the day of the fire (December 8, 2000), as recorded in the study area. The average annual rainfall at different elevations ranges from a maximum of 1,255 mm to a minimum of 645 mm. The mean annual temperature varies between 10.3°C and 16.1°C, depending on the specific location. At lower elevations, the average minimum and maximum monthly temperatures are 3°C and 26.5°C, respectively. In contrast, at higher elevations, based on the thermal gradient, the minimum and maximum temperatures are − 20.5°C and 6.3°C, respectively. Due to the region’s topographical features and its classification under the Amberger climate system, the climate is categorized as cold, humid, and mountainous. Table 2 the 3-hour intervals of temperature, rain, wind speed, and wind direction on the day of the fire Time Temperature (degrees Celsius) Rain intensity Wind speed Wind direction 3:30 8.8 0 0 - 6:30 7.2 0 3 SW 9:30 13.2 0 2 SE 12:30 14.4 0 5 NE 15:30 14.0 0 2 SE 18:30 12.4 0 1 SW 21:30 12.1 0 2 SW 00:30 13.0 0 2 NW Vegetation Canopy density In this study, the selection of appropriate fuel models was based on the similarities between the field-observed vegetation characteristics and the descriptions of standard fuel models (Anderson 1982; Scott and Burgan 2005 ), as presented in Table 3 . A map depicting the vegetation density types is shown in Fig. 5 . Table 3 classification of vegetation canopy in study area Vegetation type Vegetation canopy (%) Canopy class Without vegetation 0 0 Shrub lands 1–25 1 Low-density forest 25–50 2 Moderate-density Forest 50–75 3 High-density forest 75–100 4 Fire Spread Barriers The FARSITE model operates under the assumption that fire spread is directly influenced by the availability and type of combustible material. If combustible material is present, the fire will continue to spread, and the model does not inherently account for factors that limit or extinguish the fire. As a result, barriers such as roads, rivers, and firebreaks are incorporated into the model to restrict fire spread, improving the accuracy of the simulation. In this study, roads and rivers were extracted from 1:25,000 maps provided by the Iran National Cartographic Center and prepared for use in the model Fire History: The fire reality map was created using a 25-year history of fire events, covering the period from 1997 to 2023. Information from the Natural Resources, Range, and Watershed Organization of Mazandaran Province in Iran, along with field visits with environmental and forest experts, was used to prepare this map. Additionally, remnants of past fires in the study area were recorded using GPS technology and GIS Arc software. To conduct the field measurements, a team of forest experts, rangers and environmentalists walked for 8 hours through the Dohezar forest to reach the Darishan forest area, where all fire points were recorded with GPS. As we observed, many large, old trees had burned, leaving only half of their trunks behind, while shrubs and grass had grown in their place, clearly indicating the effects of past fires. Based on the field survey, the starting and ending points of the wildfire were identified through expert assessments and observations. The affected area exhibited a triangular shape, with the apex positioned at a lower elevation and the broader base at a higher elevation. Akey factor in containing the fire was a substantial natural barrier, resembling a large rock formation or mountain. This barrier, composed of stone, created a significant gap between the trees, effectively acting as a firebreak. The dimensions of the rock formation were measured by the team, with a width of 10 meters and a height of 40 meters, which played a crucial role in halting the spreading of fires. During our field visit, we also noticed that many shepherds had built small fires in the forest to stay warm but had left them unattended without extinguishing the fire, which could be one of the primary causes of fire ignition in the area. The primary outputs of the fire spread simulation using the FARSITE model are maps showing the fire spread at predetermined time intervals, as well as the final fire spread map. The simulated fire spread map generated by the FARSITE model is presented in Fig. 6 . To evaluate the accuracy of the simulation, the Kappa coefficient and consistency ratio were computed. The consistency ratio represents the proportion of the fire area that has been simulated as burned, while the Kappa coefficient measures the similarity between the observed and simulated fire spots. It accounts for both underestimation and overestimation, ethe fire spread falling below or above the actual area burned. The Kappa coefficient measures the similarity between observed and simulated fire spots, calculating both underestimation and overestimation (i.e., the estimation falling below or above the actual fire spread). The results of FARSITE model, Kappa coefficient, and consistency ratio are shown in Table 4 . Table 4 Results of the FARSITE model, Kappa coefficient, and consistency ratio analysis. Simulated fire are (ha) Observed fire area (ha) Underestimation (ha) Overestimation (ha) Consistency (ha) Kappa 35.78 72.63 37.7 0.85 34.93 0.68 The main outputs of the fire simulation process are fire spread maps at predetermined time intervals and a final fire spread map. To evaluate the accuracy of the simulation, the Kappa coefficient and consistency ratio were used. The consistency ratio measures the proportion of the actual fire-affected area that was accurately simulated as burned, while the Kappa coefficient evaluates the similarity between observed and simulated fire spots, considering both overestimation (excessively predicted burned area) and underestimation (areas missed by the simulation). Table 4 presents the results of the FARSITE simulation. Of the 72.63 hectares of observed burned area, 34.93 hectares were correctly predicted, resulting in a high consistency ratio. The overestimated area was minimal (0.85 ha), but the model underestimated 37.7 hectares, suggesting potential limitations in capturing certain fire spread patterns. A Kappa coefficient of 0.68 reflects a moderate level of correspondence between the observed and simulated fire spread. While the model performs well in estimating fire spread, the high underestimation suggests a need for refining input parameters or integrating additional environmental variables (e.g., real-time moisture conditions) to improve prediction accuracy. Possible reasons for overestimation and observed burned areas Errors in input data (e.g., vegetation, fuel moisture). Limitations of the FARSITE model in specific terrains. Field measures challenges Fuel models which are generated for the USA Lack of information in databases Incorporating such data into fire spread modeling could enhance the accuracy of simulations (Arca et al. 2007; Arroyo et al. 2008; Jahdi 2014). However, due to limited information and the challenges of field-measuring large historical fires, such as the inaccessibility of certain regions and the removal of old burn marks in forest areas, this remains difficult. Conversely, access to documented data on the behavior of larger fires facilitates the testing of more significant changes in fire spread. The fuel models used in this study were developed in the United States and, while they closely align with the conditions in the study area, directly measuring fuel parameters (such as fuel load and moisture content) could yield more accurate models (Jahdi, 2014). One limitation of FARSITE is that it does not account for the effects of the two-dimensional fire shape on acceleration. Instead, it models fire acceleration as a point-source for simplicity (Phillips, 2006). Modifying the scale and type of the fuel model could also enhance the model's ability to simulate fires in this region. Additionally, using a 10-meter Digital Elevation Model (DEM) would provide higher resolution and offer a more accurate representation of the fuel heterogeneity (Philips 2006). This highlights the need to create fuel µodels tailored to the teµperate conditions of northern Iran's forests or Hyrcanian Forest fuel µodel. Developing these custoµized µodels requires adjusting fuel paraµeters based on field observations (Nyatondo 2010 ). In comparison with other research, recent studies have increasingly focused on fire risk assessment and fire spread simulation (Wang 2023; Wu 2022; Burge et al. 2020; Hodges and Lattimer 2019 ; Marjani and Mesgari 2023 ; Qi et al. 2022 ; Kanagaraj et al. 2023; Zigner et al. 2022). Previous research indicates that the accuracy of the FARSITE model can be improved by using fuel models specifically designed for simulating fire spread and behavior in various vegetation types (Fernandes 2009; Salis 2008; Bacciu 2009). Further studies focused on actual fire behavior in the region are necessary to validate and enhance the FARSITE model, with thorough calibration potentially increasing the accuracy of predictions (Jahdi 2014). Given the robust wildfire database in northern forests of Iran, conducting simulations in other regions and comparing results could significantly refine this approach. Additionally, improving wildfire simulations can be achieved by developing local and regional fuel models and collecting more detailed weather data. A comparative analysis with similar FARSITE simulations in Mediterranean climates suggests that such underestimations can result from the challenges in calibrating fuel models and integrating accurate local meteorological data. Studies in these regions have also encountered challenges in accurately modeling fire spread, highlighting the need for more refined fuel models and localized data to enhance predictions. Fire spread models, supported by accuracy assessments, are vital for understanding fire behavior, evaluating associated risks, assisting in active fire suppression, and guiding fuel treatment strategies, such as the creation of fuel breaks (Finney 2006 ; Shinneman et al. 2019). The use of generalized fuel bed maps, such as those provided by LANDFIRE, can lead to discrepancies between simulated fire spread and observed patterns. The model accuracy assessment and alternative methods for selecting Fire Behavior Fuel Models (FBFMs) used in some studies represent advancements over field-vegetation map-based accuracy assessments in models like MTT and FARSITE, which were initially developed for semiarid environments, such as Mediterranean shrublands or Iranian grasslands (Arca et al. 2007; Jahdi et al. 2015 , 2016; Salis et al. 2016 , 2021). However, research by Samuel Price and Matthew J. Germino (2022) suggests that such methods may not be suitable for all site-specific applications. Studies using the FARSITE simulator have demonstrated that areas with a high risk of fire often align with regions that have previously burned, supporting the model's validity. As one of the most widely used fire spread simulators, FARSITE can produce valuable results for Iranian forests, provided that a properly calibrated fuel model is used. The simulator’s user-friendly interface, transparency, and adaptability to different forest fire types increase its usefulness (Eskandari 2013). However, J. Germino's (2022) research suggests that such methods may not be applicable for most site-specific applications. Conclusion