تحلیل مکانی - زمانی توزیع آتش‌سوزی‌ها در ایران با استفاده از داده‌های ماهواره‌ای: شناسایی مناطق پرخطر و دوره‌های بحرانی

ABSTRACT
Wildfires represent a critical environmental challenge in Iran, posing severe threats to natural resources and ecosystems. This study aims to analyze the spatiotemporal patterns of wildfires across Iran from 2003 to 2023, leveraging satellite data to identify high-risk provinces and critical periods. The findings contribute to recommendations for optimal fire station placement. The methodology integrates time-series satellite data, spatiotemporal analyses using Google Earth Engine and GIS, and multivariate regression modeling with kernel density estimation to identify influential environmental factors and areas with frequent fires. Results reveal that out of a total burned area of 95,600 square kilometers, the largest burned area was recorded in 2010 (7,722 km²) and the smallest in 2022 (3,031 km²). A negative correlation of -0.24 between monthly precipitation and burned area suggests that increased rainfall is associated with reduced wildfire incidence. The provinces of Khuzestan, Fars, and Ilam recorded the highest burned areas, with Khuzestan alone experiencing over 25,000 km² of burned terrain. Summer, particularly in the years 2010 and 2007, marked peak fire occurrence and extent. By accurately identifying spatiotemporal patterns, this study provides critical insights for targeted placement of fire stations and effective wildfire management in high-risk regions.
Extended Abstract
Introduction
Wildfires are a major threat to environmental, economic and social stability in many parts of the world. Over time, the frequency and intensity of wildfires are increasing due to climate change, human activities and other natural factors. Iran is a country with diverse ecosystems and climatic conditions. The frequency of wildfires in Iran has also increased in recent decades. Understanding the spatio-temporal patterns of fires is important for developing effective techniques and strategies with minimal negative impacts. In this paper, the spatial distribution of wildfires in Iran between 2003 and 2023 has been studied using MODIS satellite data in high-risk periods and regions where environmental protection can be envisaged by optimising resource allocation.
In the light of the above discussion, the general objectives of the present research would henceforth be: 

The spatio-temporal analysis of forest fires in Iran for the last twenty years.
To identify provinces with high risk of forest fires where management and resources could be focused.
To identify high-risk seasons and periods when forest fires are most likely to occur.
Provide actionable programme recommendations: Develop and organise preventive measures to reduce the impact of forest fires.

 
Methodology
This study has developed a method that combines satellite data analysis, geographic information system techniques and statistical methods to analyse the patterns of wildfires in Iran. The main dataset used in this study is from MODIS/006/MCD64A1, a globally recognised dataset for high-resolution global burned area information. The period considered in this paper is from 2003 to 2023; therefore, it includes all fire events recorded in this period as documented by the MODIS data.
Data Collection
Of these, the MODIS Burned Area Product MCD64A1 was selected because it has a very high spatial resolution of 500 metres and is frequently updated, making it suitable for monitoring wildfires over a long period of time. The subsequent extraction of burned areas per Iranian province was made possible by the cloud-based Google Earth Engine platform, which streamlines the way users can process and analyse large geospatial datasets.
Data Analysis and Processing
A combination of Google Earth Engine was used to analyse the extent, frequency and severity of wildfires in different provinces of Iran. Seasonal and annual totals were used to aggregate the data to identify patterns of occurrence. A spatio-temporal map showing the spatial occurrence of wildfires was created, from which hotspots were derived. Overlaying the burned area data with environmental and climatic variables using GIS techniques, such as vegetation type, elevation, slope, aspect, geographical units, mean temperature and rainfall, provided an indication of the conditions most associated with each land type.
 
