Forest fire risk uncertainty analysis based on Dempster Shafer's theory of evidence (Study area: part of Kermanshah forests) | ||
| مدیریت مخاطرات محیطی | ||
| Article 6, Volume 8, Issue 4, January 2022, Pages 413-430 PDF (1.34 M) | ||
| Document Type: Applied Article | ||
| DOI: 10.22059/jhsci.2022.334588.690 | ||
| Authors | ||
| Zeinab Neisani Samani1; Ali Asghar Alesheikh* 2; Najmeh Neysani Samany3 | ||
| 1Ph.D. Student, Department of GIS, Faculty of Geodesy and Geomatics Engineering, K. N. Toosi University of Technology, Iran | ||
| 2Department of GIS, Faculty of Geomatics Engineering, K. N. Toosi University of Technology, Tehran, Iran | ||
| 3Associate Professor, Faculty of Geography, Department of RS and GIS, University of Tehran, Iran | ||
| Abstract | ||
| Introduction Every fire accident is accompanied by an uncertainty factor that is caused by human or natural factors. In general, fire hazards are assessed through a variety of predictive models based on the history of large fires. However, researchers and experts point to uncertainty-based fire modeling as one of the challenges to verifying the results. Therefore, the ultimate goal of this paper is to model the fire potential of fire in forest areas by considering the uncertainty due to the weighting of effective spatial criteria based on the Dempster-Shafer Intuition Theory. The innovation of the present study is the application of the Dempster-Shafer Intuition Theory of intuition to reduce the uncertainty caused by the weight of criteria by experts. The necessity of conducting this research, considering the large number of sub-criteria, is managing uncertainty by using the simultaneous use of belief and justification functions, which highlights the possibility of managing uncertainty due to the importance of fire prediction discussion. Research Methods First, the effective criteria along with the sub-criteria related to the review of previous research as well as the opinions of experts were identified; Then, a weighting process was performed using the Dempster-Shafer Intuition Theory. To weigh the effective criteria, the opinions of 30 experts were used. Finally, the weighted overlap of the criteria is done and the desired fire hazard map is prepared. The results are evaluated using the ROC curve. Determining effective criteria To identify and determine the effective criteria and sub-criteria, the results of previous researches and experts' opinions were used. Finally, 4 criteria include topography, climate, human, and vegetation were considered. Quantitative criteria such as rainfall, temperature, and distance from residential areas that have a direct impact on fire risk were considered for normalization. Criteria and sub-criteria were defined at 3 levels. Then, using the opinion of 30 experts, criteria and sub-criteria were ranked at different levels about each other and based on their importance in fire risk based on the characteristics of oak forest basins located in Islamabad west and Hamil cities. Since there is uncertainty in the opinion of experts, Dempster-Shafer's theory of intuition and the rule were used to integrate opinions, eliminate uncertainty and calculate the importance of each of the sub-criteria and criteria in this study. The theory of intuition and the Dempster-Schaffer rule This theory is developed by discussing existing beliefs about a situation or a system of situations. As the belief structure of the control theory is related to the classical probability model. Among the introductory concepts of evidence, the following can be noted: Diagnosis framework, Function to belief and justification, Belief range, Laws of the composition of evidence Research scope The study area is part of the oak forests of Kermanshah belonging to the west Islamabad and Hamil regions. The climate of this region is temperate Mediterranean and the average annual rainfall is 478 mm. Discussion and findings To implement the proposed research method, the maps of the criteria were prepared. Experts were used for the desired intervals of the sub-criteria and were modified if necessary. The criteria and sub-criteria were ranked about each other at different levels. Finally, using the Dempster-Shafer Intuition Theory, the final weights for the criteria and sub-criteria at each level were calculated. The effective parameters of the second level have been calculated using the existing sub-criteria in the third level. The fire hazard for the research area was classified into 4 different classes. Comparison of the resulting map with fires that occurred in the study area shows that high-risk and dangerous classes have a high overlap with points. The ROC method was also used to evaluate the results. Conclusion In the present study, Dempster-Schaefer's theory of intuitive reasoning was used to eliminate the uncertainty in the opinions of experts and to determine the final weight of the criteria and sub-criteria affecting the fire risk. Criteria and sub-criteria effective in causing fire were determined and then classified at different levels. Then, using weighted overlap methods, the criteria according to the final weight obtained from combining the opinions of experts using Dempster-Shafer Intuition Theory, fire hazard map of oak forests in west Islamabad was obtained. Finally, the obtained hazard map was normalized. Then, according to the mean values and standard deviation of the normalized hazard map, the research area was classified into 4 classes in terms of fire risk. The results showed that the most effective indicators in the occurrence of fire are vegetation and then humans, respectively. Comparison of the resulting map with fires that occurred in the research area shows that high-risk and dangerous classes have a high overlap with points. Also, model testing by the ROC curve shows the high accuracy of the model with a value of 92%. | ||
| Keywords | ||
| Modeling; Uncertainty; Fire; Dempster-Shafer; Weighted Overlay | ||
| References | ||
|
[1]. Bui, D.; Hoang, N-D.; & Samui, P. (2019). Spatial Pattern Analysis and Prediction of Forest Fire Using New Machine Learning Approach of Multivariate Adaptive Regression Splines and Differential Flower Pollination Optimization: A case study at La Cai province (Vietnam). Journal of Environmental Management. 10.1016/j.jenvman.2019.01.108.
[2]. Khan, M.; & Anwar, S. (2019). Paradox Elimination in Dempster–Shafer Combination Rule with Novel Entropy Function: Application in Decision-Level Multi-Sensor Fusion. Sensors. 19. 4810. 10.3390/s19214810.
[3]. Pique, M.; Olabarria, J.; & Reynolds, K. (2019). Strategic and tactical planning to improve suppression efforts against large forest fires in the Catalonia region of Spain. Forest Ecology and Management. 432. 612-622. 10.1016/j.foreco.2018.09.0.
[4]. Murthy, K.; Sinha, S.; Kaul, R.; & Vaidyanathan, S. (2019): A fine-scale state-space model to understand drivers of forest fires in the Himalayan foothills. Forest Ecology and Management, 432: 902–911. https://doi.org/10.1016/j.foreco.2018.10.009 | ||
|
Statistics Article View: 557 PDF Download: 357 |
||