This study assesses the suitability of the FARSITE fire behavior simulation model for use in Iranian forests.The simulation results showed good agreement between simulated and observed fire environments, with the spread pattern and fire shape aligning well with the actual fire spot. Although minor overestimation was observed, significant underestimation, especially in the northern region, occurred, likely due to inaccuracies in the fuel model or inadequate meteorological data. Despite this, the overall results were reliable, suggesting that FARSITE can effectively simulate fire spread in the study area. GIS-based fire spread models like FARSITE generate polygonal shapes that closely match natural fire spread patterns and incorporate factors influencing fire behavior, thus enhancing model credibility. The findings provide valuable insights for fire management in the Hyrcanian Forests, supporting decisions on firebreak placement, resource allocation, and the formulation of prevention strategies, such as controlled burns. To improve future simulations, integrating high-resolution remote sensing data, refining fuel models to better reflect local vegetation, and using more accurate meteorological data are recommended. These improvements will enhance fire risk management not only in Iran but also in similar forest ecosystems worldwide. This research contributes to the advancement of wildfire management strategies, ecological conservation, and fire simulation model development, providing valuable guidance to policymakers, forestry officials, and conservation agencies in implementing effective fire prevention and mitigation strategies specifically tailored to Iran's forest landscapes. References Adab H, Kanniah KD, Solaimani K (2013) Modeling Forest fire risk in the northeast of Iran using remote sensing and GIS techniques. Natural Hazards 65:1723-1743. Ajin RS, Ana-Maria L, Jacob MK, Vinod PG, Krishnamurthy RR (2016) The Risk Assessment Study of Potential Forest Fire in Idukki Wildlife Sanctuary using RS and GIS Techniques. Int J Adv Earth Sci Engin 5:308-318. Aliani H, Babaie Kafaky S. Saffari A. Monavari S. M (2016) Determining an appropriate method for the purpose of land allocation for ecotourism development (case study: Taleghan County, Iran). Environ Monit Assess .188:651 DOI 10.1007/s10661-016-5651-2. Albini FA (1979) Spot fire distance from burning trees: a predictive model. USDA Forest Service, General Technical Report INT, 56 pp. Alonso-Betanzos A, Fontenla-Romero O, Guijarro-Berdiñas B, Hernández-Pereira E, Canda E, Jimenez J, Legido JL, Muñiz S, Paz-Andrade C, Paz-Andrade ML (2002) A neural network approach for forest fire risk estimation. Presented at the 15th 2nd Int Conf on Artif Intell, ECAI' Lyon, France Anderson DH, Catchpole EA, De Master NJ, Parkes T (1982) Modeling the spread of grass fires. J Austral Math Soc (Series B) 23:451-466 Bowman DMJS, French BJ, Williamson GJ, Prior LD (2021) Fire, herbivores and the management of temperate Eucalyptus savanna in Tasmania: Introducing the Beaufront fre – mammalian herbivore feld experiment. Ecol Manag Restor 22:140–151 Bui DTB, Pradhan O, Lofman I, Revhaug O, Dick B (2012) Landslide susceptibility mapping at Hoa Binh province (Vietnam) using an adaptive neuro-fuzzy inference system and GIS. Comput Geosci 45:199-211 Chandra S (2005) Application of Remote Sensing and GIS Technology of Forest Fires: A Case Study in the Garhwal Himalayan Region. In: Geo-Inform Disaster Manag, 1239-1254 Chandra S, Aurora MK (2006) Forest fire risk zonation mapping using remote sensing technology (proceedings paper). J SPIE (abstract), 1-2 Chuvieco E, Kasischke ES (2007) Remote sensing information for fire management and fire effects assessment. J Geophys Res: Biogeosci 112:1-8. https://doi.org/10.1029/2006JG000230 Cruz MG, Alexander ME (2013) Uncertainty associated with model predictions of surface and crown fire rates of spread. Environ Modell Softw 47:16-28 Eskandari S, Oladi Ghadikolaei J, Jalilvand H, Saradjian S (2013) Detection of Fire High-Risk Areas in Northern Forests of Iran Using Dong Model. World Appl Sci J 27(6):770-773 Finney M (1998) FARSITE: Fire Area Simulator-model development and evaluation. USDA For Serv Res Pap RMRS-RP-4. Finney MA (1998) Efforts at comparing simulated and observed fire growth patterns. Final Report INT-95066-RJVA. Systems for Environmental Management, Missoula, MT, USA. Finney MA (2006) An Overview of FlamMap, Fire Modeling Capabilities. USDA For Serv Proc, pp 213-220 Finney MA, Bradshaw L, Butler BW (2006) Modeling surface winds in complex terrain for wildland fire incident support. Final Report for JFSP Funded Project #03-2-1-04. U.S. Department of Agriculture, Forest Service, Fort Collins, CO, USA, pp 1-10. Finney MA, Grenfell IC, McHugh CW, Seli RC, Trethewey D, Stratton RD, Brittain S (2011) A method for ensemble wildland fire simulation. Environ Model Assess 16:153-167 Ghobadi GJ, Gholizadeh B, Dashliburun OM (2012) Forest fire risk zone mapping from geographic information system in Northern Forests of Iran (Case study, Golestan Province). Int J Agric Crop Sci 4(12):818-824 Goleiji E, Hosseini SM, Khorasani N, Monavari SM (2017) Forest fire risk assessment-an integrated approach based on multicriteria evaluation. Environ Monit Assess. https://doi.org/10.1007/s10661-017-6225-7, pp 189-612 Hodges JL, Lattimer BY (2019) Wildland Fire Spread Modeling Using Convolutional Neural Networks. Fire Technol 1–28. https://doi.org/10.1007/s10694-019-00846-4 Jahdi R, Salis M, Darvishsefat A, Mostafavi A, Alcasena F, Etemad V, Lozano O, Spano D (2015) Calibration of FARSITE simulator in northern Iranian forests. Nat Hazards Earth Syst Sci 15:443-459 Jain P, Coogan SCP, Subramanian SG, Crowley M, Taylor SW, Flannigan MD (2020) A review of machine learning applications in wildfre science and management. Environ Rev 0(ja). https://doi.org/10.1139/er-2020-0019 John B, Matthew B, Matthias I et al (2020) Convolutional LSTM Neural Networks for Modeling Wildland Fire Dynamics. https://doi.org/10.48550/arXiv.2012.06679 Jordan MI, Mitchell T (2015) Machine learning: Trends, perspectives, and prospects. Sci 349:255–260. Keane RE, Burgan R, van Wagtendonk J (2001) Mapping wild land fuels for fire management across multiple scales: integrating remote sensing, GIS, and biophysical modeling. Int J Wildland Fire 10:301-319 Kevin T, Brett S, Derek C (2008) Phoenix: development and application of a bushfire risk management tool. Aust J Emerg Manag 23:47-54. https://doi.org/10.1007/978-3-540-74492-4_21 Li Z, Huang Y, Li X, Xu L (2020) Wildland fre burned areas prediction using long short-term memory neural network with attention mechanism. Fire Technol 57:1–23 Liu S, Cao L, Lin C, Dai Y, Li X, Li S, Sun S, Li D (2024) Fire spread prediction model based on multi scale convolutional neural network. Multimed Tools Appl. https://doi.org/10.1007/s11042-024-20400-x. Marjani M, Mesgari MS (2023) The large-scale wildfre spread prediction using a multi-kernel convolutional neural network[J]. ISPRS Ann Photogram Remote Sens Spatial Inform Sci 10:483–488 Marozas V, Racinskas J, Bartkevicius E (2007) Dynamics of ground vegetation after surface fires in hemi boreal Pinus sylvestris forests. For Ecol Manage 250(1-2):47-55 Marshall G, Thompson DK, Anderson K, Simpson B, Linn R, Schroeder D (2020) The impact of fuel treatments on wildfire behavior in north American boreal fuels: a simulation study using FIRE[1]TEC. Fire-Switz 3(2):18 Mirdeylami T, Shataee Sh, Kavousi MR (2014) Forest fire risk zone mapping in the Golestan national park using weighted linear combination (WLC) method. Iran J For 4(4):377-390 (In Persian) Mohammadi F, Shabanian N, Pourhashemi M, Fatehi P (2010) Risk zone mapping of forest fire using GIS and AHP in a part of Paveh Forests. Iran J For Poplar Res 18(4):569-586 Nasanbat E, Lkhamjav O (2016) Wild Fire Risk Map In The Eastern Steppe Of Mongolia Using Spatial Multi-Criteria Analysis. Int Arch Photogramm Remote Sens Spatial Inf Sci, pp 1-9 Nyatondo UN (2010) Fire spread modeling in Majella National Parks, Italy. Master of Science, Int Inst Geo-Inform Sci Earth Obs, Enschede, the Netherlands, 91 p Paz Sh, Carmel Y, Jahshan F, Shoshany M (2011) Post-fire analysis of pre-fire mapping of fire risk: A recent case study from Mt. Carmel. For Ecol Manage 262:1184-1188 Price S, Germino MJ (2022) Modeling of fire spread in sagebrush steppe using FARSITE: an approach to improving input data and simulation accuracy. Fire Ecol. https://doi.org/10.1186/s42408-022-00147-2. Pyne SJ, Andrrew PL, Laven AR (1996) Introduction to wildfire. 2nd ed. John Wiley and Sons, New York, 769 p Qi P, Chang J, Chen X et al (2022) Identifcation of Rock Properties of Rock Wall Cut by Road header Based on PSO-VMD-LSSVM[J]. Front Earth Sci 10:884633 Radke D, Hessler A, Ellsworth D (2019) FireCast: Leveraging Deep Learning to Predict Wildfire Spread. International Joint Conference on Artifcial Intelligence. https://doi.org/10.24963/ijcai. 2019/636. Richard GD (1995) A general mathematical framework for modeling two-dimensional wildland fire spread. Int J Wildland Fire 5:63-72 Rothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. USDA For Serv Intermountain For Range Exp Stn Res Pap INT-115 Rothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. Research Paper INT-115. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT, USA. Rothermel RC (1991) Predicting behavior and size of crown fires in the Northern Rocky Mountains. USDA For Serv Res Pap INT-438, Ogden, Utah, USA, 67 p Rwanga SS, Ndambuki JM (2014) Application of geographical information systems and FARSITE in fire spread modeling. Int J Environ Sustain Dev 13(2):185-203 Salis M (2007) Fire Behavior simulation in Mediterranean Maquis using FARSITE (Fire Area Simulator). PhD Doctoral Thesis, Univ Degli Studi Di Sassari Salis M, Arca B, Alcasena F, Arianoutsou M, Bacciu V, Duce P, Duguy B, Koutsias N, Mallinis G, Mitsopoulos I, et al. (2016) Predicting wildfire spread and behaviour in Mediterranean landscapes. Int J Wildland Fire 25:1015-1032. Scott JH, Burgan RE (2005) Standard fire behavior fuel models: A comprehensive set for use with Rothermel's surface fire spread model. USDA Gen Tech Rep RMRS-GTR-153 Stocks BJ, Alexander ME, Lanoville RA (2004) Overview of the International Crown Fire Modeling Experiment (ICFME). Can J For Res 34:1543–1547 Wang Z (1992) General forest fre risk rating system. J Natl Disast 03:39–44. Yang S, Li N (2021) Research progress of forest fire spread model. Gansu Sci Technol 37(03):45-47 You X, Zheng Z, Yang K et al (2023) A PSO-CNN-Based Deep Learning Model for Predicting Forest Fire Risk on a National Scale[J]. Forests 15(1):86 Zhang ZX, Zhang HY, Zhou DW (2009) Using GIS spatial analysis and logistic regression to predict the probabilities of human-caused grassland fires. J Arid Environ 74:386-393 Zigner K, Carvalho LMV, Peterson S, Fujioka F, Duine GJ, Jones C, Roberts D, Moritz M (2020) Evaluating the ability of FARSITE to simulate wildfires influenced by extreme, downslope winds in Santa Barbara, California. Fire MDPI. Cite Share Download PDF Status: Published Journal Publication published 18 Jun, 2025 Read the published version in Natural Hazards → Version 1 posted Reviewers agreed at journal 07 Apr, 2025 Reviewers invited by journal 07 Apr, 