Results and Discussion
Spatial Distribution of the Fires
There were strong spatial contrasts in the frequency of wildfires between Iranian provinces. Khuzestan, Fars and Ilam were identified as the most fire-prone regions, as their large burnt areas occurred consistently throughout the two decades studied, due to arid climatic conditions, better vegetation cover and a high frequency of wildfire incidents.
The low threat in Yazd, Qom and Semnan provinces can be explained by the large area of arid environment and sparse vegetation.
Distribution across time and seasonality
The temporal analysis highlighted the strong seasonality of forest fires, dominated by the summer months with high temperatures and low humidity. Spring events may be related to the accumulation of dry vegetation and the rise in temperature from March to May. Among the seasonal patterns, the lowest level of activity occurred in winter, from December to February, due to generally cooler and wetter conditions that neutralised the fire hazard.
Places and Time-Rime of High Risk
It is this integration of spatial and temporal analysis that has made it possible to outline high-risk areas and critical periods for fires in Iran. Regionally, the most vulnerable parts of the country included the south-western and western provinces of Khuzestan, Fars and Ilam respectively, essentially defining the areas where wildland fire management needs to be prioritised.
Within these regions, late spring and summer have been identified as the period when conditions are most conducive to fire outbreaks.
Factors of Wildfire Risk
The result of this study indicated that climatic and environmental factors are dominant in describing fire hazard patterns in Iran. High temperatures with low humidity and abundant vegetation are the causes of the main factors of fire hazards. This factor is influenced by human activities, especially land clearing and agricultural fields, and unintentional fires.
All these natural factors interact with others of anthropogenic origin and therefore require a holistic approach to dealing with wildland fires, both in terms of prevention and response.
Consequences for Wildfire Management
The method also makes it possible to identify high-risk areas and critical periods, which is very important for policy makers and land managers. If resources were directed to these specific areas and times, it is likely that better wildfire prevention and control could be achieved. The establishment of firefighting units, early warning systems and community involvement in high-risk provinces can avert most of these influences on wildfires.
The integration of traditional fire management with modern remote sensing and GIS technologies could help to increase the effectiveness of such an attempt to manage wildfires.
Recommendations
Based on the findings of this research study, the following recommendations are made to improve the management of wildfires in Iran:

Focus on high-risk provinces: Concentrate efforts and resources on the high-risk southwestern and western provinces. Increase the number of fire stations, firefighting equipment and regular training of personnel to improve preparedness and response capabilities.
Seasonal fire monitoring and early warning: Apply seasonal monitoring, supported by satellite data analysis, to detect early signs of fire activity. For early detection of fire risks, establish an early warning system to alert authorities and people in advance to take necessary precautions in time.
Public awareness and participation: Educate the public about forest fires through community awareness to prevent the recurrence of fires. Educational campaigns such as safe agricultural practices, responsible land use and prompt reporting of fire incidents will be emphasised. Local communities, if involved in fire management plans, would feel more responsible and cooperative in protecting natural resources.
Policy and legislation: Develop and implement policies and legislation on wildfire prevention, land management and fire management. Activities that increase the risk of wildfires, such as land clearing and uncontrolled burning, should be controlled. Increased penalties for unlawful acts that cause fires could act as a deterrent to certain behaviours, with better compliance with prescribed fire safety regulations.
Research and development: There is a need to further develop the models used to predict wildland fires, including an understanding of the new dynamics of fire behaviour brought about by climate change. International organisations, government agencies and research institutions increase the likelihood of innovative solutions to wildland fires.

 
Conclusion
The wildfire distribution analysis developed here for Iran indicates the high-risk regions and their critical management periods. As fire risks are spatially and temporally distributed, the results will be very important for targeted wildfire management strategies. The result will be an overall capability for Iran to prioritise high-risk areas, develop early warning systems and engage local communities in ways that will ensure maximum fire prevention and control capability for Iran, while better protecting its natural resources and significantly reducing the socio-economic impacts of such destructive events.
It is this integration of advanced satellite remote sensing and GIS technologies that would be fundamental in achieving the stated goals and objectives of moving towards fire management practices that create harmony, resilience and sustainability.
 
Funding
There is no funding support.
 
Authors’ Contribution
All of the authors approved thecontent of the manuscript and agreed on all aspects of the work.
 
Conflict of Interest
Authors declared no conflict of interest.
 
Acknowledgments
We are grateful to all the scientific consultants of this paper.