2025 Editor assigned by journal 04 Apr, 2025 First submitted to journal 03 Apr, 2025 Editorial decision: Minor revisions 29 Jan, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5764687","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":439363384,"identity":"a2b89623-4ff9-4b31-a7e8-d2f26ef49372","order_by":0,"name":"Elham Goleiji","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYLCCBwwSzGzsDUCWgQWRWhJAWngOgLRIEK0FCCQgJGHV/AzsFx8kVFiw80k+v7rhR4EEA397dwJeLZINPMUGCWeADpPOKbvZA3SYxJmzG/BqMTjAkyaR2AbWknaDB6jFQCIXvxZ7sJZ/QC2SZ9Ju/iFGiwED+zGJxAagFgn2Y7eJskXiMA+zQcIxUCDnsN2WMZDgIegX/vb2hw8+1NQly7cff3bzzR8bOf72XvxaGJh5DEBUMgMDhMGDXzkYsD8AkXYwxigYBaNgFIwCDAAAB2w92xOtJoMAAAAASUVORK5CYII=","orcid":"https://orcid.org/0009-0007-8245-0863","institution":"The University of British Columbia","correspondingAuthor":true,"prefix":"","firstName":"Elham","middleName":"","lastName":"Goleiji","suffix":""},{"id":439363385,"identity":"d8f3310b-d9e7-479d-b46f-2e33388cd656","order_by":1,"name":"Hamide Aliani","email":"","orcid":"","institution":"Tehran Municipality","correspondingAuthor":false,"prefix":"","firstName":"Hamide","middleName":"","lastName":"Aliani","suffix":""},{"id":439363386,"identity":"1455bdee-fe74-4c5b-b17e-bb203bea76bf","order_by":2,"name":"Seyed Mohsen Hosseini","email":"","orcid":"","institution":"Tarbiat Modares University","correspondingAuthor":false,"prefix":"","firstName":"Seyed","middleName":"Mohsen","lastName":"Hosseini","suffix":""}],"badges":[],"createdAt":"2025-01-04 16:56:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5764687/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5764687/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11069-025-07404-y","type":"published","date":"2025-06-18T15:57:48+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":80283914,"identity":"75ca7fad-b890-48c4-9f89-50b41b63e1e2","added_by":"auto","created_at":"2025-04-10 06:25:27","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":161058,"visible":true,"origin":"","legend":"\u003cp\u003eLocation of study area (Darishan forest)\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/0dc2abb6ba5ec5b05ca7f87b.jpeg"},{"id":80283897,"identity":"c68a42a6-5b77-432e-be19-e186700fff46","added_by":"auto","created_at":"2025-04-10 06:25:26","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":148694,"visible":true,"origin":"","legend":"\u003cp\u003emap of slope of study area\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/47e0f20182a3520e18d722e1.jpeg"},{"id":80283895,"identity":"28b999d5-961c-46a8-bc93-944730acda27","added_by":"auto","created_at":"2025-04-10 06:25:26","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":76781,"visible":true,"origin":"","legend":"\u003cp\u003emap of elevation of study area\u003c/p\u003e","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/58b42e3e37e8eb72861a73ec.png"},{"id":80283894,"identity":"5fee1787-0c4e-4ff5-bda3-04564f44e171","added_by":"auto","created_at":"2025-04-10 06:25:26","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":264146,"visible":true,"origin":"","legend":"\u003cp\u003eAspect map of the study area\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/fb3b98c84f5535a944a02f6c.jpeg"},{"id":80283925,"identity":"da64dc41-0cbc-4cab-876b-0e26aa73b428","added_by":"auto","created_at":"2025-04-10 06:25:27","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":184989,"visible":true,"origin":"","legend":"\u003cp\u003eMap of vegetation density types in the study area.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/dd1eae7176be7813d6ec1f1a.jpeg"},{"id":80284471,"identity":"7adfead6-fc95-423c-9e59-3b733bfff3b1","added_by":"auto","created_at":"2025-04-10 06:33:26","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1043120,"visible":true,"origin":"","legend":"\u003cp\u003eMap of the simulated fire spread using FARSITE model\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/b702bb69e822a07bac826d05.png"},{"id":80284470,"identity":"48dd196d-96d0-4080-b250-4a1a611d5322","added_by":"auto","created_at":"2025-04-10 06:33:26","extension":"jpeg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":264146,"visible":true,"origin":"","legend":"\u003cp\u003eAspect map of the study area\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/9afb431421cb4a8869240f87.jpeg"},{"id":80283919,"identity":"275c5726-f1e1-4a72-b773-1bc0b2d2a28d","added_by":"auto","created_at":"2025-04-10 06:25:27","extension":"jpeg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":184989,"visible":true,"origin":"","legend":"\u003cp\u003eMap of vegetation density types in the study area.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/cd1acb5e66eb39c9a3ccc814.jpeg"},{"id":80283918,"identity":"1c6a1dd2-330e-4cf4-ac42-67ec02d545c9","added_by":"auto","created_at":"2025-04-10 06:25:27","extension":"jpeg","order_by":16,"title":"Figure 16","display":"","copyAsset":false,"role":"figure","size":161058,"visible":true,"origin":"","legend":"\u003cp\u003eLocation of study area (Darishan forest)\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/d9b6966e6db1dac7e006c39f.jpeg"},{"id":80283909,"identity":"fbfa1eb4-cd31-4a27-b43e-507c4f008a8f","added_by":"auto","created_at":"2025-04-10 06:25:26","extension":"jpeg","order_by":17,"title":"Figure 17","display":"","copyAsset":false,"role":"figure","size":148694,"visible":true,"origin":"","legend":"\u003cp\u003emap of slope of study area\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/f4c371b7c55cdef13597a708.jpeg"},{"id":80284474,"identity":"d28b5446-f7e3-49c6-9396-48532f48aeee","added_by":"auto","created_at":"2025-04-10 06:33:27","extension":"png","order_by":29,"title":"Figure 29","display":"","copyAsset":false,"role":"figure","size":76781,"visible":true,"origin":"","legend":"\u003cp\u003emap of elevation of study area\u003c/p\u003e","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/3f154611d1b2ac010e934f03.png"},{"id":85231420,"identity":"cc26ed04-789a-4ad0-9ed7-0233a5537c44","added_by":"auto","created_at":"2025-06-23 16:07:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3903982,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5764687/v1/da44731f-5fd4-4d5a-9c05-1b36da617cea.pdf"}],"financialInterests":"","formattedTitle":"Integrating FARSITE and GIS for Enhanced Forest Fire Spread Prediction and Simulation","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWildfires represent a major environmental hazard, particularly in forested and grassland ecosystems, where they can cause extensive ecological and economic damage. The occurrence and intensity of wildfires are shaped by multiple factors, such as climate fluctuations, fuel properties, and terrain conditions. (Wang 2023; Ajin 2016; Mohammadi 2010; Goleiji 2017; Zigner 2020; Mirdeylami 2014). Forests, as one of the most vital renewable natural resources, Play a vital role in preserving biodiversity, regulating carbon cycles, and stabilizing regional climate patterns. (Zhang 2009; Jahdi 2015; Ghobadi 2012; Aliani 2016). However, wildfires pose a significant threat to these ecosystems, making fire modeling and predictive analysis essential for effective wildfire management (Nasanbat \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Marozas 2007; Paz 2011; Adab 2013).\u003c/p\u003e \u003cp\u003eThe modeling of fire spread dynamics is an essential component of wildfire risk assessment, as it enables the prediction of fire behavior under varying environmental conditions. The spread of fire is determined by a range of interconnected factors, such as vegetation type, fuel load, weather conditions, and topographical features. To aid in the analysis of fire propagation, several simulation models have been created, encompassing physical, empirical, and machine learning-based methods such as FIRESTAR (Cruz and Alexander \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), FIRETEC (Marshall et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and the Phoenix RapidFire model (Kevin et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2008\u003c/span\u003e) Depend on the core principles of combustion and heat transfer. Empirical models, including the widely used Rothermel model (Rothermel \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1972\u003c/span\u003e), the McArthur model in Australia (Stocks et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2004\u003c/span\u003e), and the Canadian Fire Danger Rating System (CFFDRS) (Bowman et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), establish relationships between fire behavior and environmental factors based on observational data.\u003c/p\u003e \u003cp\u003eIn recent years the use of machine learning techniques has transformed fire prediction and risk assessment across various scientific disciplines. (Hegde and Rokseth 2020; Chen et al. 2019; Jain et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Jordan and Mitchell \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Notably, convolutional neural networks (CNNs) have emerged as effective tools for forecasting the future status of fire-affected areas (Li et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Radke et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The development of advanced machine learning models, such as the Convolutional Long Short-Term Memory (ConvLSTM) networks, has demonstrated superior performance in predicting fire behavior compared to traditional methods (Burge et al. 2020).\u003c/p\u003e \u003cp\u003eFire simulation models, including FARSITE and Minimum Travel Time (MTT; Finney \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e1998\u003c/span\u003e, 2002), are essential for assessing and managing fire risks (Keane 2001). Numerous studies have focused on fire risk assessment and fire spread simulation, highlighting the importance of accurate fire behavior modeling for effective fire management strategies.\u003c/p\u003e \u003cp\u003eIn contrast to North America and Europe, countries like Iran frequently lack detailed data on fire history, fuel characteristics, and environmental conditions, which can affect the precision of fire modeling systems. The reliability of these models hinges on three critical factors: the quality of input data, the theoretical framework of fire behavior models, and the efficiency of fire spread algorithms. (WenBin and Perera 2008; Jahdi 2015).\u003c/p\u003e \u003cp\u003eFARSITE is a widely used fire spread simulation tool, but its effectiveness in Iran remains underexplored. While studies like Salis (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2007\u003c/span\u003e), Nyatondo (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), and Jahdi (2015) have validated its use in Mediterranean and montane regions, its application in Iranian forests lacks thorough validation. This study integrates FARSITE with GIS to enhance localized fire spread predictions, leveraging environmental data for more precise and context-specific modeling.\u003c/p\u003e \u003cp\u003eWhile FARSITE has been widely used for fire spread simulations, its application to the unique environmental conditions of Iranian forests remains limited. This study integrates GIS techniques with FARSITE to assess its accuracy under varying conditions, identifying strengths and limitations. By developing detailed vegetation maps and fuel models, we aim to improve fire spread predictions and management strategies. Additionally, the findings will offer valuable insights for policymakers and fire management agencies, contributing to enhanced fire prevention and mitigation efforts. This research fills a critical gap in fire modeling literature for northern Iranian forests and advances the understanding of fire dynamics in these ecosystems.\u003c/p\u003e \u003cp\u003eThis study hypothesizes that integrating GIS with the FARSITE model will improve the precision of fire spread simulations in the study area. We aim to evaluate FARSITE\u0026rsquo;s applicability in modeling historical fire events, map regional vegetation types, and develop fuel models tailored to local conditions. The hypotheses are:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eFARSITE simulations will accurately replicate over 50% of the observed fire spread patterns in the Darisan Forest.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eIncorporating local environmental data, including vegetation and climate, will significantly improve model accuracy.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eThe developed fuel models will effectively represent fire behavior across different vegetation types.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThis study seeks to answer the following research questions:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e\u0026bull; How accurately does FARSITE simulate fire spread in the Darisan wildfire event?\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e\u0026bull; What key factors drive fire propagation in this forest ecosystem?\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e\u0026bull; How can GIS integration enhance fire spread prediction and support wildfire management?\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThis study examines the effectiveness of the FARSITE model in simulating wildfire spread during the Darisan event, identifies the primary factors influencing fire propagation, and evaluates the contribution of GIS in enhancing fire prediction and management. The objectives are:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eEvaluating FARSITE\u0026rsquo;s accuracy in predicting fire spread.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eIdentifying environmental and topographical factors influencing wildfire behavior.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eEnhancing fire modeling through GIS-based spatial analysis.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eBy achieving these goals, this research aims to advance wildfire modeling and improve understanding of fire dynamics in complex forest ecosystems.\u003c/p\u003e"},{"header":"Case study","content":"\u003cp\u003eThe study area (Darishan forest) is in Dohezar forest area in Northwestern of Iran (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This study focuses on the Darishan Forest within the Dohezar region (UTM Zone 39N). The selection of the case study area is based on two primary factors.\u003c/p\u003e \u003cp\u003eFirst, the Darishan wildfire was the largest and most destructive fire event in local history, making it a critical case study for understanding fire spread dynamics and validating simulation models. Comprehensive fire records are available, including ignition locations, burned area perimeters, and wildfire weather data, Sourced from regional natural resources and watershed management organization of province.\u003c/p\u003e \u003cp\u003eSecondly, the study area is located within the Hyrcanian Forests in northwestern Iran, known for their distinctive ecological features harboring significant vegetation species such as Fagus orientalis, Carpinus betulus, Quercus castanaefolia, Alnus subcordata, Acer insign, Tilia begonifolia, Buxus hyrcana, Juglans regia, and Parrotia persica. This region boasts a rich diversity of flora, contributing to high biodiversity and significant conservation value and providing critical habitat for rare and endemic species and playing vital roles in climate regulation, soil conservation, and water resource management.\u003c/p\u003e \u003cp\u003eThe mix of varied vegetation types, substantial fuel loads, and intricate topography highlights the need for localized fire spread simulations. With the growing challenges of deforestation and climate change, studying wildfire behavior in this region is essential for refining fire management strategies, improving predictive models, and promoting sustainable forest conservation efforts.\u003c/p\u003e"},{"header":"Methodology","content":"\u003cp\u003eAmong the available fire simulation models, the Fire Area Simulator (FARSITE) has been widely adopted for predicting fire behavior in diverse ecosystems (Finney \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e1998\u003c/span\u003e, 2002). FARSITE integrates key environmental variables, including fuel models, weather conditions, and topographical data, to generate spatiotemporal fire spread projections. Previous applications of FARSITE have demonstrated its effectiveness in simulating fire spread under various landscape conditions, including Mediterranean forests (Salis \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) and mountainous terrain (Nyatondo \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn this study, we employ FARSITE to model fire spread behavior in the Darishan area of the Dohezar and Sehezar forests region, incorporating detailed geospatial datasets through Geographic Information Systems (GIS). The integration of GIS allows for the precise mapping of fuel types, vegetation distribution, and terrain features, thereby enhancing the accuracy of fire simulations. By leveraging high-resolution spatial data, we aim to assess the extent to which FARSITE can accurately reconstruct the fire spread patterns observed in historical wildfire event in study area.\u003c/p\u003e\u003ch3\u003eMathematical Basis for Fire Spread Modeling\u003c/h3\u003e\u003cp\u003eFire modeling relies on fundamental mathematical equations that describe the energy released during combustion (Albini \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1979\u003c/span\u003e). Fire spreads as energy from the flame’s transfers to adjacent flammable materials, raising their temperature until ignition occurs. Accurately modeling this energy transfer enhances the ability to predict the spatial and temporal progression of fire (Rothermel \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e1991\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe most important mathematical equations used in fire modeling focus on the energy generated during a fire (Albini \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1979\u003c/span\u003e). The process of fire spread involves the release of energy from the flames and the transfer of part of this energy to adjacent flammable materials, which increases their temperature until they reach the ignition point and begin to burn (Albini \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1979\u003c/span\u003e). Modeling the flow of energy enhances the ability to predict the propagation and spatio-temporal evolution of fire (Rothermel \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e1991\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eMany field studies and observations generally support the elliptical shape of fire propagation (Richard \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e1995\u003c/span\u003e). For simulation purposes, it is suggested to consider points or vertices at regular intervals around the fire perimeter, each representing a new small source from which the fire begins to spread. These two-dimensional vertices, defined by X and Y coordinates, are considered as the peripheral points of fire propagation (Richard \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e1995\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis fire growth simulation model, grounded in Huygens' principle of wave propagation, suggests that environmental conditions are calculated individually for each vertex on the fire's perimeter. The elliptical shape and direction of the fire front are then determined according to the time taken for each calculation step. Richard developed formulas based on Huygens’ principle to model fire propagation, which can be incorporated into the FARSITE simulation (Richard \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e1995\u003c/span\u003e). These formulas describe the elliptical wave of the fire front using vertices along the fire perimeter.\u003c/p\u003e\u003cp\u003eHuygens' principle posits that each vertex acts as the origin of an expanding oval, growing independently.\u003c/p\u003e\u003cp\u003eThe data required for each vertex includes:\u003c/p\u003e \u003cp\u003ea) The position of each vertex on the fire front, denoted by xsx_sxs​ and ysy_sys​\u003c/p\u003e \u003cp\u003eb) The maximum extent of fire spread (θ), determined by the slope vector and wind direction\u003c/p\u003e \u003cp\u003ec) The elliptical shape of the fire, primarily determined by the dominant conditions at a given vertex, specified by the parameters aaa, bbb, and ccc.\u003c/p\u003e\u003cp\u003eThe equation used in the FARSITE model to calculate fire spread rate was developed by Rothermel (Rothermel 1979). It computes the rate of fire spread (RRR) in meters per minute for each vertex on a plane parallel to the Earth's surface.\u003c/p\u003e\u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:R=\\frac{{I}_{R}\\xi\\:\\left(1+{{\\Phi\\:}}_{W}+\\:{{\\Phi\\:}}_{S}\\right)\\:}{{\\rho\\:}_{b}\\epsilon\\:{\\mathcal{Q}}_{ig}}\\)\u003c/span\u003e \u003c/span\u003e Eq.\u0026nbsp;(1)\u003c/p\u003e\u003cp\u003eIn the above equation, III represents the intensity of the reaction (kJ/min/m²), ϕ\\phiϕ is the flux ratio of propagation, ρ\\rhoρ is the volume of dry material (kg/m³), NNN is the number of factors affecting heat generation, TignitionT_{\\text{ignition}}Tignition​ is the temperature before ignition (kJ/kg), and θ\\thetaθ and vvv are the slope and wind speed, respectively.\u003c/p\u003e\u003cp\u003eParameters Affecting the FARSITE Model\u003c/p\u003e\u003ch3\u003eThe Process of Implementing the FARSITE Model\u003c/h3\u003e\u003ch2\u003eStep 1: Preparing Spatial Data by ArcGIS\u003c/h2\u003e\u003cp\u003eTo implement the fire spread model, the first step is preparing spatial data. ArcGIS software was used to process and prepare the necessary spatial data. These data layers were subsequently transformed into ASCII format for integration with the simulation model.\u003c/p\u003e\u003ch3\u003eStep 2: Creating the Landscape Layer by combination of GIS\u003c/h3\u003e\u003cp\u003eThe converted spatial layers were imported into the model and integrated to generate a comprehensive \"landscape\" layer. This layer included essential spatial characteristics of the study area, such as topography, vegetation, and fuel properties.\u003c/p\u003e\u003ch2\u003eStep 3: Inputting Meteorological Data\u003c/h2\u003e\u003cp\u003eMeteorological data was collected from the Synoptic Station near the study area at 3-hour intervals. These data were then incorporated into the model to account for weather conditions influencing fire behavior.\u003c/p\u003e\u003ch3\u003eStep 4: Sensitivity Analysis and Fire Simulation\u003c/h3\u003e\u003cp\u003eA sensitivity analysis was conducted to assess the model’s response to key parameters. Using FARSITE version 4.1.0, fire spread was simulated based on spatial and meteorological data.\u003c/p\u003e\u003ch3\u003eStep 5: validity of model\u003c/h3\u003e\u003cp\u003eFinally, for validity of model, The Kappa index is calculated.\u003c/p\u003e\u003ch2\u003eInput Data required\u003c/h2\u003e\u003cp\u003eSimulations in the FARSITE model require a set of spatial data, which are classified into three main environmental factors that influence fire behavior:\u003c/p\u003e\u003cul\u003e \u003cli\u003e \u003cp\u003eEcological factors /topography\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eWeather conditions\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eVegetation\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e\u003cp\u003eThese factors are often referred to as the \"fire environment triangle.\"\u003c/p\u003e\u003ch2\u003eEcological factors/ Topography\u003c/h2\u003e\u003cp\u003eEcological factors, including topographic criteria such as slope, geographic aspect, and elevation, play an essential role in fire risk assessment. Topography can significantly impact vegetation composition and distribution, microclimate, soil moisture, and, ultimately, fire behavior (Syphard 2008). The relationship between altitude and fire risk is influenced by temperature, humidity, and vegetation. As the altitude increases, the risk of fire tends to decrease. In contrast, steeper slopes can affect soil moisture content and heighten fire risk. In dense areas, the slope of the terrain brings the flames closer to the ground, thereby increasing the likelihood of fire (Whelan 1995). Areas receiving more sunlight are typically more prone to intense fires (Franklin et al. 2000).\u003c/p\u003e\u003cp\u003eTopography exhibits significant spatial variation but is treated as constant over time in simulations. It is defined by three distinct layers: height, slope, and slope direction. Height refers to the altitude above sea level and is required to calculate the burned surface during the simulation process. Slope and slope direction are used to calculate fire spread, flame angle, and adjust the fire spread surface on a horizontal plane.\u003c/p\u003e\u003cp\u003eTopographic data, including height, slope, and direction, were derived from 1:25,000 scale maps supplied by the Iran National Cartographic Center in raster format. The data were then used to create a digital elevation model (DEM), which provided elevation, slope, and aspect data for the simulation. These data were then converted to ASCII format using ArcGIS software to prepare them for input into the FARSITE model.\u003c/p\u003e\u003ch2\u003eWeather condition\u003c/h2\u003e\u003cp\u003eFor weather data, climate statistics from the station with the most similar atmospheric conditions to the study area were used. included key meteorological variables, such as temperature, relative humidity, wind speed, wind direction, and precipitation.\u003c/p\u003e\u003cp\u003eTemperature, relative humidity, and precipitation data from a specified location were incorporated into the FARSITE model using a weather stream file.\u003c/p\u003e\u003cp\u003eTo maintain consistency and accuracy, climate statistics were obtained from the synoptic station that most closely matched the atmospheric conditions of the study area. Of the available stations, the Ramsar synoptic station, located 25 km west of the study area's boundary, was identified as the most representative. Its 3-hour interval data formed the basis for the weather inputs in the simulation. This dataset included crucial meteorological variables such as temperature, relative humidity, wind speed, wind direction, and rainfall.\u003c/p\u003e\u003ch2\u003eVegetation\u003c/h2\u003e\u003cp\u003eVegetation acts as a flammable fuel bed, driving fire spread and significantly influencing fire characteristics, including intensity, spread rate, and flame height (Pyne 1996). Among environmental factors, vegetation is the only one that can directly control fire behavior through management operations. The characteristics of vegetation can be categorized into several types, including fuel type, vegetation canopy, and forest mass height.\u003c/p\u003e\u003ch2\u003eVegetation Canopy\u003c/h2\u003e\u003cp\u003eA canopy map of the area is essential for calculating shading and wind speed reduction factors for all fuel models.\u003c/p\u003e\u003cp\u003eThe vegetation canopy refers to the horizontal area of the ground surface directly covered by tree crowns. Canopy cover maps of the region, which are vital for assessing the extent of shading on the forest floor, were acquired.\u003c/p\u003e\u003cp\u003eThe vegetation canopy represents the horizontal area of the ground surface directly covered by the tree crowns. Canopy cover maps of the region, which are crucial for determining the level of shading on the forest floor, were obtained.\u003c/p\u003e\u003cp\u003eThese maps are crucial for calculating the moisture content of combustible materials and for models that predict fire behavior. Furthermore, canopy cover affects the amount of radiant energy the forest receives. The canopy cover maps were classified based on data provided by the Forests and Rangelands, and Watershed Management Organization of Iran, and were then converted into a format compatible with the FARSITE model.\u003c/p\u003e\u003ch2\u003eFuel Model\u003c/h2\u003e\u003cp\u003eFuel is one of the primary elements in the burning process, whether the fire is at the ignition or spread stage. Mathematical equations and predictive systems designed to model fire behavior require data related to the burning area surface and fuel bed. This data includes fuel load, bulk density, fuel particle size, heat content, and moisture content limitations (Scott \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo simplify the use of these factors in models and systems, they are presented as a set of fuel models. Rothermel (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1972\u003c/span\u003e) defined a fuel model as a complete set of inputs used to define a flammable bed in a mathematical model of fire spread (Rothermel 1979). He described the fuel model as a tool that allows users to predict fire behavior and patterns with greater accuracy (Anderson 1982). At the first level, fuel models are classified into base and initial groups based on cover type, with seven distinct groups. This classification is like the one proposed by Anderson in 1982 (Anderson 1982).\u003c/p\u003e\u003cp\u003eTo select a standard fuel model, the models developed by Scott and Burgan were utilized. This updated set of fuel models, designed for fire behavior analysis, is compatible with any simulation system based on Rothermel's (1972) fire spread model (Rothermel, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1972\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSince fuel properties are typically complex, fire managers have attempted to simplify them by categorizing the physical parameters and spatial distribution of fuel into different classes, known as ‘fuel models’ (Anderson 1982; Burgan and Rothermel 1984). These fuel characteristics' spatial distribution is often represented as fuel type maps. Various classification systems are used to group vegetation types based on their fire behavior. Fifty-three standard fuel models have been developed for the Rothermel (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1972\u003c/span\u003e) fire spread model. According to Nyatondo (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), these include the original 13 fuel models described by Anderson (1982) and the additional 40 models outlined by Scott and Burgan (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). In this study, the selection of appropriate fuel models was based on the similarities between the vegetation characteristics observed in the field and the descriptions of the standard fuel models (Anderson 1982; Scott and Burgan \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) as shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The vegetation cover types identified during fieldwork were reclassified according to the selected fuel models, reflecting the fuel conditions observed on-site.\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescription of the standard fuel models used for simulation in the wildfire case studies from 53 standard fuel models (Anderson 1982; Scott and Burgan \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2005\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObserved vegetation\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFire carrying fuel type, model name and code\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFuel model number\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFuel model description\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRiver\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInsufficient wildland fuel to carry wildland fire under any condition (non-burnable, NB8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e98\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLand covered by open bodies of water such as lakes, rivers and oceans\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrass (Low density)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHumid-climate grass (Grass, GR5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGrass and herb fuel load is light; fuel bed depth is about 1 to 2 feet.\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrass (Medium density)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eContinuous humid-climate grass (Grass, GR6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e106\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLoad is greater than GR5 but depth is about the same.\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrass-shrub\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGrass and shrubs combined (Grass-Shrub, GS3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e123\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModerate grass/shrub load, average grass/shrub depth less than 2 feet\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNatural Forest (Medium density and timber-shrub)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModerate litter load with shrub component (Timber-Understory, TU2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e162\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigh extinction moisture. Spread rate is moderate; flame length low\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNatural forest/Mixed Forest (Medium density and timber-grass shrub)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGrass and shrub mixed with litter from forest canopy (Timber Understory, TU3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e163\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eExtinction moisture is high. Spread rate is high; flame length moderate\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHardwood plantation (Medium density)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModerate load broadleaf litter (Timber Litter, TL6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e186\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLess compact than TL2. Spread rate is moderate; flame length low\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHardwood plantation (High density)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVery high load, fluffy broadleaf litter (Timber Litter, TL9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e189\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpread rate is moderate; flame length moderate.\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003eIn the final step of evaluating the simulation results, the Kappa index was employed to assess the model's validity.\u003c/p\u003e\u003cp\u003eThe Kappa index is calculated using the following matrices and equations (Eq.\u0026nbsp;2).\u003c/p\u003e\u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:\\text{K}=\\frac{\\sum\\:_{\\text{i}=1}^{\\text{n}}{\\text{P}}_{\\text{i}\\text{i}}-\\sum\\:_{\\text{i}=1}^{\\text{n}}{\\text{q}}_{\\text{i}\\text{i}}}{1-\\sum\\:_{\\text{i}=1}^{\\text{n}}{\\text{q}}_{\\text{i}\\text{i}}}\\)\u003c/span\u003e \u003c/span\u003e Eq.\u0026nbsp;(2)\u003c/p\u003e\u003cp\u003eWhere Pii ​ is the observed probability and ​ qii is the expected probability.\u003c/p\u003e"},{"header":"Results and Discussion","content":"\u003cp\u003eThe study was conducted in the Darishan Forest, located within the Dohezar forest area, in watershed region 33 at latitude 36˚19\u0026prime;22\u0026Prime; N and longitude 50˚29\u0026prime;57\u0026Prime; E. The Hyrcanian Forests, which encompass Darishan, are renowned for their rich biodiversity. In December 2000, the study area was affected by a fire that burned 72.63 hectares, and this event was used to simulate fire behavior in the region.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eData that is used in study area\u003c/h2\u003e \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e \u003ch2\u003eTopography:\u003c/h2\u003e \u003cp\u003eTopographic data, including elevation, slope, and aspect, were extracted from 1:25,000 scale maps provided by the Iran National Cartographic Center in raster format. Figures\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e illustrate the slope, elevation, and aspect of the study area. The data were then used to create a digital elevation model (DEM), which supplied the elevation, slope, and aspect data for the simulation. The data were subsequently converted to ASCII format using ArcGIS software to prepare them for input into the FARSITE model.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eWeather data:\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e displays the 3-hour intervals of temperature, rainfall, wind speed, and wind direction on the day of the fire (December 8, 2000), as recorded in the study area.\u003c/p\u003e \u003cp\u003eThe average annual rainfall at different elevations ranges from a maximum of 1,255 mm to a minimum of 645 mm. The mean annual temperature varies between 10.3\u0026deg;C and 16.1\u0026deg;C, depending on the specific location. At lower elevations, the average minimum and maximum monthly temperatures are 3\u0026deg;C and 26.5\u0026deg;C, respectively. In contrast, at higher elevations, based on the thermal gradient, the minimum and maximum temperatures are \u0026minus;\u0026thinsp;20.5\u0026deg;C and 6.3\u0026deg;C, respectively. Due to the region\u0026rsquo;s topographical features and its classification under the Amberger climate system, the climate is categorized as cold, humid, and mountainous.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ethe 3-hour intervals of temperature, rain, wind speed, and wind direction on the day of the fire\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTemperature (degrees Celsius)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRain intensity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWind speed\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eWind direction\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e8.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e7.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSW\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNE\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e15:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e18:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSW\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSW\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e00:30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e13.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNW\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eVegetation Canopy density\u003c/h2\u003e \u003cp\u003eIn this study, the selection of appropriate fuel models was based on the similarities between the field-observed vegetation characteristics and the descriptions of standard fuel models (Anderson 1982; Scott and Burgan \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), as presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. A map depicting the vegetation density types is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eclassification of vegetation canopy in study area\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVegetation type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVegetation canopy (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCanopy class\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWithout vegetation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShrub lands\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u0026ndash;25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow-density forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25\u0026ndash;50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerate-density Forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e50\u0026ndash;75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh-density forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e75\u0026ndash;100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eFire Spread Barriers\u003c/h2\u003e \u003cp\u003eThe FARSITE model operates under the assumption that fire spread is directly influenced by the availability and type of combustible material. If combustible material is present, the fire will continue to spread, and the model does not inherently account for factors that limit or extinguish the fire. As a result, barriers such as roads, rivers, and firebreaks are incorporated into the model to restrict fire spread, improving the accuracy of the simulation. In this study, roads and rivers were extracted from 1:25,000 maps provided by the Iran National Cartographic Center and prepared for use in the model\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eFire History:\u003c/h2\u003e \u003cp\u003eThe fire reality map was created using a 25-year history of fire events, covering the period from 1997 to 2023. Information from the Natural Resources, Range, and Watershed Organization of Mazandaran Province in Iran, along with field visits with environmental and forest experts, was used to prepare this map. Additionally, remnants of past fires in the study area were recorded using GPS technology and GIS Arc software. To conduct the field measurements, a team of forest experts, rangers and environmentalists walked for 8 hours through the Dohezar forest to reach the Darishan forest area, where all fire points were recorded with GPS. As we observed, many large, old trees had burned, leaving only half of their trunks behind, while shrubs and grass had grown in their place, clearly indicating the effects of past fires. Based on the field survey, the starting and ending points of the wildfire were identified through expert assessments and observations.\u003c/p\u003e \u003cp\u003eThe affected area exhibited a triangular shape, with the apex positioned at a lower elevation and the broader base at a higher elevation. Akey factor in containing the fire was a substantial natural barrier, resembling a large rock formation or mountain. This barrier, composed of stone, created a significant gap between the trees, effectively acting as a firebreak. The dimensions of the rock formation were measured by the team, with a width of 10 meters and a height of 40 meters, which played a crucial role in halting the spreading of fires.\u003c/p\u003e \u003cp\u003eDuring our field visit, we also noticed that many shepherds had built small fires in the forest to stay warm but had left them unattended without extinguishing the fire, which could be one of the primary causes of fire ignition in the area. The primary outputs of the fire spread simulation using the FARSITE model are maps showing the fire spread at predetermined time intervals, as well as the final fire spread map. The simulated fire spread map generated by the FARSITE model is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eTo evaluate the accuracy of the simulation, the Kappa coefficient and consistency ratio were computed. The consistency ratio represents the proportion of the fire area that has been simulated as burned, while the Kappa coefficient measures the similarity between the observed and simulated fire spots. It accounts for both underestimation and overestimation, ethe fire spread falling below or above the actual area burned. The Kappa coefficient measures the similarity between observed and simulated fire spots, calculating both underestimation and overestimation (i.e., the estimation falling below or above the actual fire spread).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe results of FARSITE model, Kappa coefficient, and consistency ratio are shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eResults of the FARSITE model, Kappa coefficient, and consistency ratio analysis.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSimulated fire are (ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eObserved fire area (ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnderestimation (ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOverestimation (ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eConsistency (ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eKappa\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e35.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e72.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e37.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e34.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe main outputs of the fire simulation process are fire spread maps at predetermined time intervals and a final fire spread map. To evaluate the accuracy of the simulation, the Kappa coefficient and consistency ratio were used. The consistency ratio measures the proportion of the actual fire-affected area that was accurately simulated as burned, while the Kappa coefficient evaluates the similarity between observed and simulated fire spots, considering both overestimation (excessively predicted burned area) and underestimation (areas missed by the simulation).\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents the results of the FARSITE simulation. Of the 72.63 hectares of observed burned area, 34.93 hectares were correctly predicted, resulting in a high consistency ratio. The overestimated area was minimal (0.85 ha), but the model underestimated 37.7 hectares, suggesting potential limitations in capturing certain fire spread patterns. A Kappa coefficient of 0.68 reflects a moderate level of correspondence between the observed and simulated fire spread.\u003c/p\u003e \u003cp\u003eWhile the model performs well in estimating fire spread, the high underestimation suggests a need for refining input parameters or integrating additional environmental variables (e.g., real-time moisture conditions) to improve prediction accuracy.\u003c/p\u003e \u003cp\u003ePossible reasons for overestimation and observed burned areas\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eErrors in input data (e.g., vegetation, fuel moisture).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLimitations of the FARSITE model in specific terrains.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eField measures challenges\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eFuel models which are generated for the USA\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLack of information in databases\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eIncorporating such data into fire spread modeling could enhance the accuracy of simulations (Arca et al. 2007; Arroyo et al. 2008; Jahdi 2014). However, due to limited information and the challenges of field-measuring large historical fires, such as the inaccessibility of certain regions and the removal of old burn marks in forest areas, this remains difficult. Conversely, access to documented data on the behavior of larger fires facilitates the testing of more significant changes in fire spread. The fuel models used in this study were developed in the United States and, while they closely align with the conditions in the study area, directly measuring fuel parameters (such as fuel load and moisture content) could yield more accurate models (Jahdi, 2014).\u003c/p\u003e \u003cp\u003eOne limitation of FARSITE is that it does not account for the effects of the two-dimensional fire shape on acceleration. Instead, it models fire acceleration as a point-source for simplicity (Phillips, 2006). Modifying the scale and type of the fuel model could also enhance the model's ability to simulate fires in this region. Additionally, using a 10-meter Digital Elevation Model (DEM) would provide higher resolution and offer a more accurate representation of the fuel heterogeneity (Philips 2006). This highlights the need to create fuel \u0026micro;odels tailored to the te\u0026micro;perate conditions of northern Iran's forests or Hyrcanian Forest fuel \u0026micro;odel. Developing these custo\u0026micro;ized \u0026micro;odels requires adjusting fuel para\u0026micro;eters based on field observations (Nyatondo \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn comparison with other research, recent studies have increasingly focused on fire risk assessment and fire spread simulation (Wang 2023; Wu 2022; Burge et al. 2020; Hodges and Lattimer \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Marjani and Mesgari \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Qi et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Kanagaraj et al. 2023; Zigner et al. 2022). Previous research indicates that the accuracy of the FARSITE model can be improved by using fuel models specifically designed for simulating fire spread and behavior in various vegetation types (Fernandes 2009; Salis 2008; Bacciu 2009). Further studies focused on actual fire behavior in the region are necessary to validate and enhance the FARSITE model, with thorough calibration potentially increasing the accuracy of predictions (Jahdi 2014). Given the robust wildfire database in northern forests of Iran, conducting simulations in other regions and comparing results could significantly refine this approach. Additionally, improving wildfire simulations can be achieved by developing local and regional fuel models and collecting more detailed weather data.\u003c/p\u003e \u003cp\u003eA comparative analysis with similar FARSITE simulations in Mediterranean climates suggests that such underestimations can result from the challenges in calibrating fuel models and integrating accurate local meteorological data.\u003c/p\u003e \u003cp\u003eStudies in these regions have also encountered challenges in accurately modeling fire spread, highlighting the need for more refined fuel models and localized data to enhance predictions. Fire spread models, supported by accuracy assessments, are vital for understanding fire behavior, evaluating associated risks, assisting in active fire suppression, and guiding fuel treatment strategies, such as the creation of fuel breaks (Finney \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Shinneman et al. 2019). The use of generalized fuel bed maps, such as those provided by LANDFIRE, can lead to discrepancies between simulated fire spread and observed patterns.\u003c/p\u003e \u003cp\u003eThe model accuracy assessment and alternative methods for selecting Fire Behavior Fuel Models (FBFMs) used in some studies represent advancements over field-vegetation map-based accuracy assessments in models like MTT and FARSITE, which were initially developed for semiarid environments, such as Mediterranean shrublands or Iranian grasslands (Arca et al. 2007; Jahdi et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, 2016; Salis et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, 2021). However, research by Samuel Price and Matthew J. Germino (2022) suggests that such methods may not be suitable for all site-specific applications.\u003c/p\u003e \u003cp\u003eStudies using the FARSITE simulator have demonstrated that areas with a high risk of fire often align with regions that have previously burned, supporting the model's validity. As one of the most widely used fire spread simulators, FARSITE can produce valuable results for Iranian forests, provided that a properly calibrated fuel model is used. The simulator\u0026rsquo;s user-friendly interface, transparency, and adaptability to different forest fire types increase its usefulness (Eskandari 2013). However, J. Germino's (2022) research suggests that such methods may not be applicable for most site-specific applications.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study assesses the suitability of the FARSITE fire behavior simulation model for use in Iranian forests.The simulation results showed good agreement between simulated and observed fire environments, with the spread pattern and fire shape aligning well with the actual fire spot. Although minor overestimation was observed, significant underestimation, especially in the northern region, occurred, likely due to inaccuracies in the fuel model or inadequate meteorological data.\u003c/p\u003e \u003cp\u003eDespite this, the overall results were reliable, suggesting that FARSITE can effectively simulate fire spread in the study area. GIS-based fire spread models like FARSITE generate polygonal shapes that closely match natural fire spread patterns and incorporate factors influencing fire behavior, thus enhancing model credibility. The findings provide valuable insights for fire management in the Hyrcanian Forests, supporting decisions on firebreak placement, resource allocation, and the formulation of prevention strategies, such as controlled burns.\u003c/p\u003e \u003cp\u003eTo improve future simulations, integrating high-resolution remote sensing data, refining fuel models to better reflect local vegetation, and using more accurate meteorological data are recommended. These improvements will enhance fire risk management not only in Iran but also in similar forest ecosystems worldwide. This research contributes to the advancement of wildfire management strategies, ecological conservation, and fire simulation model development, providing valuable guidance to policymakers, forestry officials, and conservation agencies in implementing effective fire prevention and mitigation strategies specifically tailored to Iran's forest landscapes.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAdab H, Kanniah KD, Solaimani K (2013) Modeling Forest fire risk in the northeast of Iran using remote sensing and GIS techniques. Natural Hazards 65:1723-1743.\u003c/li\u003e\n \u003cli\u003eAjin RS, Ana-Maria L, Jacob MK, Vinod PG, Krishnamurthy RR (2016) The Risk Assessment Study of Potential Forest Fire in Idukki Wildlife Sanctuary using RS and GIS Techniques. Int J Adv Earth Sci Engin 5:308-318.\u003c/li\u003e\n \u003cli\u003eAliani H, Babaie Kafaky S. Saffari A. Monavari S. M (2016) Determining an appropriate method for the purpose of land allocation for ecotourism development (case study: Taleghan County, Iran). Environ Monit Assess .188:651 DOI 10.1007/s10661-016-5651-2.\u003c/li\u003e\n \u003cli\u003eAlbini FA (1979) Spot fire distance from burning trees: a predictive model. USDA Forest Service, General Technical Report INT, 56 pp.\u003c/li\u003e\n \u003cli\u003eAlonso-Betanzos A, Fontenla-Romero O, Guijarro-Berdi\u0026ntilde;as B, Hern\u0026aacute;ndez-Pereira E, Canda E, Jimenez J, Legido JL, Mu\u0026ntilde;iz S, Paz-Andrade C, Paz-Andrade ML (2002) A neural network approach for forest fire risk estimation. Presented at the 15th 2nd Int Conf on Artif Intell, ECAI\u0026apos; Lyon, France\u003c/li\u003e\n \u003cli\u003eAnderson DH, Catchpole EA, De Master NJ, Parkes T (1982) Modeling the spread of grass fires. J Austral Math Soc (Series B) 23:451-466\u003c/li\u003e\n \u003cli\u003eBowman DMJS, French BJ, Williamson GJ, Prior LD (2021) Fire, herbivores and the management of temperate Eucalyptus savanna in Tasmania: Introducing the Beaufront fre \u0026ndash; mammalian herbivore feld experiment. Ecol Manag Restor 22:140\u0026ndash;151\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eBui DTB, Pradhan O, Lofman I, Revhaug O, Dick B (2012) Landslide susceptibility mapping at Hoa Binh province (Vietnam) using an adaptive neuro-fuzzy inference system and GIS. Comput Geosci 45:199-211\u003c/li\u003e\n \u003cli\u003eChandra S (2005) Application of Remote Sensing and GIS Technology of Forest Fires: A Case Study in the Garhwal Himalayan Region. In: Geo-Inform Disaster Manag, 1239-1254\u003c/li\u003e\n \u003cli\u003eChandra S, Aurora MK (2006) Forest fire risk zonation mapping using remote sensing technology (proceedings paper). J SPIE (abstract), 1-2\u003c/li\u003e\n \u003cli\u003eChuvieco E, Kasischke ES (2007) Remote sensing information for fire management and fire effects assessment. J Geophys Res: Biogeosci 112:1-8. https://doi.org/10.1029/2006JG000230\u003c/li\u003e\n \u003cli\u003eCruz MG, Alexander ME (2013) Uncertainty associated with model predictions of surface and crown fire rates of spread. Environ Modell Softw 47:16-28\u003c/li\u003e\n \u003cli\u003eEskandari S, Oladi Ghadikolaei J, Jalilvand H, Saradjian S (2013) Detection of Fire High-Risk Areas in Northern Forests of Iran Using Dong Model. World Appl Sci J 27(6):770-773\u003c/li\u003e\n \u003cli\u003eFinney M (1998) FARSITE: Fire Area Simulator-model development and evaluation. USDA For Serv Res Pap RMRS-RP-4.\u003c/li\u003e\n \u003cli\u003eFinney MA (1998) Efforts at comparing simulated and observed fire growth patterns. Final Report INT-95066-RJVA. Systems for Environmental Management, Missoula, MT, USA.\u003c/li\u003e\n \u003cli\u003eFinney MA (2006) An Overview of FlamMap, Fire Modeling Capabilities. USDA For Serv Proc, pp 213-220\u003c/li\u003e\n \u003cli\u003eFinney MA, Bradshaw L, Butler BW (2006) Modeling surface winds in complex terrain for wildland fire incident support. Final Report for JFSP Funded Project #03-2-1-04. U.S. Department of Agriculture, Forest Service, Fort Collins, CO, USA, pp 1-10.\u003c/li\u003e\n \u003cli\u003eFinney MA, Grenfell IC, McHugh CW, Seli RC, Trethewey D, Stratton RD, Brittain S (2011) A method for ensemble wildland fire simulation. Environ Model Assess 16:153-167\u003c/li\u003e\n \u003cli\u003eGhobadi GJ, Gholizadeh B, Dashliburun OM (2012) Forest fire risk zone mapping from geographic information system in Northern Forests of Iran (Case study, Golestan Province). Int J Agric Crop Sci 4(12):818-824\u003c/li\u003e\n \u003cli\u003eGoleiji E, Hosseini SM, Khorasani N, Monavari SM (2017) Forest fire risk assessment-an integrated approach based on multicriteria evaluation. Environ Monit Assess. https://doi.org/10.1007/s10661-017-6225-7, pp 189-612\u003c/li\u003e\n \u003cli\u003eHodges JL, Lattimer BY (2019) Wildland Fire Spread Modeling Using Convolutional Neural Networks. Fire Technol 1\u0026ndash;28. https://doi.org/10.1007/s10694-019-00846-4\u003c/li\u003e\n \u003cli\u003eJahdi R, Salis M, Darvishsefat A, Mostafavi A, Alcasena F, Etemad V, Lozano O, Spano D (2015) Calibration of FARSITE simulator in northern Iranian forests. Nat Hazards Earth Syst Sci 15:443-459\u003c/li\u003e\n \u003cli\u003eJain P, Coogan SCP, Subramanian SG, Crowley M, Taylor SW, Flannigan MD (2020) A review of machine learning applications in wildfre science and management. Environ Rev 0(ja). https://doi.org/10.1139/er-2020-0019\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eJohn B, Matthew B, Matthias I et al (2020) Convolutional LSTM Neural Networks for Modeling Wildland Fire Dynamics. https://doi.org/10.48550/arXiv.2012.06679\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eJordan MI, Mitchell T (2015) Machine learning: Trends, perspectives, and prospects. Sci 349:255\u0026ndash;260.\u003c/li\u003e\n \u003cli\u003eKeane RE, Burgan R, van Wagtendonk J (2001) Mapping wild land fuels for fire management across multiple scales: integrating remote sensing, GIS, and biophysical modeling. Int J Wildland Fire 10:301-319\u003c/li\u003e\n \u003cli\u003eKevin T, Brett S, Derek C (2008) Phoenix: development and application of a bushfire risk management tool. Aust J Emerg Manag 23:47-54. https://doi.org/10.1007/978-3-540-74492-4_21\u003c/li\u003e\n \u003cli\u003eLi Z, Huang Y, Li X, Xu L (2020) Wildland fre burned areas prediction using long short-term memory neural network with attention mechanism. Fire Technol 57:1\u0026ndash;23\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLiu S, Cao L, Lin C, Dai Y, Li X, Li S, Sun S, Li D (2024) Fire spread prediction model based on multi scale convolutional neural network. Multimed Tools Appl. https://doi.org/10.1007/s11042-024-20400-x.\u003c/li\u003e\n \u003cli\u003eMarjani M, Mesgari MS (2023) The large-scale wildfre spread prediction using a multi-kernel convolutional neural network[J]. ISPRS Ann Photogram Remote Sens Spatial Inform Sci 10:483\u0026ndash;488\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMarozas V, Racinskas J, Bartkevicius E (2007) Dynamics of ground vegetation after surface fires in hemi boreal Pinus sylvestris forests. For Ecol Manage 250(1-2):47-55\u003c/li\u003e\n \u003cli\u003eMarshall G, Thompson DK, Anderson K, Simpson B, Linn R, Schroeder D (2020) The impact of fuel treatments on wildfire behavior in north American boreal fuels: a simulation study using FIRE[1]TEC. Fire-Switz 3(2):18\u003c/li\u003e\n \u003cli\u003eMirdeylami T, Shataee Sh, Kavousi MR (2014) Forest fire risk zone mapping in the Golestan national park using weighted linear combination (WLC) method. Iran J For 4(4):377-390 (In Persian)\u003c/li\u003e\n \u003cli\u003eMohammadi F, Shabanian N, Pourhashemi M, Fatehi P (2010) Risk zone mapping of forest fire using GIS and AHP in a part of Paveh Forests. Iran J For Poplar Res 18(4):569-586\u003c/li\u003e\n \u003cli\u003eNasanbat E, Lkhamjav O (2016) Wild Fire Risk Map In The Eastern Steppe Of Mongolia Using Spatial Multi-Criteria Analysis. Int Arch Photogramm Remote Sens Spatial Inf Sci, pp 1-9\u003c/li\u003e\n \u003cli\u003eNyatondo UN (2010) Fire spread modeling in Majella National Parks, Italy. Master of Science, Int Inst Geo-Inform Sci Earth Obs, Enschede, the Netherlands, 91 p\u003c/li\u003e\n \u003cli\u003ePaz Sh, Carmel Y, Jahshan F, Shoshany M (2011) Post-fire analysis of pre-fire mapping of fire risk: A recent case study from Mt. Carmel. For Ecol Manage 262:1184-1188\u003c/li\u003e\n \u003cli\u003ePrice S, Germino MJ (2022) Modeling of fire spread in sagebrush steppe using FARSITE: an approach to improving input data and simulation accuracy. Fire Ecol. https://doi.org/10.1186/s42408-022-00147-2.\u003c/li\u003e\n \u003cli\u003ePyne SJ, Andrrew PL, Laven AR (1996) Introduction to wildfire. 2nd ed. John Wiley and Sons, New York, 769 p\u003c/li\u003e\n \u003cli\u003eQi P, Chang J, Chen X et\u0026nbsp;al (2022) Identifcation of Rock Properties of Rock Wall Cut by Road header Based on PSO-VMD-LSSVM[J]. Front Earth Sci 10:884633\u003c/li\u003e\n \u003cli\u003eRadke D, Hessler A, Ellsworth D (2019) FireCast: Leveraging Deep Learning to Predict Wildfire Spread. International Joint Conference on Artifcial Intelligence. https://doi.org/10.24963/ijcai. 2019/636.\u003c/li\u003e\n \u003cli\u003eRichard GD (1995) A general mathematical framework for modeling two-dimensional wildland fire spread. Int J Wildland Fire 5:63-72\u003c/li\u003e\n \u003cli\u003eRothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. USDA For Serv Intermountain For Range Exp Stn Res Pap INT-115 \u0026nbsp;\u003c/li\u003e\n \u003cli\u003eRothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. Research Paper INT-115. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT, USA.\u003c/li\u003e\n \u003cli\u003eRothermel RC (1991) Predicting behavior and size of crown fires in the Northern Rocky Mountains. USDA For Serv Res Pap INT-438, Ogden, Utah, USA, 67 p\u003c/li\u003e\n \u003cli\u003eRwanga SS, Ndambuki JM (2014) Application of geographical information systems and FARSITE in fire spread modeling. Int J Environ Sustain Dev 13(2):185-203\u003c/li\u003e\n \u003cli\u003eSalis M (2007) Fire Behavior simulation in Mediterranean Maquis using FARSITE (Fire Area Simulator). PhD Doctoral Thesis, Univ Degli Studi Di Sassari\u003c/li\u003e\n \u003cli\u003eSalis M, Arca B, Alcasena F, Arianoutsou M, Bacciu V, Duce P, Duguy B, Koutsias N, Mallinis G, Mitsopoulos I, et al. (2016) Predicting wildfire spread and behaviour in Mediterranean landscapes. Int J Wildland Fire 25:1015-1032.\u003c/li\u003e\n \u003cli\u003eScott JH, Burgan RE (2005) Standard fire behavior fuel models: A comprehensive set for use with Rothermel\u0026apos;s surface fire spread model. USDA Gen Tech Rep RMRS-GTR-153\u003c/li\u003e\n \u003cli\u003eStocks BJ, Alexander ME, Lanoville RA (2004) Overview of the International Crown Fire Modeling Experiment (ICFME). Can J For Res 34:1543\u0026ndash;1547\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWang Z (1992) General forest fre risk rating system. J Natl Disast 03:39\u0026ndash;44.\u003c/li\u003e\n \u003cli\u003eYang S, Li N (2021) Research progress of forest fire spread model. Gansu Sci Technol 37(03):45-47\u003c/li\u003e\n \u003cli\u003eYou X, Zheng Z, Yang K et\u0026nbsp;al (2023) A PSO-CNN-Based Deep Learning Model for Predicting Forest Fire Risk on a National Scale[J]. Forests 15(1):86\u003c/li\u003e\n \u003cli\u003eZhang ZX, Zhang HY, Zhou DW (2009) Using GIS spatial analysis and logistic regression to predict the probabilities of human-caused grassland fires. J Arid Environ 74:386-393\u003c/li\u003e\n \u003cli\u003eZigner K, Carvalho LMV, Peterson S, Fujioka F, Duine GJ, Jones C, Roberts D, Moritz M (2020) Evaluating the ability of FARSITE to simulate wildfires influenced by extreme, downslope winds in Santa Barbara, California. Fire MDPI.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Forest fire risk management, Fire spread simulation, FARSITE, Geographical Information System","lastPublishedDoi":"10.21203/rs.3.rs-5764687/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5764687/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGiven the widespread fires in forests worldwide, which result in the destruction of ecosystems, effective management of these forests\u0026mdash;particularly in relation to fire behavior and its spread\u0026mdash;is essential. Forest fire prediction is a worldwide problem. There are different models to predict the spread of fires, especially in forest areas but these models are usually difficult to apply for real forest fire cases and the accuracy of prediction is lower than reality. The aim of this research is to identify and predict fire spread behavior and the process of fire spread in natural areas using the FARSITE model. To conduct the research, data including topography, vegetation, and meteorological information, were compiled the study area. Using GIS tools and the FARSITE model, a simulation map of fire spread behavior in the studied forest was generated. The model tested against the spread of fires under the field conditions where real fire had occurred in forest area. To assess the accuracy of the simulation, the Kappa coefficient method was employed. The simulation results showed that 34.93 hectares of the observed fire area were accurately simulated as a burned area, demonstrating a high degree of consistency with the actual fire spread. The overestimated area was minimal (0.85 hectares), while 37.7 hectares of the fire spot were not simulated as burned. The findings of this study can inform fire spread behavior prediction models and guide the development of preventive and control measures for fire risk management in forests.\u003c/p\u003e","manuscriptTitle":"Integrating FARSITE and GIS for Enhanced Forest Fire Spread Prediction and Simulation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-10 06:25:20","doi":"10.21203/rs.3.rs-5764687/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2025-04-07T16:15:59+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-07T08:22:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-04T11:38:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"Natural Hazards","date":"2025-04-03T12:13:28+00:00","index":"","fulltext":""},{"type":"decision","content":"Minor revisions","date":"2025-01-30T02:20:42+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"254576b8-7a6e-4620-8e73-1e725ebcca01","owner":[],"postedDate":"April 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-06-23T16:02:39+00:00","versionOfRecord":{"articleIdentity":"rs-5764687","link":"https://doi.org/10.1007/s11069-025-07404-y","journal":{"identity":"natural-hazards","isVorOnly":false,"title":"Natural Hazards"},"publishedOn":"2025-06-18 15:57:48","publishedOnDateReadable":"June 18th, 2025"},"versionCreatedAt":"2025-04-10 06:25:20","video":"","vorDoi":"10.1007/s11069-025-07404-y","vorDoiUrl":"https://doi.org/10.1007/s11069-025-07404-y","workflowStages":[]},"version":"v1","identity":"rs-5764687","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5764687","identity":"rs-5764687","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}